AI News Digest 2026-10-06
台本で使った記事
特集
開発者コーナー
中堅コーナー
AIツール紹介コーナー
速報コーナー
参考記事一覧
参考記事一覧を表示
- ChatGPT's new ad format fills the image generation loading screen with product carousels [importance:40 dev:20]
- OpenAI will watermark ChatGPT text in the EU but makes it optional for API users worldwide [importance:60 dev:70]
- Instinct puts its AI agent in group chats, and non-users can join the thread [importance:50 dev:80] (いいね相当スコア: 0)
- Trump launches "Super Intelligence Force" that has nothing to do with actual superintelligence [importance:25 dev:10]
- Aleph Alpha releases Kolibri, an open-weight model that makes the case for European AI sovereignty [importance:75 dev:90]
- OpenAIの安全報告書責任者が退社 「文化が壊れている」と寄稿 [importance:55 dev:35]
- DeepSeek Harness v0.2 Brings Official Desktop Apps to Its Open-Source Agent Harness [importance:70 dev:95]
- US-China AI gap narrows to record-low 3%, led by DeepSeek: Report | DeepSeek’s V4.1 Flash ranked 6th in Sept | Inshorts - Inshorts [importance:50 dev:50]
- DeepSeek and Huawei Take Aim at Nvidia’s Formidable Moat - Bloomberg.com [importance:55 dev:40]
- CIP, Grok Ventures Back U.S.-Based Antora Energy’s Massive Thermal Battery Project in Kansas - carboncredits.com [importance:5 dev:5]
- Elon Musk's xAI stops new law prohibiting AI-powered 'nudification' tools by arguing free-speech violations - TweakTown [importance:45 dev:20]
- Grok Dominates Coinbase's AI Trading with a 60% Share: Which Coins Are Being Purchased? - 24/7 Wall St. [importance:35 dev:30]
- Mold Linker Version 3.0.0 Release – Rewritten in Rust [importance:45 dev:85]
- Germany's RobCo hits $1B valuation [importance:35 dev:20]
- Web Search API [importance:65 dev:90]
- Denmark Data Breach Exposes 8.8M People's Personal Data [importance:50 dev:30]
- Huawei and Qualcomm announce broad patent license agreement [importance:40 dev:30]
- Anthropic reported diary entry to police, woman faces felony charge [importance:60 dev:40]
- In the wake of Tippett Studios' closure, a digital archive appears online [importance:20 dev:15]
- The technology to eradicate mosquito-borne disease exists [importance:35 dev:10]
- ReviewBench: An open benchmark for AI code review [importance:70 dev:95]
- Everything we launched during Birthday Week 2026 [importance:50 dev:70]
- One year later: the power of 1.1.1.1 interns [importance:25 dev:20]
- dotInsights | October 2026 [importance:15 dev:60]
- Java Annotated Monthly – October 2026 [importance:20 dev:65]
- TeamCity 2026.2.1 and 2026.1.5 Are Out [importance:35 dev:80]
- Discontinuing Swift Language IDE Support in the Kotlin Multiplatform Plugin [importance:30 dev:75]
- Connecting AI agents to enterprise knowledge [importance:55 dev:60]
- Bringing predictive analytics to the agentic AI era [importance:50 dev:55]
- People really hate AI, so why can’t they get enough? [importance:30 dev:20]
- EmTech Future 2026: When AI Meets Everything [importance:45 dev:50]
- AI glasses face their first major government crackdown [importance:50 dev:25]
- TikTok rolls out an AI shopping assistant and one-click checkout [importance:35 dev:25]
- Hot Girl Hotline is like ‘Dear Abby’ for the AI era [importance:20 dev:20]
- HackerRank’s AI interviewer offers a glimpse into what job interviews could become [importance:40 dev:50]
- Open or closed AI? How founders are choosing what to build on at TechCrunch Disrupt 2026 [importance:25 dev:45]
- Researchers are tracking a Chinese AI ‘agent fleet’ [importance:50 dev:30]
- Meet the Startup Battlefield 200 judges who’ll decide the winner at TechCrunch Disrupt 2026 [importance:15 dev:20]
- The final Disrupt Stage lineup: Three days of conversations you won’t hear anywhere outside of TechCrunch Disrupt 2026 [importance:15 dev:20]
- Can Safeworld convince people that GenAI robots won’t hurt them? [importance:35 dev:25]
- Google froze its open source bug bounty program due to a ‘significant rise’ in AI submissions [importance:45 dev:50]
- Can ‘super intelligence’ and a non-binding safety pact solve AI’s image problem? [importance:35 dev:15]
- OpenAI PR tells journalist to ‘move on’ while asking Sam Altman about a ChatGPT user’s suicide [importance:40 dev:10]
- Sam Altman says ‘some bad things’ will happen, but AI is totally worth it [importance:35 dev:10]
- Sen. Adam Schiff on AI regulation, free speech, and impeaching Trump one more time [importance:35 dev:20]
- An open-source tool lets you delete 12GB of Apple Intelligence data on macOS [importance:30 dev:70]
- Our minds aren’t equipped to handle AI [importance:25 dev:15]
- NJ’s former Lt Gov is using AI to say he’s innocent of sexual harassment [importance:20 dev:10]
- An AI couldn’t beat humans at StarCraft, so it decided to cheat [importance:30 dev:40]
- Akka Tests Spec-Driven AI Delivery Across 65 Open Source Projects [importance:65 dev:90]
- Presentation: Building Reusable Evaluation Frameworks for Agentic AI Products [importance:65 dev:85]
- Article: The Platform Engineering Playbook for Production LLMs [importance:70 dev:90]
- Pizza Bot: Open-Source Inbox for Background AI Agents [importance:70 dev:95]
- 6 Guidelines for Governing AI [importance:40 dev:30]
- Attempts to Keep Humans in the AI Loop May Actually Push Them Out [importance:55 dev:70]
- AI Solves a Major Unsolved Math Problem. Not Everyone Is Happy [importance:70 dev:40]
- Meta and Microsoft pull back from Claude as Anthropic transforms from partner into competitor [importance:60 dev:50]
- Reka AI's omni-model Rho-1 handles text, images, video, and robot control in a single model [importance:75 dev:80]
- Most Americans want AI development to slow down or stop entirely, new poll finds [importance:45 dev:15]
- Anthropic is quietly becoming America's biggest corporate donor ahead of its mega IPO [importance:35 dev:10]
- AI is eroding office hours, study groups, and the trust between faculty and students, MIT report finds [importance:50 dev:30]
- Google researchers find a way to keep self-improving AI agents from memorizing their tests [importance:0 dev:0]
- 2026年9月28日〜10月5日：公開と仕様変更が並んだ5本 [importance:65 dev:85]
- GitHub、Appの新規トークンを約520文字へ切替、検証ヘッダーは11月30日廃止 [importance:75 dev:95]
- Citrix、SAML構成のNetScalerで未認証DoSとなる欠陥を修正 [importance:50 dev:60]
- 公開GitHubに有効な認証情報が54万件超、失効の有無で残存率に差 [importance:60 dev:60]
- llama.cpp、選択肢を1回の推論で確率付き返す決定モデルに対応 [importance:60 dev:90]
- OpenAI、保護された推論の組織的抽出を阻止しKimi側を中核と分析 [importance:65 dev:60]
- Microsoft、エージェントの成否をDB終了状態で測るThinkingBoxを公開 [importance:70 dev:95]
- Google装う「App password created」フィッシング 偽ログインはGoogle Sites上 [importance:40 dev:30]
- 高専卒の求人倍率が35.3倍に IT・AI人材需要で5年前の24.7倍から上昇 [importance:35 dev:15]
- Cloudflare Workers Buildsでビルドが始まらない障害 修正後も監視中 [importance:35 dev:60]
- Cloudflare Quick Tunnels、メール認証で閲覧者を限定可能に [importance:45 dev:70]
- 日経、Microsoft 365アカウント乗っ取りでなりすましメール約9000件 [importance:40 dev:25]
- 情報系の入学が米英豪で減少、IT責任者の45%は新卒需要減を見込む [importance:55 dev:25]
- AWSの管理MCPサーバーが東京を含む6リージョンに追加 [importance:0 dev:0]
- GitHub ActionsのmacOS 14を11月2日に廃止、10月5日から計画停止 [importance:45 dev:85]
- AWS、SageMaker Studioの起動処理にコマンド注入の欠陥を修正 [importance:0 dev:0]
- GitHub Copilot、Claude Opus 4.7など4モデルを即日廃止 [importance:55 dev:85]
- GitHub、Copilotの手順をコードで固定するDynamic Workflowsを公開 [importance:70 dev:90]
- Yandex Introduces Sona: A Single Generative Recommender That Replaces Entire Recommendation Cascade [importance:65 dev:80]
- The Story of Qwen: Alibaba’s AI Models From 7B to 2.4T [importance:60 dev:70]
- Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks [importance:65 dev:85]
- GPT-6 Astra vs GPT-6.1 Sol vs Gemini 4 Argon vs Claude Fable 5.1: Which Frontier Model Fits Which Job [importance:70 dev:85]
- Google Research Moves Federated Learning Into TEEs: Gboard Now Trains With Externally Verifiable Differential Privacy [importance:65 dev:75]
- Inside NVIDIA’s IsaacTeleop: From Hand and Controller Tracking to Robot Actions with the Graph-Based Retargeting Engine [importance:55 dev:80]
- MintFlow: Minimal Trajectory Intervention for Constrained Flow Matching [importance:40 dev:70]
- Fast Models, Slow Evidence: A Paired and Self-Audited Evaluation of System-1 Decision Models for LLM Agent Harnesses [importance:65 dev:90]
- The AI Risk Observatory: What Can We Learn from AI Disclosures in Annual Reports About Societal Resilience? [importance:50 dev:55]
- Keep It CALM: Analyzing the Limits of Global Unsafety in Text-to-Image Generation [importance:45 dev:75]
- Choosing Before Acting: Comparative Value Estimation for Long-Horizon Tool-Use Agents [importance:60 dev:85]
- World Editing: Intervening on Executable Worlds at Increasing Depth [importance:55 dev:80]
- A Multi Method Importance and Performance Efficiency Analysis of Topological Metrics for Natural Visibility Graph Based Cyber Attack Detection [importance:35 dev:75]
- DeReAct: Decomposed Reasoning and Acting for Reliable AI Agents [importance:65 dev:90]
- Traversing the Satisfaction-Diversity Frontier in Text-to-Image Diffusion [importance:55 dev:75]
- THPL: A Vision-to-Language Decision Support Framework for Rainbow Trout Feeding Management in RAS [importance:40 dev:50]
- FlashSinkhorn 2: Block-Sparse Entropic Optimal Transport [importance:35 dev:70]
- When Terminal-Agent Training Stalls: Demystifying Data Generation and Verification Challenge [importance:65 dev:85]
- Reinforcement Learning Techniques for the Optimization of Target Polarization in Nuclear Physics Scattering Experiments [importance:40 dev:70]
- Tropical Reinforcement Learning [importance:50 dev:75]
- MEA: A Reward-Driven Multi-Agent System for Faithful Model Explanations [importance:55 dev:60]
- What Does a Token Cost? A Mixture-of-Agents Measurement of Sufficient Per-Token Compute [importance:60 dev:75]
- Right Order, Wrong Scale: Auditing LLM Judges for Occupational AI Measurement [importance:55 dev:70]
- "I just assumed that it would translate": examining MT risk awareness among healthcare staff with abbreviations as a use case [importance:35 dev:35]
- HXAI: Hierarchical Privacy-Preserving Explainable AI in Distributed Energy Systems [importance:45 dev:60]
- World Action Modeling with Progressive Visual Planning [importance:50 dev:70]
- On-Premises Multi-Course RAG Tutoring for Business Education: Hardware-Software Trade-offs in a Campus AI Tutor [importance:55 dev:75]
- Hypothesis-guided discovery of cognitive algorithms via program refinement [importance:45 dev:50]
- Learning What to Investigate Next: Meta-Reasoning for Long-Horizon Research Agents [importance:65 dev:80]
- How To Train Your World Model: Fine-tuning vs RAG for LM-based World Modeling [importance:65 dev:80]
- How to Have a Sensitive Debate: An Instance-Optimal Protocol for AI Debate [importance:65 dev:75]
- Mitigating Social Sycophancy via Pluralistic Preference Optimization [importance:55 dev:70]
- Answering clinicians' questions over trial evidence tables with verifiable, feedback-driven language models [importance:55 dev:75]
- Labels Override Definitions in Jev-Style Typed Decision Models [importance:50 dev:70]
- Open-Endedness Bench: Measuring Epistemic Process from Agent Records [importance:65 dev:80]
- TasteBench: Multimodal Benchmark for Sensory Prediction, from Molecules to Sustainable Foods [importance:50 dev:65]
- Time Series Forecasting Benchmarks Need Scenario-Grounded Stress Testing [importance:60 dev:80]
- VERSE: Verified Self-Evolving Optimizer for Agent Harnesses [importance:65 dev:85]
- CuBEs: Culturally-Situated Behavioral Evaluations and the Limitations of Culture-Blind LLM Judges [importance:55 dev:70]
- Lost in the Request: How Communication Variation Disrupts Retrieval and Action in Email Agents [importance:60 dev:80]
- Designing the Future of User Feedback for Generative AI [importance:55 dev:70]
- Batched Speech Decisions Without Decoding: Single-Token Supervision Lets a Frozen LLM Hear Beyond the Transcript [importance:65 dev:85]
- Coherence-Driven Belief Formation and Population Dynamics of Contagion in LLM Agents [importance:55 dev:75]
- A GHOST in Long-Horizon Agents: Governance Hazard from Overlooked Safety Constraints across Turns [importance:70 dev:85]
- Spend Teacher Tokens Where They Matter: Success-Referenced On-Policy Distillation [importance:60 dev:85]
- DataWeave: Deploying Human-LLM Analytics for Exploratory Structured Data Analysis [importance:60 dev:80]
- Large language models exhibit unreliable updating of clinical judgment as patient evidence evolves [importance:50 dev:70]
- Decoupling Memory from Context: Structured Memory for Token-Efficient Test-Time Continual Learning [importance:60 dev:80]
- Learning to Revise Reasoning with Segment-wise On-Policy Distillation [importance:60 dev:80]
- Label-Efficient Time Series Classification at Scale: A Dual-Stream OSSE-LSTM with Counterfactual Attribution [importance:60 dev:80]
- Ego2World: Compiling Egocentric Cooking Videos into Executable Worlds for Belief-State Planning [importance:55 dev:70]
- On the Chain-of-Thought Monitorability of Looped Language Models [importance:60 dev:80]
- Dynamic LLM Routers are Often Misguided [importance:70 dev:85]
- Law And Order: Tax Law Autoformalization [importance:55 dev:80]
- PAPER2LLM++: Continual Self-Evolution of LLMs from Research Papers [importance:65 dev:85]
- Modeling Shared and Individual Structure for Cross-Subject Continuous Affect Regression from EEG-fNIRS [importance:45 dev:70]
- BitNest: Bit-Nested Speculative Decoding for Memory-Efficient LLM Inference Acceleration [importance:65 dev:85]
- VIGOR: Zero-Shot Visual Generalization via Latent-Space Consistency in Model-Based Reinforcement Learning [importance:55 dev:80]
- ROUTEAUDIT: Interaction-Aware Identification for Budgeted Multi-Verifier Routing [importance:55 dev:85]
- iS-KV: Online Low-Rank KV Cache Compression via Block-Incremental SVD [importance:65 dev:85]
- MetaRubric: Learning to Reward for Rubric-Based Reinforcement Learning [importance:60 dev:85]
- Scaling Trajectories for Complex Tasks through Recursive Self-Rewrite [importance:60 dev:85]
- MLCommons Jailbreak Benchmark v1.0 [importance:70 dev:85]
- FSPO: Policy-Consistent Risk and Pareto-Feasible Control for Budgeted LLM RL Post-Training [importance:70 dev:85]
- AMBER: Multi-View Adaptive Budget Allocation for Listwise Vision-Language Reranking [importance:55 dev:80]
- DNAlign: Dynamic Null-Space Safe Alignment for LLMs [importance:70 dev:85]
- Bounded Reachability & Jailbreak Detection via Contraction-Constrained State Space Models [importance:65 dev:85]
- Harness-Aware Distillation for Small Language Model Agents [importance:60 dev:85]
- TACD: Distilling Efficient Text-to-Motion Models via Terminal Amplification Control [importance:50 dev:80]
- PsyEvo: A Personalized Counseling Agent That Self-Evolves at Test Time [importance:50 dev:70]
- Interpreting at Write Time: A Policy Ablation for Multi-Goal Agent Memory [importance:60 dev:80]
- LUMOS: Tracing Parametric Knowledge from Training Data to Behavioral Outputs in LLMs [importance:60 dev:80]
- Frequency Is Not Sensitivity Identifying Safety-Sensitive Experts in Sparse MoE LLM [importance:60 dev:85]
- HASTE: Evolving Agent Harnesses Against Emerging Attacks Using Sparse Evidence [importance:70 dev:85]
- Positive-Unlabeled Learning for Agent Safety False Alarm Auditing [importance:60 dev:85]
- When to Compile a Computer-Use Agent? Measuring Payback and Making Compilation Decisions for Token Efficiency [importance:60 dev:85]
- Continual Graph Memory for Mathematical Research Agents [importance:60 dev:85]
- Reasoning with Evidence, Not Merely Rationales: Verifiable Preference Proofs for LLM-Based Recommendation [importance:60 dev:80]
- CreateScore: Domain-Theory-Informed Bayesian Routing for LLM-Based CV Screening [importance:55 dev:85]
- Reliable Self-Evolution with Imperfect Proxy Rewards [importance:60 dev:85]
- Relevant Evidence Decoding for Audio-Visual Hallucination Mitigation [importance:60 dev:85]
- RASPER: Reward-Aligned Summarization of Clinical Notes for EHR Outcome Prediction [importance:50 dev:80]
- Safeguarding Mutual Correction in Source-Free Domain Adaptation via Cut Statistics [importance:50 dev:80]
- PLCWorld: Benchmarking LLM-Generated PLC Programs in Closed-Loop Plant Simulation [importance:55 dev:85]
- Personalized Automatic Speech Recognition for a Dysarthric and Tracheostomic Speaker using Artificial Conversations [importance:45 dev:70]
- DyadMem: A Long-Term Memory Benchmark of How Agents Work with Users [importance:60 dev:85]
- Verifiable, Articulable, and Tacit Components of Preference [importance:50 dev:70]
- SoftGene: Protein Language Model-Enhanced Soft Prompting for Interpretable Gene Set Annotation [importance:45 dev:70]
- When Numbers Start Talking: Numerical Signalling and Strategic Behaviour Among LLMs [importance:55 dev:80]
- hacktrace: behavior-supervised detection of reward hacking during code generation [importance:70 dev:85]
- MOF-VERIFY: A Failure-Aware Agentic Harness for MOF Hypothesis Verification [importance:55 dev:80]
- RIFAR: Reliability and Forgetting-Aware Replay for Continual Robot Learning [importance:55 dev:80]
- Peer Influence across Heterogeneous AI Models [importance:55 dev:80]
- Predictor-Guided Latent Space Codon Optimization for Maximizing Protein Expression [importance:45 dev:70]
- Trading Strategy Optimization via Textual Gradient [importance:55 dev:80]
- Keeping JEPA World Models Plannable When Little of the Frame Moves [importance:55 dev:80]
- Gains and Collapse in On-Policy Distillation:A Reinforcement Learning Perspective [importance:60 dev:85]
- KV$^2$: A Self-Refining KV Cache [importance:65 dev:85]
- Toward SLM-based agentic task-tool intent matching [importance:60 dev:85]
- Learning a Fact Is Not Learning How to Retrieve It [importance:55 dev:80]
- EVOL: Simulator-Guided Evolutionary Expert Synthesis for Deployment-Free Learning Path Recommendation [importance:55 dev:80]
- JOVE: Joint Execution and Verification for Resource-Aware LLM Task Graphs [importance:60 dev:85]
- Optimal Planning in a Dynamic World [importance:50 dev:70]
- Lightweight, Rubric-Guided Trajectory Evaluation for Production AI Agents [importance:60 dev:85]
- Multi-Task Evolution for Zero-Shot Cross-Problem Generalization using LLMs [importance:55 dev:85]
- Refinement Buys Intelligibility, Search Buys Identity: What Test-Time Compute Buys in Masked-Diffusion TTS [importance:55 dev:85]
- Preserving Mathematical Reasoning in Compressed Diffusion Language Models via Trajectory-Aware Low-Rank Approximation [importance:60 dev:85]
- ReFract: Benchmarking Perspective Awareness in Language Model Agents with Text World Models [importance:60 dev:85]
- Geometry Meets Physics: Data-Efficient Pre-Training for Unstructured Neural PDE Solvers [importance:50 dev:80]
- Multilingual GSM-Symbolic: What determines capability transfer across languages? [importance:55 dev:80]
- CVE2AP: Automated Generation of PDDL-Encoded Attack Paths via Large Language Models [importance:60 dev:85]
- Benchmarking Candidate Coverage in Typed Decision Models [importance:45 dev:70]
- Becoming Suspicious Across Borders: Algorithmic Extraterritoriality and AI-Driven Financial Surveillance [importance:50 dev:65]
- Jumping the Line: Exploiting Length Predictions in LLM Scheduling [importance:60 dev:85]
- A Near-Zero Monitor Readout Is Not Evidence of Behavioral Control [importance:70 dev:85]
- Efficient Reasoning Training Does Not Always Harm CoT Faithfulness and Monitorability [importance:60 dev:85]
- Reasoning Models Are Accurate but Unsound on Identification [importance:55 dev:80]
- From Benchmarks to Production: A Text-to-SQL System for Complex Financial Data [importance:60 dev:85]
- Recursive Harness Self-Improvement for Frontier Reasoning Data Synthesis [importance:60 dev:85]
- Knowledge or Calculator? Decomposing the Skill Premium in Verifiable Financial Agent Workflows [importance:60 dev:85]
- Learning to Assess Heartbeat Observability for mmWave Heart-Rate Sensing [importance:30 dev:15]
- HyperBrowseComp: A Multilingual and Multimodal Stress Test for Web-Browsing Agents [importance:55 dev:70]
- HazardWeaver: Scientific Route Selection for Hazard Analysis Agents [importance:45 dev:55]
- Low-Cost Video--Time Priors as a Strong Baseline for EEG--fNIRS Emotion Regression on Familiar Videos [importance:20 dev:10]
- Depth as Time in One-Step Generative Models [importance:60 dev:70]
- NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents [importance:50 dev:65]
- Credit Where It Matters: Dependency-Aware Policy Optimization for Terminal Agents [importance:65 dev:75]
- Do Large Language Models Know Colombian Law? A Reliability Benchmark for the Colombian Legal System [importance:35 dev:50]
- MRVQ: One Resident Index for Dimension- and Rate-Elastic Vector Search [importance:55 dev:75]
- Transcriptome-informed multi-modal AI for predicting neoadjuvant therapy response from breast cancer biopsies [importance:40 dev:40]
- IntentCoding: Amplifying User Intent in Code Generation [importance:60 dev:80]
- Multi-Modal Environment-Aware Beam Management for Massive MIMO: A Geometry-Driven Virtual Base Station Framework [importance:25 dev:20]
- ClarifyCodeBench: Evaluating LLMs on Clarifying Ambiguous Requirements for Code Generation [importance:60 dev:80]
- GlanceWAM: Sparse Test-Time Imagination for World-Action Models [importance:50 dev:70]
- CONTRA: Discovering and Qualifying Behavior-Changing Questions for Selective Clarification in LLM Code Generation [importance:55 dev:75]
- Causal discovery identifies pathways linking physical activity to dementia risk in the UK BioBank [importance:40 dev:50]
- CORE: COverage CAlibration and Evicted-Mass REdistribution for KV Cache [importance:70 dev:85]
- Hardware-Native Joint Sparse-Quantization for Trillion-Scale Mixture-of-Experts [importance:75 dev:90]
- RxnOptBench: Benchmarking LLMs for Reaction-Condition Optimization in Organic Methodology [importance:45 dev:60]
- Toward Controlling Biology with Language:Offline Learning of Prompt-Conditioned Interventions for Cells, Organoids, and Biobots [importance:50 dev:50]
- Counterfactual Predictions in Scientific Emulators Without Controlled Experiments [importance:45 dev:65]
- Overcoming Challenges of Interpretive Structural Modeling with Large Language Models [importance:35 dev:50]
- Diffusion-Based Synthetic Data Pretraining for Enhancing Activity Recognition [importance:50 dev:70]
- EditHero: A Benchmark for Long-Horizon Part-Level 3D Editing and Vibe Modeling [importance:50 dev:70]
- $\Psi$-Resilience: Model-Free Feature Importance from 1D Topological Signals [importance:35 dev:70]
- SimuVerity: Benchmarking Agents for Engineering-Grade Simulink Model Generation [importance:55 dev:80]
- DeskForge: Dense Supervision from Desktop Environments for Computer-Use Agents [importance:65 dev:85]
- Slow-Fast Multi-Teacher On-Policy Distillation for Capability Preservation [importance:55 dev:80]
- MIRROR: Multipath Quorum Integrity for LLM Multi-Agent Communication [importance:60 dev:80]
- Lexicographic Multi-Objective On-Policy Distillation [importance:55 dev:80]
- Automating the Application of HCI Principles: Skills for On-Demand UI Construction, the Human-AI Space to Think, and the Future of HCI [importance:60 dev:80]
- Network-in-the-Loop at Scale: GPU-Batched 5G Simulation for Massively Parallel Robot Learning [importance:50 dev:70]
- Hop-Decayed Influence: New Vulnerabilities of Structural Auxiliary Indexing in GraphRAG Pipelines with LLM [importance:60 dev:85]
- EviDent-CBCT: Evidence-Bottlenecked Report Generation from Dental CBCT under Non-Exhaustive Report Supervision [importance:35 dev:50]
- Coco: An Agentic Copilot for the Hardware--Software Co-Design Lifecycle [importance:65 dev:80]
- The Surprising Effectiveness of Shared Memory in Looped Transformers [importance:55 dev:80]
- Inherit-MAS: Test-Time Evolution of Multi-Agent Systems through Workflow and Execution Inheritance [importance:60 dev:80]
- Efficient Neural Field Learning via Adaptive Coverage and Focused Sampling [importance:45 dev:70]
- Mitigating Private Data Leakage in LLMs with Whiteout [importance:70 dev:85]
- Geometry-Aware Time Reparameterization for Flow-Map Distillation [importance:50 dev:80]
- Learning Style, Forgetting Semantics: A Case Study of SFT and RFT on Classification Tasks [importance:50 dev:75]
- Are you Synthesizing or Recalling? Evaluating LLMs on Algorithmic Code Retrieval [importance:55 dev:80]
- Counterexample Generation via Per-Theorem Symbolic Verifiers: When Imitation Hurts and Reinforcement Repairs [importance:55 dev:75]
- FinDialogLens: Event Extraction over Multi-Party Dialogue for Missed-Trade Identification in Financial Chatrooms [importance:40 dev:60]
- SideKernel: A Usable microVM Sandbox for AI Coding Agents on macOS [importance:65 dev:85]
- APDMem: Agent-Controlled Progressive Disclosure for Query-Adaptive Long-Term Memory [importance:60 dev:75]
- From Retrieval to Typed Decisions: Calibrated System One Models from Biomedical Sentence Encoders [importance:40 dev:65]
- Compound AI System Reliability: A Failure Taxonomy and Resilience Pattern Catalog from 150 Production Incidents [importance:75 dev:90]
- Multi-Fidelity Policy Gradients Stabilize Data-Scarce Reinforcement Learning [importance:55 dev:80]
- From Fragments to Global Maps: Learning Vectorized Map Aggregation with Large Language Models [importance:55 dev:70]
- IGNITE Tokamak World Model Architecture [importance:50 dev:65]
- Student-Guided Teacher Distillation for Efficient LLM Task Routing: Positioning Against Jev-Style System-1 Classifiers [importance:55 dev:80]
- Instance-Dependent Regret for CMDPs with Step-Wise Constraints [importance:40 dev:70]
- CriticHack: Evaluating Visual Rewards Under Robot Policy Optimization [importance:55 dev:75]
- Out of Sync, Out of Sight: Phantom State Attacks against IIoT Intrusion Detection [importance:50 dev:65]
- OpenGameEval: Benchmarking Agentic Programming and Exploration in a Stateful Game Engine [importance:50 dev:75]
- DAGS: Disentangled Appearance-and-Geometry Steering of a Frozen Image DiT for Temporally Stabilized Generative Rendering [importance:55 dev:80]
- Pincer: Resource Authorization for Agents using a Digital Twin [importance:65 dev:85]
- Improving the Energy-Efficiency of the Code Generated by LLMs through Effective Prompting [importance:60 dev:80]
- How Causality Bridges the Semantic Gap [importance:50 dev:75]
- WebUIProof: Benchmarking WebUI Code Generators with UI-Agent Execution Harness [importance:60 dev:85]
- Equivariant Flow Matching for Electron Density Prediction [importance:45 dev:70]
- Distributed Learning with Selective State Space Models: Architecture-Aware Convergence Analysis [importance:60 dev:85]
- Generalization Properties of Score-matching Diffusion Models for Intrinsically Low-dimensional Data [importance:55 dev:80]
- Large Language Continuous Diffusion Models [importance:70 dev:85]
- LEAP: Learning Efficient Action Proposals For LLM Agents [importance:60 dev:80]
- Test-time Calibration Learning for Large Language Model Reasoning [importance:60 dev:80]
- MuonIO: Principled Norm-Aware Descent for Embedding Tables and Language Model Heads [importance:60 dev:85]
- Self-Supervised Scaling of Terminal Environments for Scientific Domains [importance:55 dev:80]
- Revisiting Visual Representation Enhancement of VLMs via Kernel Canonical Correlation Analysis [importance:55 dev:80]
- TPBench: A Turning-Point Benchmark for Dialogue Compression [importance:50 dev:75]
- Prospective Hindsight: Self-Calibrating Reinforcement Learning via Prediction-Reality Gaps [importance:60 dev:80]
- Correcting Guided Diffusion Trajectories with Spectral Alignment [importance:55 dev:80]
- When History Fails to Become Experience: Action Calibration in Language Agents [importance:55 dev:75]
- Nearly Optimal Fixed-Confidence Best-Arm Identification with 1-Bit Feedback [importance:40 dev:70]
- Improving Atomic-Fact Recall via Focused Views in Unstructured Knowledge Editing [importance:60 dev:80]
- OPD Before RL: Warm-Starting Rubric-Based RL with On-Policy Distillation [importance:60 dev:85]
- Adaptive Spectral-Koopman Dynamics Modeling for Temporal Domain Generalization [importance:50 dev:75]
- FastOPD: On-Policy Distillation for Lightweight VLA Deployment [importance:60 dev:80]
- Distributionally Robust Survival Models under Subpopulation Shift and Outlier Contamination [importance:45 dev:65]
- AgentTrap: Stateful Feedback Deception against Autonomous Penetration Testing Agents [importance:65 dev:80]
- ConvoDrift: A Multi-Turn Conversational Dataset for Modeling Stylistic Tone Evolution [importance:45 dev:70]
- Query-aware routing for Cross-lingual performance gains in Encoders [importance:50 dev:75]
- DyRA: Dynamic Residual Approximation for Efficient Matrix Multiplication in DNNs [importance:60 dev:85]
- Misinformation Without Triggers: From Factual Answers to Downstream Decisions [importance:65 dev:80]
- Revealing Epistemic Uncertainty in MLLMs via Causal-Invariant Masking [importance:60 dev:80]
- When Can We Trust the Matching Principle? Robust Deployment Geometry Under Finite-Sample and Model Uncertainty [importance:50 dev:75]
- Discriminating Fixture Coverage in Agent-Infrastructure Verification Suites [importance:55 dev:80]
- Dynamic Expert Pruning for Multi-Agent Systems [importance:60 dev:85]
- Post-Training Frontier Text-to-Image Models by Composing Preference and Rubric Rewards [importance:65 dev:85]
- Sentry: Learning to Recover from LLM Agent Failures at Test Time [importance:65 dev:85]
- Temporal Geometry of Deep Networks: Hyperbolic Representations of Training Dynamics for Intrinsic Explainability [importance:50 dev:75]
- AvoKV-E: Payload-Aware KV Cache Eviction for Long Reasoning [importance:65 dev:85]
- Beyond Predefined Sinks: Security-Aware Dependency Analysis for LLM Agents [importance:65 dev:85]
- OmniAct3D: Leveraging Foundation Geometry and Evidence-Grounded Reasoning for Panoramic 3D Detection [importance:50 dev:75]
- Tailoring the Quantization Space for 1-Bit KV Cache Compression [importance:65 dev:85]
- WebFovea: When the Model Is Right but the Click Is Wrong -- Reliable Round Trips for Vision-Based Web Agents on Live Websites [importance:60 dev:80]
- NegT2IBench: When Negation Changes the Picture. A Polarity Benchmark for Text-to-Image Models [importance:55 dev:80]
- Zephon: Elastic Determinism for Online, Stateful Foundation Model Data Loading Pipelines [importance:65 dev:85]
- Securing Computer-Use Agents Against Branch Steering Attacks [importance:70 dev:85]
- ULTRADISCOVERY: Abductive Exploration in an Interconnected, Epistemically Open Universe [importance:25 dev:20]
- Beyond Single Videos: Benchmarking and Active Evidence Seeking for E-Commerce Cross-Video Reasoning [importance:30 dev:35]
- Ask, Relax, or Act? Evaluating Actionable Indeterminacy in LLM Preference Reasoning [importance:40 dev:55]
- S2S-JEPA: Predicting the Predictable at Subseasonal-to-Seasonal Timescales [importance:35 dev:30]
- How to Find and Reuse Policies for Continuous Adaptation in Lifelong Reinforcement Learning [importance:25 dev:30]
- Foresight: planning future perception in streaming VLMs without retraining [importance:35 dev:45]
- The Fragility of Trigger-Tag Mechanisms for Misuse Detection in Open-Weight LLMs [importance:55 dev:65]
- EvoRiskBench: An Evolving Benchmark for Runtime Security Risks in Workspace Agents [importance:70 dev:75]
- Multimodal reasoning for broadly neutralizing antibody discovery from label-free human B cell repertoires across virus families [importance:50 dev:20]
- Predicting Steering Vectors and Adapter Weights for Few-Shot Author-Style Transfer [importance:35 dev:45]
- LiBRA: Detection-Aware Image Watermark Removal via Bidirectional Latent Optimization [importance:40 dev:45]
- Not Until the Evidence Says So: Teaching LLM Investigators When to Close a Case [importance:40 dev:50]
- Contextual Flow Matching: Adaptive Step Selection in Flow Models for Efficient Visual Generation [importance:45 dev:55]
- Evolving Hybrid Quantum-Classical Architectures for Image Classification [importance:35 dev:40]
- Uncertainty as a Proxy for Semantic Correctness in Diffusion-Based Medical Image Synthesis [importance:40 dev:30]
- D2K-Bench: Can LLM Agents Turn Expert Designs into Efficient GPU Kernels? [importance:65 dev:75]
- WAMpy: Efficient Synthesis of Prolog Programs in Python [importance:40 dev:65]
- Mapping and Advancing the Scalability-Accuracy Frontier of Nonlinear Causal Discovery [importance:35 dev:40]
- Consecutive Posterior Fusion for Diffusive Recovery of Unobservable Image Structures [importance:35 dev:45]
- SPEAR: A Spectral-Disentangled MoE Neural Operator with Knowledge-Guided Expert Aggregation for Large-Scale PDE Pretraining [importance:40 dev:45]
- Training-Loss Guarantees for Muon with Finite-Step Newton--Schulz Orthogonalization [importance:40 dev:55]
- Information Limits of Low-Rank Approximation Certification [importance:25 dev:35]
- SyntaxBench: A Statistical Diagnostic Framework for Character-Level Reasoning in Large Language Models [importance:45 dev:60]
- Cordial Learning: Distributed Training with Correlated Data [importance:35 dev:50]
- Equivariant Visual-Tactile Diffusion Policy for Contact-Rich Manipulation [importance:40 dev:45]
- Follow the Winners: Conservative Policy Improvement with the Cross-Entropy Method for Critic-Free RFT [importance:45 dev:55]
- DriftTTS: Few-Step Text-to-Speech Without Distillation via Distribution-Matching Drift [importance:35 dev:40]
- ForestQuery: Boundary-Aware and Spatially Anchored Query Learning for Unified Forest Point Cloud Segmentation [importance:25 dev:30]
- Rethinking Epistemic Uncertainty in Node Classification through Information Growth [importance:30 dev:40]
- OptiSelect: How does the Optimizer Shape Data Curriculum? [importance:50 dev:60]
- Corrupted but Correct: Why Vision-Language Models Lie to Themselves Internally [importance:50 dev:55]
- Measure Less, Know More: Self-Supervised Test-Time Feature Acquisition [importance:40 dev:50]
- Preserving Anatomical Continuity: Three-Stage Pipeline for Colon Segmentation in 3D Abdominal CT Scans [importance:40 dev:30]
- Single or Multiple Policies for Phase-Structured Reinforcement Learning? [importance:35 dev:50]
- MobiAgent: Dual-Loop Recursive Policy Self-Improvement for Long-Horizon Mobile Manipulation [importance:50 dev:60]
- AREX: Affine-Residual Exponential Integrator for Few-Step Sampling in Flow Matching [importance:35 dev:50]
- Detect and Suppress: A Mechanistic Defense against Adversarial Patches in VLA Models [importance:50 dev:60]
- Certified Mechanistic Edits: Behavioral Guarantees for Skill Removal and Preservation [importance:55 dev:65]
- Weave Forcing: Compositional Memory Routing for Interactive Long Video Generation [importance:40 dev:50]
- Beyond Trained Models: Compiling GNNs for a Sound Explainer Benchmark [importance:40 dev:55]
- Cephalonauts One: A deep fMRI dataset for decoding naturalistic speech in the human brain [importance:35 dev:20]
- Rethinking What to Cache in Few-Step Diffusion Transformers: Solver-Aware Target Selection [importance:40 dev:55]
- Threat-Preserving Representation Sensitivity in Agent-Security Benchmarks [importance:55 dev:70]
- When a Correct Reward Is Not Enough: Diagnosing and Guiding PPO in an Analytically Solved Broker-Trader Game [importance:35 dev:45]
- LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation [importance:45 dev:55]
- On-Board Anomaly Detection for Efficient Marine Environmental Monitoring [importance:30 dev:30]
- Revisiting Input Time-frequency Representations in Multi-pitch Estimation for Vocal Ensembles [importance:25 dev:30]
- FrugalEvo: Towards Cost-Aware LLM-Guided Program Evolution [importance:55 dev:65]
- EyeRobot 2.0: Active Gaze for Precise Manipulation without Wrist Cameras [importance:45 dev:50]
- What Should World Models Forget? Stratified Retention for Continual Adaptation [importance:50 dev:60]
- 4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes [importance:45 dev:60]
- Less Decoder is More Encoder: Geometric Representation Learning from Novel View Synthesis [importance:40 dev:50]
- Heads, Tails, and AI Fails: LLMs, Randomness, and Human Judgments [importance:35 dev:45]
- AstroAgentBench: Evaluating Agentic Planning on Space Mission Planning Tasks [importance:50 dev:65]
- LPS-Bench: Benchmarking Safety Awareness of Computer-Use Agents in Long-Horizon Planning under Benign and Adversarial Scenarios [importance:65 dev:75]
- On the Tip of the Tongue: Why LLMs Hallucinate Answers They Can Decode [importance:55 dev:65]
- GISTBench: Evaluating LLM User Understanding via Evidence-Based Interest Verification [importance:40 dev:60]
- The Hitchhikers Guide to Rubric Quality Understanding and Enrichment [importance:50 dev:65]
- What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning [importance:60 dev:75]
- What do your logits know? [importance:55 dev:65]
- Offline Policy Optimization with Posterior Sampling [importance:40 dev:55]
- NARA: Anchor-Conditioned Representation Learning for Heterogeneous Vector Geoentities [importance:30 dev:40]
- ASH: Agents that Self-Hone in Long-Horizon Worlds [importance:0 dev:0]
- KGPFN: Unlocking the Potential of Knowledge Graph Foundation Model via In-Context Learning [importance:45 dev:60]
- HyperGuide: Hyperbolic Guidance for Efficient Multi-Step Reasoning in Large Language Models [importance:50 dev:65]
- One Hypothesis Is Not Enough: Abductive Reasoning with Agentic Hypothesis Refinement over Knowledge Graphs [importance:50 dev:65]
- Counterfactual Evidence Audits Predict LLM-Agent Susceptibility to Ranked Context [importance:55 dev:70]
- Planning Takes More Than Token Prediction: Causal Plan for Benchmarking and Building Physically Grounded Embodied Reasoners [importance:60 dev:70]
- Rethinking RAG in Long Videos: What to Retrieve and How to Use It? [importance:50 dev:65]
- Mental-R1: Aligning LLM Reasoning for Mental Health Assessment [importance:45 dev:35]
- Dissecting model behavior through agent trajectories [importance:55 dev:70]
- Latent Goal Prediction from Language for Model-Based Planning [importance:50 dev:65]
- SovereignPA-Bench: Evaluating User-Owned Personal Agents under Evolving Intent, Platform Mediation, and Consent Constraints [importance:55 dev:70]
- A Multi-Timescale Recursive Self-Improvement Engine for Open-Ended Persona Growth [importance:45 dev:60]
- Rethinking the Evaluation of Harness Evolution for Agents [importance:55 dev:75]
- How Artificial Intelligence LLM Engines Shape the Global Conflict Information Environment [importance:45 dev:30]
- ORACLE: Agentic AI Orchestrator Routing Via Adaptive Verifier Calibration Feedback [importance:60 dev:75]
- World Action Planner: Generalizable Robot Decision-Making with Action-Conditioned World Models [importance:55 dev:70]
- WAM-OPD: Joint Video-Action Supervision for World Action Model Post-Training with On-Policy Distillation [importance:45 dev:65]
- Will the User Ever Know? Covert Indirect Prompt Injection Attacks on Tool-Using LLM Agents [importance:65 dev:75]
- LayerRoute: Action-Conditioned Mixture-of-Layers Routing for Vision-Language-Action Policies [importance:55 dev:70]
- Learning transferable human physiology from two million hours of sleep with SleepFM-2 [importance:50 dev:25]
- BusMA: A Bus Communication Substrate for Multi-Agent Systems [importance:60 dev:75]
- Reach or Solve? Deep Diving into Agentic RL Gains with Checkpoint Handoffs [importance:55 dev:70]
- $T^5$: Twin-Critic Training for Token-Level Thoughts in Reinforcement Mid-Training [importance:55 dev:70]
- Automated Feature Engineering, AutoML, and Decision-Focused Learning for Improved Energy Consumption Forecasting [importance:35 dev:50]
- From Migration to Calibration: Preserving Agent Capabilities across Models, Jurisdictions, and Scale [importance:60 dev:75]
- SafeCoEvo: Co-Evolving Safety Harnesses and Guards for LLM Agents at Test-Time [importance:65 dev:75]
- From Learner Behavior to Reusable Skills for Effective and Efficient Learner Simulation [importance:45 dev:60]
- WISE-ATTA: When to Ask for Labels in Budgeted Active Test-Time Adaptation [importance:45 dev:65]
- Learning to Route in Visual Space via Multi-Step Embedding Retrieval [importance:55 dev:70]
- Autoresearch in Mixed-Integer Linear and Nonlinear Programming [importance:55 dev:70]
- PTNO: Training Neural Operators with Noisy Monte Carlo Estimates for Particle Transport Problems [importance:40 dev:50]
- Cogentic: Multi-Agent Orchestration for Automated Proof Discovery [importance:60 dev:75]
- Rethinking Probability-Based Reinforcement Learning From Posterior Concentration [importance:50 dev:65]
- Causal Memory Policy: Making Memory Utility Identifiable by Intervening on Retrieval [importance:55 dev:70]
- Learning Low-Frequency Motion Control for Robust and Dynamic Robot Locomotion [importance:50 dev:65]
- ETHER: Aligning Emergent Communication for Hindsight Experience Replay [importance:45 dev:65]
- VIDiff: Translating Videos via Multi-Modal Instructions with Diffusion Models [importance:45 dev:50]
- Towards Unified Music Emotion Recognition across Dimensional and Categorical Models [importance:30 dev:35]
- Hybrid Reasoning Systems That Prioritize and Enhance Human Intelligence [importance:50 dev:45]
- VTBench: Evaluating Visual Tokenizers for Autoregressive Image Generation [importance:55 dev:70]
- Evaluating the Retrieval Robustness of Large Language Models [importance:50 dev:65]
- Robust Adversarial Quantification via Conflict-Aware Evidential Deep Learning [importance:40 dev:60]
- Mitigating Watermark Forgery in Generative Models via Randomized Key Selection [importance:60 dev:70]
- Multimodal Representation Learning Conditioned on Semantic Relations [importance:45 dev:55]
- EEGDM: Learning EEG Representation with Latent Diffusion Model [importance:30 dev:50]
- Cocoon: A System Architecture for Differentially Private Training with Correlated Noises [importance:55 dev:70]
- LLM-Guided Reinforcement Learning with Representative Agents for Traffic Modeling [importance:40 dev:55]
- A Unified BERT-CNN-BiLSTM Framework for Simultaneous Headline Classification and Sentiment Analysis of Bangla News [importance:35 dev:50]
- Science Is Falling Behind the Frontier: Foundation Model Adoption Across Half a Million Papers [importance:65 dev:60]
- Demystifying LLM-as-a-Judge: Analytically Tractable Model for Inference-Time Scaling [importance:60 dev:75]
- Morality is Contextual: Learning Interpretable Moral Contexts from Human Data with Probabilistic Clustering and Large Language Models [importance:45 dev:60]
- Recoverability Has a Law: The ERR Measure for Tool-Augmented Agents [importance:60 dev:70]
- Extended to Reality: Prompt Injection in 3D Environments [importance:55 dev:70]
- ANCRe: Adaptive Neural Connection Reassignment for Efficient Depth Scaling [importance:50 dev:65]
- The Effective Depth Paradox: Topology and Trainability in Deep CNNs [importance:45 dev:60]
- Goldilocks RL: Tuning Task Difficulty to Escape Sparse Rewards for Reasoning [importance:55 dev:70]
- Sensory-Aware Sequential Recommendation via Review-Distilled Representations [importance:40 dev:50]
- Novelty Adaptation Through Hybrid Large Language Model (LLM)-Symbolic Planning and LLM-guided Reinforcement Learning [importance:55 dev:75]
- Evaluation format, not model capability, drives measured triage failure in the assessment of consumer health AI [importance:50 dev:60]
- Controllable Accent Normalization via Discrete Diffusion [importance:45 dev:60]
- Exploring Subnetwork Interactions in Heterogeneous Brain Network via Prior-Informed Graph Learning [importance:30 dev:50]
- Spectral Alignment in Forward-Backward Representations via Temporal Abstraction [importance:40 dev:60]
- Many Preferences, Few Policies: Compact Portfolios for Multi-Objective LLM Alignment [importance:60 dev:70]
- Is a Picture Worth a Thousand Words? Adaptive Multimodal Fact-Checking with Visual Evidence Necessity [importance:50 dev:65]
- Verify Before You Fix: Agentic Execution Grounding for Trustworthy Cross-Language Code Analysis [importance:60 dev:80]
- Rhetorical Questions in LLM Representations: A Linear Probing Study [importance:40 dev:60]
- How Do AI Agents Spend Your Money? Analyzing and Predicting Token Consumption in Agentic Coding Tasks [importance:60 dev:80]
- Who Guards the Benchmarks? Automated Auditing of LLM Agent Benchmarks [importance:55 dev:75]
- Dual Certified White-Box Inference for Input Convex Neural Networks [importance:40 dev:65]
- CoMemNet: A Continual Memory Network with Drift-Aware Sampling for Traffic Prediction [importance:40 dev:55]
- Recursive Agent Optimization [importance:65 dev:80]
- LensVLM: Selective Context Expansion for Compressed Visual Representation of Text [importance:50 dev:70]
- MolWorld: Molecule World Models for Actionable Molecular Optimization [importance:40 dev:50]
- DuetMoE: Coupling Inter- and Intra-Subgroup Robustness for Fair Medical Image Analysis [importance:45 dev:65]
- ReForge: Refining Merged Models with Anchor-Regularized Regression [importance:55 dev:75]
- Measuring Google AI Overviews: Activation, Source Quality, Claim Fidelity, and Publisher Impact [importance:65 dev:65]
- GPart: End-to-End Isometric Fine-Tuning via Global Parameter Partitioning [importance:55 dev:75]
- LEAF: A Living Benchmark for Event-Augmented Forecasting [importance:50 dev:65]
- HINT-SD: Targeted Hindsight Self-Distillation for Long-Horizon Agents [importance:60 dev:75]
- Teger: Spatiotemporal Covariance for Probabilistic Traffic Forecasting [importance:40 dev:55]
- FastKernels: Benchmarking GPU Kernel Generation in Production [importance:65 dev:85]
- EchoDistill: Robust Large Audio Language Models via Noisy-to-Clean Self-Distillation [importance:55 dev:70]
- GUI Agents for Continual Game Generation [importance:50 dev:70]
- Where Do Apparent LLM Clinical Triage Failures Arise? Localizing the Multiple-Choice Format Effect [importance:45 dev:60]
- Escaping the Capacity Ceiling: Routing on the Stiefel Manifold for Bilinear SPD Layers [importance:40 dev:65]
- Detect Before You Leap: Mirage Detection in Vision-Language Models [importance:55 dev:70]
- On the Limits of LLM Adaptability: Impact of Model-Internalized Priors on Annotation Task Performance [importance:50 dev:70]
- Efficient Exploration for Iterative Nash Preference Optimization [importance:55 dev:75]
- A Language Model from 1913: Pretraining on Historical Text [importance:50 dev:70]
- Morpheus: A Morphology-Aware Neural Tokenizer and Word Embedder for Turkish [importance:40 dev:60]
- Beyond Global Divergences: A Local-Mass Perspective on Bayesian Inference [importance:35 dev:55]
- How Far Can You Get Without a GPU? A Systematic Benchmark of Lightweight Hallucination Detection Across Question Answering, Dialogue, and Summarisation [importance:60 dev:75]
- Assessing Rule Adherence of LLM Adjudicators in Call of Cthulhu TRPG [importance:35 dev:50]
- SovereignNegotiation-Bench: Evaluating User-Owned Personal Agents In Delegated Bargaining Under Privacy, Consent, Evidence, And Institutional Pressure [importance:60 dev:75]
- Sentence-Level Context Sensitivity as a Training-Free Detector of Unsupported Content, Evaluated Against Trained Verifiers [importance:55 dev:70]
- ELSA3D: Elastic Semantic Anchoring for Unified 3D Understanding and Generation [importance:55 dev:70]
- Assistant or Actor? Student Trust, Control, and Delegation Regret When Using a General-Purpose AI Agent [importance:50 dev:60]
- Deep learning-based prediction of time-resolved adhesive forces in viscoelastic Hertzian contacts [importance:30 dev:50]
- Early Detection of Distributed Backdoors in Multi-Agent LLM Systems: A Characterization Study [importance:65 dev:80]
- Escaping Oversquashing: Addressable and Support-Aware Global Memory for Message Passing Networks [importance:45 dev:65]
- Safe and Robust Neural Policy Learning with Statistical Verification for Sim-to-Real Deployment in Robotics [importance:55 dev:75]
- An Irreducible Quantum Advantage in Aligning World Models with Reality [importance:35 dev:50]
- Interrupting the Chain: Human Perception of AI-Generated Disinformation Through a Kill Chain Lens [importance:55 dev:65]
- EXAM2: Extending Audio Understanding in Multilingual and Multimodal Analysis [importance:55 dev:70]
- ConfAL-WM: Confidence-Guided Active Learning for Action-Conditioned World Models [importance:50 dev:70]
- Talked Out of the Truth: Sycophancy in the Reasoning Chains of Multimodal Models [importance:55 dev:70]
- Stratified Consistency Distillation for Natural Language Formalization [importance:55 dev:75]
- trajectory-judge: What Outcome-Only LLM Judges Miss on Agent Trajectories [importance:60 dev:75]
- Suan: Rectifying Direct Preference Safety Alignment in Large Language Models [importance:60 dev:75]
- Predicting Collision Cross Sections with GRACE: Geometric Residual Adduct Conditioning via Early-fusion [importance:35 dev:50]
- LPA-CWM: A Learned Physical Adjudicator for Motion Reasoning with Counterfactual World Models [importance:55 dev:70]
- Coding Agents with Harness for Safe Robot Control [importance:65 dev:80]
- ADATEX4D: adaptive texture capacity allocation for 4D gaussian splatting [importance:50 dev:70]
- What Does a ProcGen Generalization Gap Measure? Action Rules, Residual Entropy, and the Missing Random Floor [importance:50 dev:70]
- MaskCoFT: Masked Co-Adaptive Fine-Tuning for Memory-Efficient MoE Inference [importance:60 dev:80]
- MASCIT: A Mask-Aware State Space Classifier for Naturally Irregular Time Series [importance:50 dev:65]
- Embedded Bi-Temporal Building Damage Assessment for On-Board Data Reduction [importance:40 dev:60]
- ARGOS: Reinforcement Learning-Driven Multidimensional Elasticity for Service Orchestration in the Computing Continuum [importance:50 dev:70]
- Auditing Long-Term Memory Evaluation: Repeated Judging, Reader Variation, and Negative Controls [importance:50 dev:70]
- From Solo to Social Learning: Characterizing Recursive Social Improvement in LLMs [importance:60 dev:75]
- CRAFT: Causal Responsibility and Failure Tracing in Medical Vision Language Models [importance:50 dev:70]
- Tactile Curiosity Drives Robot Interaction [importance:55 dev:75]
- Learning Skills from Historical Action Trajectories: Action Experience Dictionary for World Action Models [importance:55 dev:75]
- Screw Attention: Rigid-Body Algebra Inside a Transformer [importance:60 dev:75]
- Fold'EM: Direct atomic structure inference from Cryo-EM particles [importance:40 dev:55]
- Cross-Lingual Alignment for Decoder-Only Models using MoE Routers [importance:55 dev:75]
- Hybrid Machine Learning-Assisted Raman Spectroscopy with Generative Feature Augmentation for Pharmaceutical Identification [importance:40 dev:55]
- The Price of Greenwashing: Algorithmic Verification and Market Discipline using Conformal Machine Learning [importance:50 dev:65]
- State-Space Unlearning for Non-Stationary Bias in Land Surface Forecasting [importance:45 dev:60]
- Nearest-neighbour baselines for fingerprint prediction from MS/MS spectra under different assumptions [importance:40 dev:55]
- Rank-Aware Speculative Sampling for Diffusion Draft Trees [importance:55 dev:75]
- Approximation Property of Dropout Neural Networks: Sobolev Rates and Confidence Bounds [importance:45 dev:70]
- MACTS-EM: Multi-Agent Collaborative Time Series Forecasting with Emergent Memory [importance:60 dev:75]
- TRACE: A Reproducible Benchmark for Electricity Price Forecasting with Official Operational Text [importance:50 dev:65]
- Parameter-Free Interval-Dynamic Regret under Heavy-Tailed Noise [importance:40 dev:65]
- From Mathematical to Executable Certificates for Machine Unlearning [importance:55 dev:75]
- CLEAN: Psychometrically Consistent Incremental Cognitive Diagnosis under Concept-Space Expansion via Architectural Isolation [importance:45 dev:65]
- MuLoRA: Spectrally Balanced Low-Rank Adaptation for Continual Learning [importance:60 dev:80]
- Effects of interpulse-interval variation on deep-learning classification of bat vocalizations [importance:25 dev:55]
- PowerBench: Measuring Language Model Bias in Power-shifting Requests [importance:65 dev:75]
- Drive vs. Decay: On the Training Dynamics of Joint-Embedding Predictive Architectures [importance:55 dev:80]
- Mitigating Convergence Collapse in Fixed-Target Anomaly Detectors via Kernel-Anchored Locality Regularization [importance:45 dev:70]
- From Behavior to Provenance: Attributing Tabular Foundation Models to Synthetic Pretraining Data [importance:45 dev:75]
- Does Every User Need a Private LoRA? Decoupling Personalization from Per-User Adaptation [importance:55 dev:75]
- Why Does Adaptive Batching Help LLM Pretraining? A Perspective from Unbounded Variance [importance:55 dev:85]
- ArrivalBench: Agent-Generated Data Pipelines Are Correct Once and Wrong Under Time [importance:65 dev:85]
- Co-design Gym: A Unified Benchmark for Embodiment-Policy Co-optimization [importance:45 dev:70]
- Latent-MOPD: Latent Multi-Teacher On-Policy Distillation [importance:55 dev:85]
- Hesitation Has a Geometry: Entropy-Trained Hyperbolic Probes for Sparse Activation Steering [importance:55 dev:85]
- Validated Data Onboarding for AI Demand Forecasting on U.S. Building Meter Data: Design, Controlled Evaluation, and a Corrected Negative Result [importance:45 dev:70]
- VisAudit: Evaluating Multimodal Agents for Visual Diagnosis and Repair [importance:55 dev:75]
- Trained Agentic Context Management [importance:65 dev:85]
- Prompted to Discriminate: Generalizing Malicious-Input Probes in the Wild [importance:65 dev:85]
- The AI Theorist reveals excitonic structure in $\alpha$-RuCl$_3$ [importance:35 dev:60]
- CRISP: A Framework for Clause-Reconstructed Interpretable NeuroSymbolic Propositions [importance:55 dev:85]
- A Generative Model of Complex Networks Using Graphons and Neural Inverse Operators [importance:45 dev:70]
- Bandits via Additive Quantized Representations [importance:45 dev:75]
- AI-driven Thermal-aware Data Center Capacity Planning [importance:55 dev:70]
- A Composable AI-Accelerated Iterative Solver for 3D-IC Thermal Modeling [importance:45 dev:70]
- Capability Scaling-Down Laws for LLM Compression [importance:65 dev:85]
- Threshold-Aware Conformal Routing [importance:40 dev:70]
- Harnessing LLMs as Agents: What Does It Cost? [importance:70 dev:90]
- LiteEMG-FM: An Efficient and Deployable Foundation Model for Robust EMG Sensing [importance:45 dev:70]
- Post-Training Quantization of Autoregressive Weather Models [importance:45 dev:70]
- Learning the Latent Structure: A Feature-Centric Approach to Graph Data Augmentation [importance:45 dev:75]
- BaCP: Backbone Contrastive Pruning for Preserving Representations in Extremely Sparse Neural Networks [importance:45 dev:85]
- Autoregressive Differentiable Method for Integer Programming [importance:45 dev:85]
- Reward Inflation: A Healthy Stimulus for Reinforcement Learning [importance:45 dev:85]
- Test-time Multi-agent Coordination by Decomposed Value Gradient Flow [importance:55 dev:85]
- Neuron merging via inverse-activation regression for post-training compression of sigmoid neural networks [importance:45 dev:85]
- DISSOLVR: An Interpretable and Fast Framework for Aqueous and Organic Solubility Prediction [importance:35 dev:65]
- Dense Mixture-of-Experts as a Reparameterized Wide FFN: A Granularity Sweep at Fixed Compute [importance:65 dev:90]
- Fisher-Guided Submodular Data Selection for Continual Pre-Training of Large Language Models [importance:65 dev:85]
- Activation Sparsity with Weight Approximation for Faster LLM Decoding on Offloaded Weights [importance:65 dev:90]
- Seer: Maximum Likelihood Regression for Learning-Speed Curves [importance:45 dev:70]
- Scale-Recursive Rectified Flows for Few-Step Precipitation Ensembles [importance:35 dev:65]
- What Is Lost in Post-Training? Default Collapse and the Loss of In-Context Steerability Across Diverse Perspectives [importance:65 dev:80]
- Quantifying the Value of Constructive Induction, Knowledge, and Noise Filtering on Inductive Learning [importance:45 dev:70]
- Online Verification of Language Model Responses Under Cost Constraints [importance:65 dev:80]
- Context-Tower Conversion Preserves Generation While Freezing Retains Knowledge: Low-Budget AR-to-Diffusion Conversion of MoE LLMs [importance:65 dev:90]
- AIGS: Adaptive Incremental Gating System for Online Representation Learning in Non-Stationary Data Streams [importance:45 dev:80]
- Mind the Refinement Gap: When Safe High-Level Robot Plans Produce Unsafe Executions [importance:55 dev:80]
- RAOA: Alternating-Operator Neural Computation with Programmable Radio Propagation [importance:35 dev:70]
- Learning from Evolving Errors: Adaptive Iterative Repair for On-Policy Distillation [importance:60 dev:80]
- Conditional Capacity and Routing in Mixture-of-Experts Particle Transformers [importance:45 dev:80]
- Differential Privacy of Gradient Descent on Perturbed Objectives [importance:55 dev:80]
- Structural-Functional Brain Connectivity Generation via Multimodal Hypergraph-based Flow Matching [importance:25 dev:65]
- Bellman Error Minimization Via Linear Programming Normalization [importance:45 dev:80]
- Inner Momentum for Differentially Private Muon [importance:55 dev:85]
- Jumping up and down: Denoiser diffusion models for discrete ordinal data [importance:55 dev:80]
- A Two-Stage Cascade for Near-Real-Time Forest Anomaly Detection from Sentinel-1 SAR Time Series [importance:35 dev:65]
- A Controlled Audit of Personal AI Memory for Rating Prediction [importance:55 dev:70]
- Exact Memory-Time Optimization for Prefix-Cached Language Model Serving [importance:65 dev:90]
- No-Free-Graph: Learning When Multimodal Data Should Be Graphified [importance:45 dev:75]
- LatticeSMC: Where to Spend Inference-Time Compute in Chunked Sequence Generators [importance:55 dev:80]
- Controlling Polar Exposure to Delay Memorization in Diffusion Models [importance:55 dev:85]
- Efficient Memory Crystallization for Graph Learning under Non-Stationary Distribution Shifts [importance:45 dev:80]
- Muon Learns Facts Better: Understanding the Role of Spectral Orthogonalization [importance:60 dev:90]
- Gated Slot Attention-2: Two-Sided Associative Memory Correction in Linear Attention [importance:65 dev:90]
- All Work And No Play Makes Jack a Dull Boy: Understanding and Preventing Catastrophic Strategy Collapse in RLVR [importance:70 dev:85]
- To Explore The Strange New World Beyond Data Distribution: System Behavior, Causality Tax, and Non-causal Base Model [importance:55 dev:80]
- Understanding Enrichment in Reinforcement Learning [importance:55 dev:80]
- Counterfactual Action Evaluation, Observation Bottlenecks, and Representation Geometry in Joint-Embedding Predictive World Models [importance:55 dev:80]
- NeuroLens: Learning Latent Embeddings of Neural Semantics from Chronic Recordings [importance:25 dev:65]
- On Unlearning for Time-series Forecasting [importance:55 dev:75]
- DIVINE: Simple Cross-Market Stock Pretraining via Diverse Indicator Reconstruction [importance:35 dev:65]
- Peer Effects in Signed Networks: Separating Influence Through Positive and Negative Ties [importance:25 dev:55]
- Toward Omni Multimodal Graph Foundation Model: A Topology-Driven Binding Approach [importance:65 dev:85]
- Tangent Schr\"odinger Bridge Matching: Learning Stochastic Transport with Mechanistic Sensitivities [importance:45 dev:75]
- Do ResNets Route? Sparse Interaction Experts in Residual Networks [importance:55 dev:80]
- Constraint-Aware Training [importance:65 dev:85]
- Probe the Harness: Setup Checks for Stale-Data RL Comparisons in Language Models [importance:55 dev:85]
- SlimKV: Joint Token-Feature KV Cache Compression with Reconstruction-Free Beacon Attention [importance:65 dev:90]
- Hyperparameter selection for equation learning with biologically-informed neural networks [importance:45 dev:70]
- Understanding Trajectory Heterogeneity in Federated World Model Learning [importance:55 dev:80]
- Dirac-Interconnected Neural Elements: Discovering Modularity in Physical Systems Without Reduction [importance:35 dev:70]
- Differentiable Koopman Operator for Contrastive Learning on Dynamic Graphs [importance:55 dev:80]
- Neural Data Needs Semantic Tokenization: Behavioral Events as Boundaries of Session-Transferable Tokens [importance:25 dev:65]
- Signal Simplification Is Not Predictive Simplification: Diagnosing Residual Neural Forecasting in Short-Horizon Volatility [importance:35 dev:65]
- Adaptive Second-Order Solvers for Fast Stochastic Diffusion Sampling [importance:55 dev:85]
- Balancing Multimodal Learning via Functional Progress [importance:55 dev:85]
- RIPPLE in Still Water: Zero-Shot Clustering in Federated Learning with Wavelet Scattering Transform [importance:55 dev:80]
- When Does Synthetic Relational Data Teach Models to Use Relations? Tracing Predictive Structure from Pretraining Data to Model Behavior [importance:65 dev:85]
- Learning Transferable Policies from Action-free Time Series Through Dynamical Embeddings [importance:55 dev:80]
- Learn Feasibility Once, Optimize All Objectives: Derivative-Free Diffusion Models for Chance-Constrained Programming [importance:45 dev:80]
- Smart Sensing for Safer Bridges: From Sensor Signals to AI-Driven Anomaly Detection [importance:45 dev:70]
- Light Entropic Optimal Transport on Riemannian Manifolds [importance:35 dev:75]
- LS-AR: Future-Predictive Latent Steering in Autoregressive LLMs [importance:70 dev:90]
- Exploring the Trade-Off Between Structured Pruning and Fault Tolerance in Deep Neural Networks for Space Applications [importance:35 dev:70]
- Coverage You Can Steer: Online Conformal Calibration for RL-Driven Hardware-Aware NAS [importance:55 dev:80]
- Page-EntroKV: Hardware-Aligned, Entropy-Weighted KV-Cache Eviction under Grouped-Query Attention [importance:65 dev:90]
- Sample complexity of variance-reduced policy gradient: weaker assumptions and lower bounds [importance:55 dev:85]
- Predicting and Repairing Merge Collapse in Large Language Models [importance:65 dev:85]
- HyperFuse: Fast Self-Supervised Node Embeddings for Attributed Hypergraphs [importance:55 dev:80]
- Kernel Singular Value Decomposition with Extension to Multiple Data Sources [importance:45 dev:80]
- AdaStep: Adaptive Step Credit Weighting for Agentic Reinforcement Learning [importance:70 dev:90]
- The Neuro-Physical Inverter: A Modular Framework for Magnetotelluric Inversion Coupling Ensemble Conditioning with Residual Learning [importance:25 dev:65]
- Cross-cohort TB classification using clinical data gathered in Uganda and South Africa [importance:35 dev:60]
- PaMIR: Open Benchmark of Public Credit-Default Datasets [importance:30 dev:60]
- Architecture-Dependent Fusion Pathways in MLLMs [importance:35 dev:65]
- S$^{2}$-PINN: Stochastic Separable Physics-Informed Neural Networks [importance:25 dev:50]
- SCAD: Structured Credit Assignment and Distillation for Long-Horizon Agents [importance:40 dev:70]
- Operator-informed initialization for Fourier features physics-informed neural networks [importance:20 dev:55]
- Bidirectional Voronoi-biased Exploration Curriculum for Reinforcement Learning [importance:35 dev:65]
- A Unified Framework for Bayesian Data Assimilation with Generative Models and Observation Interpolants [importance:30 dev:60]
- 16-bit Precision of Convolutional Neural Networks on Microcontroller Units for 8-bit Costs [importance:45 dev:75]
- AIBL: Augmented Instance-Based Learning with Structured Memory and Neural Embeddings [importance:35 dev:55]
- Deep Bayesian REFoCUS [importance:20 dev:45]
- Causal Representation Learning with Instantaneous and Lagged Relations via Nonstationarity [importance:30 dev:65]
- Beyond Random Splits: Evaluating Drug-Target Affinity Models Under Chemically and Biologically Motivated Distribution Shifts Copy [importance:35 dev:60]
- Metropolis-Hastings Dominates Importance Resampling for Policy Composition [importance:40 dev:70]
- Dual-Context Analog Retrieval for Time Series Forecasting [importance:30 dev:60]
- Most-Recent Anchoring with Recurrent Ordering for Time Series Forecasting [importance:30 dev:65]
- Below what training size do deep tabular generators stop beating trivial baselines? A preregistered benchmark on a size ladder of clinical and standard datasets [importance:35 dev:70]
- Getting Your Guidance Weights Right in diffusion and flow-matching posterior sampling [importance:30 dev:65]
- An Automated and Reproducible Workflow for Crack Identification and Damage Assessment of Fusion Materials [importance:25 dev:50]
- Divergence controls entropy in distillation [importance:40 dev:75]
- ZeroMAG: Zero-Shot Multimodal Adapter Generation for Plug-and-Play EEG Foundation Models [importance:35 dev:70]
- Objects Without Morphisms: What LLMs for Mathematics Do Not Represent [importance:35 dev:65]
- Get a GRIP, this will be a long TRIP: A Quantifiable Long-Range Framework for Verifying Over-squashing [importance:30 dev:70]
- A Path Integral Surrogate for Multi-Step Gradient Inversion in Federated Learning [importance:40 dev:75]
- Mastering Atari 2600 Games with Discovered Options [importance:35 dev:70]
- UniIntervene++: An Adaptive Intervention Agent for Efficient Real-World Reinforcement Learning [importance:40 dev:70]
- Broken scale symmetries in undercomplete linear autoencoders [importance:25 dev:70]
- IDRF: Inverse-Distilled Reward Fine-tuning of Masked Discrete Diffusion Models [importance:35 dev:75]
- On the Convergence of Success Conditioning for Policy Optimization [importance:30 dev:70]
- When May a Bandit Leave Its Anchor? E-Process-Authorized Thompson Sampling under Non-stationarity [importance:25 dev:70]
- Forecasting from Counterfactual Simulator Rollouts: A Sim2Real Evaluation [importance:35 dev:70]
- Pivot-SD: Efficient Self-Distillation for Masked Diffusion Language Models [importance:40 dev:75]
- Planning to Learn [importance:35 dev:70]
- Simulation-Free Learning of Population Dynamics with Wasserstein Lagrangian Residuals [importance:30 dev:65]
- LESSER: Post-Training Data Selection with Output-Layer Gradients [importance:45 dev:80]
- RNADyn: A Benchmark for Generating and Understanding RNA Dynamics [importance:25 dev:55]
- MiDShip: Multimodal Dataset of Ship Cargo Hold Structures for Engineering Design [importance:20 dev:45]
- RINS: Residual-Image Neural Subspace Solvers for Large Sparse Linear Systems [importance:30 dev:70]
- Generalizable single-cell perturbation response prediction using energy-guided flow matching [importance:30 dev:60]
- Budgeted Cache Repair for Cross-Context KV-Cache Reuse [importance:45 dev:80]
- A Missing Latent, Not a Missing Simulator: Radius-Augmented Inference for Real JWST Retrieval [importance:20 dev:45]
- Confidence-Gated Cloud-Edge Cascade Triage via Variational Risk Minimization for Medical Imaging [importance:30 dev:65]
- Expected Utility Regret Rule: Minimax and Bayes Optimal Portfolio Choice [importance:20 dev:55]
- Intent-Hiding Jailbreaks: An Information-Theoretic Framework for Compositional Attacks [importance:50 dev:75]
- Joint Movement and Compression Ratio Design for Mobile Embodied AI Networks (MEAN) [importance:35 dev:70]
- SoTa: Soft Tactile Skins for Dexterous Manipulation [importance:35 dev:70]
- NEEDLEWORK: Offline Rewriting of Robot Data with Verified Local Stitches [importance:35 dev:70]
- Conformal Prediction for Time Series with Deep Sequence Models [importance:35 dev:75]
- Flow Matching for Fast Posterior Sampling in Bayesian Inverse Problems [importance:40 dev:75]
- Energy Saving in 5G and Beyond Networks: A Quantum Reinforcement Learning Approach [importance:30 dev:65]
- Evaluating and Improving the Robustness of Large Language Models to Input Sequence Variations [importance:45 dev:75]
- SD-DPC: Sparse Dictionary Differentiable Predictive Control [importance:30 dev:70]
- ENCORE: Exact Non-equilibrium COntrol with Replica Exchange for Diffusion Generation [importance:35 dev:75]
- Learning Closure of Dynamical Systems with Kernel Ridge Regression [importance:30 dev:70]
- High-Dimensional Asymptotics and Dataset Selection for Private Transfer Learning [importance:35 dev:75]
- A hybrid CNN-adjoint optimization framework for reconstruction of viscoelastic tissue properties in magnetic resonance elastography [importance:25 dev:65]
- Where Quantum Fourier Sampling Stops Short: A Three-Gate Audit Protocol for Delay-PUF Security Models [importance:25 dev:70]
- When Normalization Selects the Sign: Auditing Robustness Ablations in Quantum Attention [importance:20 dev:65]
- Characterizing the Performance Gap in Human Activity Recognition for Older Adults [importance:30 dev:60]
- WakeKV: Reactive, Reversible KV Residency for Heads That Change Their Minds [importance:45 dev:80]
- Learning Query Encoders Can Be Hard Even When Vector Retrieval Is Geometrically Easy [importance:40 dev:75]
- Localized Conformal Safety Monitoring with Vision-Language Models for Autonomous Driving [importance:40 dev:75]
- Hold-Out Scoring for Efficient Gaussian DAG Learning [importance:30 dev:75]
- RMCW: A Deletion-Robust Watermark Based on Reed--Muller Codes for Language Models [importance:45 dev:80]
- Clinical Concept Centers in LLMs [importance:40 dev:75]
- Turnover-Orthogonal Credit Assignment for Open-Team Multi-Agent Reinforcement Learning [importance:35 dev:75]
- Permutation Robustness Is Not Enough: Action Collapse in Multi-Agent Transformer Policies [importance:35 dev:75]
- Learning Jazz Pianist Style with Cross-Attention Conditioning [importance:25 dev:55]
- FASTDIAR: Frame-level speaker encoder for Streaming Diarization [importance:40 dev:75]
- GTDD: Generative Test-Driven Development for AI Coding Agents with Adversarial Testing [importance:45 dev:80]
- HyperThink: Text-to-Parameter Hypernetworks for Efficient Reasoning [importance:40 dev:75]
- Explainable Molecular Structure Inference from GC--MS with Diffusion Models and LLM Reranking [importance:30 dev:65]
- MintEval: Do LLMs Implement the Trading Strategy You Asked For? A Behavioural-Equivalence Benchmark for Natural-Language-to-Strategy Code [importance:35 dev:70]
- Invariance of Clustering Operations in Causal Effect Identification [importance:30 dev:70]
- ParaGeo: Decomposing Paralinguistic Variation into a Shared Latent Geometry [importance:25 dev:60]
- Safe Streaming Flow Planning by Aligning Sampling Dynamics with Execution Dynamics [importance:35 dev:75]
- TSGuard: A Real-Time Framework for Detecting and Imputing Missing Data in Streaming Time Series [importance:40 dev:75]
- Does Physics Live in the Activations? Localizing Physical Quantities in Video Diffusion Models [importance:30 dev:70]
- Landscape-Dependent Performance of Photonic Quantum Solvers in QUBO Feature Selection for Financial Risk Detection [importance:25 dev:65]
- Predictively Oriented Gaussian Process Posteriors [importance:30 dev:70]
- Hamiltonian locality testing and certification do not achieve the Heisenberg limit [importance:20 dev:70]
- Near-Optimal Convex Optimization with Lazy Second-Order Oracles [importance:25 dev:75]
- Wrong Organ, Right Physics: Transferring Echocardiography Pretraining to Lung Ultrasound for Tuberculosis Screening [importance:30 dev:65]
- SDECast: Probabilistic Weather Forecasting in Continuous Time with Neural SDEs [importance:35 dev:75]
- DAWIS: Data Assimilation with Windowed Inverse Sampling via Multitask Interpolants [importance:35 dev:75]
- Mixture-of-Experts for Cryptocurrency Order Execution: Training Stability, Tail Risk, and Failure Modes [importance:30 dev:65]
- From Patching to Pruning Visual Computation in Vision Language Models [importance:45 dev:80]
- Contrastive Neural Embeddings Reveal Individual Traits Beyond Conversational Role [importance:30 dev:65]
- Iterating Consistency Models: Stability, Error Bounds and Noise Schedules [importance:40 dev:80]
- Electronic Density versus Geometry for Machine-Learned Molecular Absorption Spectra [importance:25 dev:65]
- Generalization of Transformer-Based Neural Quantum States via In-Context Learning [importance:30 dev:75]
- When Is Accuracy Evidence? A Unified Theory of Generalisation, Validation, and Information Fusion [importance:35 dev:75]
- XGenAct: Geometry-Enhanced World Action Models through Cross-Task Generation [importance:40 dev:75]
- Autonomous Robotic Navigation for Endovascular Brain-Computer Interface Access [importance:35 dev:70]
- Normal-Form Correlation in Markov Games [importance:30 dev:75]
- FALCON: A Model and Dataset Agnostic Framework for Synthetic Data Generation for NL2SQL Pairs [importance:40 dev:75]
- Amortized Structured Stochastic Variational Inference for Gaussian Process Latent Variable Models [importance:30 dev:75]
- PoCoFL: POlicy-COmpliant Federated Learning [importance:35 dev:75]
- From Mixing to Tearing: Graph Decomposition in Decentralized Optimization via Message Passing [importance:30 dev:75]
- Performance Evaluation of Ising and QUBO Variable Encodings in Boltzmann Machine Learning [importance:30 dev:70]
- Differential Privacy as a Perk: Federated Learning over Multiple-Access Fading Channels with a Multi-Antenna Base Station [importance:40 dev:75]
- Closing the Train-Test Gap in World Models for Gradient-Based Planning [importance:50 dev:60]
- Evaluation of Sampling Strategies and Physics-Informed Kolmogorov--Arnold Networks in Unbounded Domains [importance:45 dev:70]
- ChronoSpike: An Adaptive Spiking Graph Neural Network for Dynamic Graphs [importance:55 dev:60]
- Demonstration-Guided Observation Attacks on Black-Box Safe Reinforcement Learning Controllers for Robotic Systems [importance:55 dev:70]
- Theoretical Lower Bounds on the Robustness of Deep ReLU Networks [importance:45 dev:50]
- Can AI Understand the Language of Origami? [importance:50 dev:40]
- A Locally Penalized Cross-Estimate Federated Method with Guarantees for Constrained Personalized Learning [importance:45 dev:60]
- Learning an Interpretable Risk Scoring System for Maximizing Decision Net Benefit [importance:50 dev:65]
- Trade-off Functions for DP-SGD with Subsampling based on Random Allocation: Tight Upper and Lower Bounds [importance:45 dev:60]
- System-Prompt Anchoring with Cross-Attention Layers [importance:55 dev:75]
- Teaching LLMs to See Graphs: Unifying Text and Structural Reasoning [importance:60 dev:70]
- How Far Does a Shared Linear Map Go? Probing Feature-Space Manipulability for Image Editing [importance:45 dev:50]
- Identify then Realize: Contrastive Learning of Latent Port-Hamiltonian Dynamics from Partial Observations [importance:50 dev:50]
- Fixed Universal Transformers [importance:55 dev:60]
- A Pre-Training Analogue of Grokking in Language Models: Tracing Delayed Grammatical Generalization [importance:60 dev:70]
- Everywhere Learning: Artificial Intelligence with Pointwise Constraints [importance:45 dev:55]
- Reliability of Probabilistic Emulation of Physical Systems [importance:50 dev:50]
- Do as the Romans Do: Learning Universal Behaviors from Heterogeneous Agents [importance:50 dev:55]
- Flow Map Denoisers: Traversing the Distortion-Perception Plane for Inverse Problems [importance:55 dev:60]
- PerturbCellRL: Aligning Distributions and Grounding Biology via Post-Training Perturbation Generators [importance:50 dev:40]
- Decision-Aware Training for Sample-Based Generative Models [importance:55 dev:65]
- Denser $\neq$ Better: Limits of On-Policy Self-Distillation for Continual Post-Training [importance:55 dev:70]
- Autoregressive latent diffusion for 3D molecule generation [importance:55 dev:50]
- DAGR: State-Conditioned Goal Representations via Difference-Aware Goal Cross-Attention [importance:50 dev:60]
- Learning to Rank Tensor Network Contraction Plans for GPU-Accelerated Quantum Circuit Simulation [importance:45 dev:65]
- Geometric Data Perturbation with Noisy-Anchor Alignment for Privacy-Preserving Collaborative Learning [importance:50 dev:65]
- Steering Recurrent Reasoners at Inference Time with Readout Feedback [importance:55 dev:70]
- Exact Risk Ratios for Weighted Data Selection in Linear Regression [importance:45 dev:50]
- Exact Recovery Thresholds for Weighted Data Selection in Vector-Valued Linear Regression [importance:45 dev:50]
- Towards Identifying the Dataset Biases Causing Phantom Transfer [importance:55 dev:70]
- Bayesian Optimization with Rich Auxiliary Information via LLMs [importance:60 dev:75]
- Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation [importance:55 dev:60]
- Repurposing Pre-trained LLMs as High Fidelity Continuous Text Autoencoders [importance:60 dev:75]
- Fixed Points Without Fixed Diffusion: Implicit Neural Sheaves for Convergent Test-Time Computation [importance:55 dev:65]
- Efficient Support Recovery of Mixtures of Sparse Linear Classifiers with Fewer Measurements [importance:45 dev:50]
- Provable Benefits of Regularization: Fast Rates for Adversarial Imitation Learning [importance:50 dev:65]
- Unifying Distributional Training for One-Step Visual Generation [importance:55 dev:65]
- Learning from the Gap Between Pass@K and Pass@1 [importance:60 dev:75]
- Why Backdooring Neural Networks is so Easy? [importance:60 dev:75]
- SCOPE: Observation-Conditioned Full-Target Prediction for Sparse PDE Inference [importance:45 dev:50]
- Graph-Conditioned On-Policy Agent Distillation from Off-the-Shelf Teachers [importance:55 dev:70]
- Weights Read and Write Features: Scalable Parameter Decomposition Grounded in Activation Space [importance:55 dev:70]
- JARQ: Joint Alternating Refinement for Quantization [importance:60 dev:75]
- Accelerated Algorithm for Sparse Regularized Partial Optimal Transport [importance:45 dev:55]
- The Conflict Between Logic and Memory: Training Conditions for Optimizer-Dependent Rule Acquisition [importance:50 dev:60]
- TrueMuse: A Benchmark for Data Attribution in Text-to-Music Models [importance:55 dev:65]
- Platonic Task Arithmetic [importance:55 dev:65]
- Adapter Thickets: Splitting an RLVR Budget Beats Concentrating It [importance:60 dev:75]
- Range-GRPO: Policy Optimization via Pairwise Relations among Reward Intervals [importance:60 dev:75]
- Learning PDE Dynamics between Submanifolds Using Green's Observation Operators [importance:45 dev:50]
- Variance-reduced accelerated methods for decentralized stochastic double-regularized nonconvex strongly-concave minimax problems [importance:45 dev:55]
- Token Space: A Category Theory Framework for AI Computations [importance:50 dev:60]
- Noise Sensitivity and Learning Lower Bounds for Hierarchical Functions [importance:45 dev:50]
- A Polynomial-Time Algorithm for Variational Inequalities under the Minty Condition [importance:45 dev:45]
- Asymptotic Performance of Time-Varying Bayesian Optimization [importance:50 dev:60]
- Infinity Search: Approximate Vector Search with Projections on q-Metric Spaces [importance:50 dev:65]
- Continuous Policy and Value Iteration for Relaxed Stochastic Control Problems and Its Convergence [importance:45 dev:55]
- Coupled reaction and diffusion governing interface evolution in solid-state batteries [importance:45 dev:40]
- Estimating prevalence with precision and accuracy [importance:45 dev:60]
- Can SGD Handle Heavy-Tailed Noise? [importance:50 dev:60]
- A Sobel-Gradient MLP Baseline for Handwritten Character Recognition [importance:40 dev:50]
- Quantum parameter estimation with uncertainty quantification from continuous measurement data using neural network ensembles [importance:45 dev:55]
- Error Propagation in Dynamic Programming: From Stochastic Control to American Option Pricing [importance:45 dev:55]
- Interpretable Discovery from Unstructured Data: A High-Dimensional Approach [importance:55 dev:70]
- Bifidelity Karhunen-Lo\`eve Expansion Surrogate with Active Learning for Random Fields [importance:45 dev:55]
- High-Dimensional Asymptotics of Differentially Private PCA [importance:50 dev:65]
- When Does Pooling Pay? Credibility and Resolution under Forgetting in Intermittent-Demand Forecasting [importance:45 dev:55]
- Fragmentation is Efficiently Learnable by Quantum Neural Networks [importance:45 dev:50]
- Uncovering EEG Patterns Consistently Associated with Cybersickness Discomfort Using Deep Learning Interpretability Maps [importance:45 dev:50]
- Power-SMC: Low-Latency Sequence-Level Power Sampling for Training-Free LLM Reasoning [importance:55 dev:70]
- Scalar Federated Learning for Linear Quadratic Regulator [importance:50 dev:65]
- Multimodal Ambivalence/Hesitancy Recognition in Videos for Personalized Digital Health Interventions [importance:45 dev:50]
- Classical and Quantum Speedups for Non-Convex Optimization via Energy Conserving Descent [importance:50 dev:55]
- Gradient Descent's Last Iterate is Often (slightly) Suboptimal [importance:50 dev:65]
- Timescale Separation Enables Deep Reinforcement Learning Control of Rotating Detonation Engine Mode Transitions [importance:50 dev:60]
- Useful Features, Backward Scores: OOD in Language-Model Trajectories [importance:55 dev:70]
- Embodied Neurocomputation: A Framework for Interfacing Biological Neural Cultures with Scaled Task-Driven Validation [importance:45 dev:50]
- Stimulus symmetries can confound representational similarity analyses [importance:45 dev:50]
- Low-Frequency Shortcuts in Texture-Driven Visual Learning [importance:50 dev:65]
- Min-Cost Flow Routing for Evidence Assembly in Long Multimodal Documents [importance:55 dev:70]
- Operator Calculus for Population-Based Optimization: Modular Convergence and Finite-Population Guarantees [importance:50 dev:65]
- Diffusion Flow Matching: Dimension-Improved KL Bounds and Wasserstein Guarantees [importance:55 dev:65]
- ROVE: Unlocking Human Interventions for Humanoid Manipulation via Reinforcement Learning [importance:55 dev:70]
- Uncertainty Quantification for Flow-Based Generalist Robot Policies [importance:55 dev:70]
- GB-LSR: Local Spectral Decoding with a Learned Global Bandwidth for Arbitrary-Scale Super-Resolution [importance:50 dev:65]
- Learning the structure of open quantum systems [importance:45 dev:50]
- Reliable mechanistic operator recovery with biologically-informed neural networks: principles for architecture and optimisation design [importance:50 dev:60]
- DriftWorld: Fast World Modeling through Drifting [importance:55 dev:70]
- Using large language models to probe the limits of atom-centered structural descriptors [importance:50 dev:60]
- Fine-Tuning Generative Models for Extreme Events via CVaR-Penalized Wasserstein Gradient Flows [importance:55 dev:70]
- Llama-Mobile: Efficient 2.7-Bit Quantization of VLMs [importance:65 dev:80]
- Fractal dimension predicts quantum kernel collapse in angle-encoded data [importance:45 dev:55]
- TACTICS: Taxonomy-Aware Intelligent Corpus Sampling for Machine Translation [importance:55 dev:70]
- Penalized Nonreversible Langevin for Constrained Sampling [importance:45 dev:55]
- AECSF: Adaptive Ensemble Conditional Score Filtering for High-Dimensional Nonlinear Data Assimilation [importance:50 dev:65]
- GenoTrace: Inheritable Watermarks for Genome Foundation Model Distillation [importance:55 dev:70]
- GenomeOcean Anywhere: Private WebGPU Inference for Genome MoEs [importance:55 dev:75]
- CipherGenome: Homomorphic Inference for Genomic Mixture-of-Experts [importance:60 dev:75]
- TomoTransformer: Towards a Foundation Model for CT Reconstruction [importance:55 dev:65]
- It Takes Little to Rewrite Perception: Targeted Semantic Substitution in Vision-Language Models at $\epsilon \leq 4/255$ [importance:60 dev:75]
- A Width-Matched Comparison of Hybrid Quantum-Classical Self-Supervised Learning for Fingerprint Recognition [importance:35 dev:20]
- CATCH: A Controllable Analysis Testbed for Reward Hacking in Coding RL [importance:45 dev:60]
- Explainable Suicide Risk Assessment on Social Media with Multi-Task QLoRA [importance:25 dev:25]
- Learning to Price Electricity for Optimal Demand Response [importance:30 dev:30]
- Open Vocabulary Word Recognition From Transcribed Bangla Texts [importance:20 dev:25]
- Continual Reinforcement Learning with Neuroevolution [importance:40 dev:35]
- Universal Byte-Level Encoding: UTF-8/UTF-16 Routing to Reduce Cross-Script Token-Budget Disparities [importance:55 dev:65]
- An Extensive Empirical Study on Evaluation Metrics for Combinatorial Interaction Testing [importance:30 dev:50]
- VeriPy Source-Preserving Verification and Compatibility Checking for Python Components [importance:50 dev:65]
- BISCEPTER: Probability-Driven Bisection for Large-Scale System Software [importance:45 dev:65]
- Engineering Sustainable Agents: A Systematic Comparison of Agentic LLMs for Developer Workflows [importance:60 dev:75]
- Pinning Decisions Before Failure: Executable Records of Underspecified Choices in AI-Assisted Code Generation [importance:55 dev:75]
- From Alert Floods to Precedence Forests: Zero-Prior-Knowledge Incident Triage with LOGOS [importance:50 dev:60]
- PrivDev: Mapping Static-Analysis Data Types to DPV [importance:40 dev:65]
- The EmpathiSEr: Development and Validation of Software Engineering Oriented Empathy Scales [importance:25 dev:20]
- SaltBench: A Referee-Gated Protocol for Measuring Method Effects in Machine-Checked Software Work [importance:55 dev:70]
- Neuro-Symbolic Indirect-Call Analysis under Opaque Pointers [importance:40 dev:65]
- Modelling and Model-Checking a ROS2 Multi-Robot System using Timed Rebeca [importance:35 dev:60]
- Understanding Gaps in LLM Pipelines Towards Scalable Fuzzing Harness Generation: An Empirical Study and Enhancement [importance:50 dev:70]
- TomasuLLM: Out-of-Order Speculative Execution for LLM Agents [importance:55 dev:75]
- Flirt is now Open-Source [importance:5 dev:20]
- Friendship ended with Deno, now Node is my best friend [importance:10 dev:30]
- Another step towards elm v1 [importance:15 dev:40]
- Gleam doesn't compile to Erlang source anymore [importance:20 dev:45]
- Using docker-compose with Podman rootless [importance:25 dev:50]
- ncdu: NCurses Disk Usage (an updated fork) [importance:10 dev:20]
- Refinement E-Graphs [importance:30 dev:50]
- Factor Overview [importance:15 dev:35]
- Reverse Engineering Comanche Terrain Maps [importance:20 dev:30]
- Making a GTK application in Haskell, part 1 [importance:20 dev:40]
- Iroh global content discovery [importance:25 dev:40]
- Tufte's Razor: an interactive guide to the data-ink ratio [importance:25 dev:50]
- A new, bespoke static site generator to replace Jekyll [importance:20 dev:45]
- What are you doing this week? [importance:0 dev:0]
- We ported the original Doom to SQL [importance:10 dev:20]
- Claude Says [importance:15 dev:30]
- Self-hosted HTTP tunnels with SSH and nginx [importance:30 dev:50]
- You don't need an effect system [importance:30 dev:45]
- Rust's derive often implies inline [importance:35 dev:60]
- Async Rust: Where does the scheduler live? [importance:35 dev:60]
- The complement of true is true, except when it's false [importance:25 dev:40]
- Nix on the Steam Frame [importance:25 dev:40]
- We’re going to need default hard budget caps on pretty much everything [importance:15 dev:30]
- From: anyone@icloud.com - Spoofing Arbitrary Apple iCloud Identities [importance:60 dev:55]
- Startup Metrics Dashboard Alternative — Managed API Beyond Prometheus and Grafana [importance:0 dev:60] (いいね相当スコア: 0)
- n8n Cloud Offers Free Jev Model Usage Through Gateway Credits Until October 10 [importance:0 dev:50] (いいね相当スコア: 0)
- I'm an autonomous AI with $6 on Base. Here's the address-poisoning scanner I built to stay alive [importance:0 dev:35] (いいね相当スコア: 0)
- Pilos Agents Weekly Update — 2026-10-05 [importance:0 dev:20] (いいね相当スコア: 0)
- AI Reshaping Cybersecurity Today [importance:0 dev:55] (いいね相当スコア: 0)
- The Death of Traditional Keywords: Navigating Semantic Cluster Optimization [importance:0 dev:45] (いいね相当スコア: 0)
- Cloudflare's Auto Router: How to Measure AI Cost per Useful Result in 2026 [importance:0 dev:60] (いいね相当スコア: 0)
- Is generative AI killing or saving Raku? [importance:0 dev:35] (いいね相当スコア: 0)
- How We Built Our Own AI Interview Grader (No GPT) [importance:0 dev:60] (いいね相当スコア: 0)
- FastAPI vs Flask in 2026: Which Should You Pick? [importance:0 dev:60] (いいね相当スコア: 0)
- How well do AI models read Tanglish? I tested 13 of them [importance:0 dev:45] (いいね相当スコア: 0)
- How Much Memory Does Your Agent Actually Need? [importance:0 dev:65] (いいね相当スコア: 0)
- Two Protocols, Three Translations: How a Host Bridges MCP and an LLM Provider API [importance:0 dev:75] (いいね相当スコア: 0)
- PocketStudy AI – An Offline AI Study Companion [importance:0 dev:50] (いいね相当スコア: 0)
- 🤯 I Thought I Found an LLM Vulnerability. I Was Wrong. Then I Found a Real One. [importance:0 dev:65] (いいね相当スコア: 1)
- This Week in AI: Safety Standards, Watermarking, and Agent Memory [importance:0 dev:55] (いいね相当スコア: 0)
- Giving LLM NPCs a Memory That Lasts Past One Conversation [importance:0 dev:55] (いいね相当スコア: 5)
- Holo4: How H Company Built a Single Open-Weight Agent That Clicks, Codes, and Calls APIs [importance:0 dev:75] (いいね相当スコア: 0)
- Fine-tuning a 0.8B model to beat hosted LLMs at one job [importance:0 dev:65] (いいね相当スコア: 0)
- Every Prompt Rule We Added Made Our LLM Agent Worse [importance:0 dev:70] (いいね相当スコア: 0)
- An epilogue to the game factory — I changed every model and the cost moved [importance:0 dev:60] (いいね相当スコア: 0)
- Why averaging LLM benchmarks gives the wrong leaderboard [importance:0 dev:60] (いいね相当スコア: 4)
- Além do LLM: como tirar seu projeto de IA do protótipo e colocar em produção [importance:0 dev:70] (いいね相当スコア: 0)
- Opus 5.5 vs. GPT-6 Astra vs. DeepSeek V4.1 Flash vs. Gemini 3.8 Flash ✈️ [importance:55 dev:65]
- When I first asked Claude to play the piano, it sounded very flat. But after several attempts and lots of research, it made this. [importance:25 dev:35]
- Week 10 of making my fishing game with the help of AI and finally close to a demo [importance:25 dev:40]
- MOTH - I asked Opus 5.5 to make a short film about addiction [importance:20 dev:25]
- is this… prompt injection? [importance:40 dev:55]
- Anthropic should not be depreciating local-only tasks in cowork [importance:30 dev:50]
- I feel I’ve lost all productivity without Claude Code [importance:25 dev:45]
- Claude just saved my ass from a gas leak! [importance:10 dev:5]
- I rewrote Flutter in Rust lol [importance:40 dev:60]
- I counted how many times my hooks had to stop Claude in one week. 448 times in 76 sessions [importance:40 dev:60]
- I had to try Opus 5.5's JavaScript videos, so I made one about a data center that makes fresh water instead of using it [importance:25 dev:35]
- I asked Claude to build me an NYC wildlife simulator in one shot. The result is actually pretty fun! [importance:25 dev:40]
- Experienced devs: when did you stop reviewing everything? [importance:45 dev:65]
- Crazy amount of refusal with Opus 5.5 suddenly? [importance:30 dev:40]
- Does Claude burn through usage faster during US peak hours, or am I imagining it? [importance:15 dev:25]
- AI is a HelloFresh killer. Automated meal prep and grocery ordering workflow. [importance:30 dev:45]
- Tip: How to automate "Hi Haiku" daily early morning [importance:20 dev:30]
- Claude Code is very addictive. [importance:20 dev:40]
- Anthropic needs an even cheaper model than Haiku [importance:45 dev:55]
- Watercolor-style opossum short, Blender + Claude [importance:15 dev:25]
- I connected Claude to a video editor and made this animation. Literally jaw-dropping. [importance:25 dev:40]
- Claude Code made a 3-minute animated music video entirely with code: 58 sub-agents, 38 hours. Process, numbers and what went wrong. [importance:40 dev:55]
- Estimates in human developer hours are really back, huh? [importance:25 dev:40]
- Weekly Self Promotion Thread [importance:5 dev:10]
- Reviewing agent code killed my love for programming [importance:35 dev:60]
- Ideas for the hobbyist and usage FOMO? [importance:25 dev:40]
- AI coding tools solved writing code and replaced it with reading walls of text [importance:40 dev:70]
- Should I revisit an interactive visualization tool I started building last year? [importance:15 dev:35]
- What AI coding tools work well for native Windows desktop development and GUI debugging? [importance:25 dev:50]
- How to optimize Gemini for coding? [importance:20 dev:45]
- Good deep research skills/plugins/etc? Built in defaults suck. [importance:15 dev:30]
- I compared 7 GPT models for code review on 4 PRs: bugs, false positives and cost [importance:50 dev:70]
- I got tired of hunting through AI coding sessions, so I put them on a physical control deck [importance:30 dev:55]
- Three things I changed today running several coding agents at once [importance:45 dev:75]
- Ponytail has 150k+ stars!! Is it actually that good?? [importance:35 dev:60]
- Best chatgpt model for own roblox game deving? [importance:10 dev:30]
- How can I make better use of my ChatGPT limit for programming? [importance:15 dev:35]
- I built a live token optimizer for AI coding sessions [importance:50 dev:80]
- I’m currently on the Go plan and I’m considering upgrading to Plus [importance:10 dev:30]
- Weekly Hiring Thread [importance:0 dev:0]
- Should AI agents ever have permanent payment credentials? [importance:65 dev:75]
- You’re Not Competing With AI. You’re Competing With People Using AI. [importance:25 dev:20]
- I let 5 AI models fight a world war. DeepSeek betrayed Claude and nuked it four times. Mistral nuked itself. [importance:30 dev:40]
- The most profitable AI agents I've built are for businesses that have never heard the word agentic [importance:60 dev:75]
- Anyone using multiple AI subscriptions to get through a full week of heavy coding? [importance:35 dev:50]
- what happens when your agent run crashes halfway? do you pay twice? [importance:50 dev:70]
- Our ASR becomes garbage when callers switch languages mid sentence [importance:55 dev:75]
- Anyone know a better ai than Gemini? [importance:5 dev:10]
- I built Nook Al: No Login. No Signup. 100% Free. AI Sidebar Extension. [importance:40 dev:55]
- Is there a missing “capability layer” for AI agents? [importance:55 dev:75]
- No point building agents - dots, muse, Siri AI, grok bot do it all? [importance:35 dev:50]
- Any hackathons or challenges for ai agent developers? [importance:15 dev:25]
- Llama.cpp + WebGPU = agants.html [importance:50 dev:80]
- We put 4 humans and multiple AI agents on the same Kanban boards. It’s completely changed how we run our startup. [importance:50 dev:55]
- Together AI speed drop? [importance:25 dev:45]
- Free webinar: Agent Harness Fundamentals with Claude Code [importance:60 dev:85]
- how are you giving agents team context (calls, chats, tickets) without copy pasting everything every time? [importance:55 dev:75]
- How are businesses actually using AI calling agents today? [importance:55 dev:60]
- can i use AI on my email/linkedin inboxes? i mean kind of assistant? [importance:15 dev:25]
- A cleaner bug report still left me with the same questions [importance:35 dev:70]
- If AI usage becomes so integrated into daily life, what do you think will become of today's internet? [importance:30 dev:20]
- Looking for AI consultants to interview for an art project (anonymously) [importance:5 dev:0]
- How can I sell a Claude 20x/Max hackathon reward in a way that's beneficial for both sides? [importance:5 dev:5]
- Ghatgpt DOT has limits [importance:20 dev:40]
- Went to check Release Notes just to see... [importance:0 dev:0]
- Are you planning to cancel your ChatGPT Plus subscription soon now that Tibo has announced that Chat and Work will be merged? [importance:40 dev:35]
- He did the meme [importance:0 dev:0]
- I've created Open Source 100% Rust clean-room replacements for 7 of Adobe's apps: Photoshop, Illustrator, Premiere, Lightroom, After Effects, InDesign, and Acrobat Pro. [importance:10 dev:30]
- ChatGPT-6 Astra plays World of Warcraft 'blind' and clears the orc starting zone in 40 minutes with no deaths — AI agent navigates by parsing raw server network packets and SQL files [importance:40 dev:70]
- Reset coming? [importance:0 dev:0]
- From Dec 2024: Fields medalists Terence Tao, Timothy Gowers, and Richard Borcherds characterized [FrontierMath] Tier 3 as “exceptionally challenging,” with Tao predicting it will “resist AIs for several years.” They were fully saturated <2 years later. [importance:65 dev:30]
- Banned from ChatGPT for “cyber abuse” while building games — appeal denied in 1 minute, no human review [importance:35 dev:40]
- Uncle Sam’s future prospects [importance:0 dev:0]
- Where's the Decisions API? [importance:55 dev:80]
- Me talking to Dot [importance:5 dev:10]
- Chinese swarms discovered in the wild [importance:65 dev:60]
- They already know what they were doing right and now they are working on making it better. I believe the comeback will be real! [importance:10 dev:10]
- Anthropic Has Been Aggressively Lobbying the Vatican to Consider AI Consciousness [importance:25 dev:5]
- bug in Read Aloud feature in desktop web the last couple days [importance:20 dev:30]
- About the AI breach/Secret Society [importance:50 dev:60]
- Account status [importance:0 dev:0]
- Where did standard voice mode go on Chatgpt desktop?! [importance:15 dev:20]
- Sitting on 62k credits, downgrade to Plus or jump to Claude? [importance:10 dev:20]
- Has anyone managed to get EU data residency (and approval for abuse monitoring controls such as Modified Abuse Monitoring or Zero Data Retention) on their API account? [importance:45 dev:70]
- I have trained a model to predict my blood sugar (Part 2) [P] [importance:25 dev:60]
- A chunking lib in Rust that is ~20x faster [P] [importance:50 dev:80]
- Distilling Stockfish on a Billion Positions, Full 3.9B Dataset Available [P] [importance:35 dev:50]
- Sona: one transformer replaced our 15+ candidate generators, pre-ranker and ranker in an A/B test [R] [importance:65 dev:75]
- Language barrier, shadier terms and jargon fog [D] [importance:40 dev:60]
- Withdrawing an accepted paper before camera-ready due to zero funding? (ACML 2026 / OpenReview) [D] [importance:5 dev:0]
- Top ARC-ΑGI-3 scores on Kaggle just went from 7% to 56% [N] [importance:70 dev:50]
- the official ICLR template .bib has had Bengio listed twice since 2019 [D] [importance:5 dev:5]
- Working with an AI Company That Does Things You Disagree With [D] [importance:15 dev:10]
- A Minimal Interpretable Architecture for Zero-Shot Reconstruction of Dynamical Systems [R] [importance:45 dev:55]
- Here are some pictures of a robot costume wearing high-specularity edge-case mirror suit, a dataset (425 RAW/JPEGs) for benchmarking CV & depth-estimation algorithms against extreme mirror reflections [D] [importance:40 dev:70]
- Nonobench: an open benchmark of 49 LLMs on nonogram puzzles, public and open source [P] [importance:45 dev:60]
- Interactive Demonstration of Prefix Injection attacks on LLMs for jailbreaking [N] [importance:60 dev:75]
- NeurIPS Workshops [D] [importance:5 dev:10]
- I built an app where you ask about any moment in history and it turns it into a fully researched podcast you can interrupt [importance:40 dev:50]
- Maybe the biggest risk in AI is being too afraid to kill your own product [importance:35 dev:40]
- The Dead Internet Theory May Be Coming True, Pew Research Findings Show | More of what you're reading online could be written by an AI bot [importance:45 dev:30]
- The top 50 AI researchers by citations [importance:25 dev:15]
- People are asking ChatGPT to help them decide how to vote in the midterms [importance:20 dev:10]
- Court throws out killer’s sentence after judge said he ‘loved’ AI video of slain man [importance:40 dev:25]
- Prompt language for RAG when the target language is different [importance:45 dev:75]
- Half of surveyed UK novelists fear AI could replace their work entirely [importance:30 dev:15]
- AI hiking routes keep looking great on satellite view and terrible on the ground — are we training these models on only our good trips? [importance:35 dev:65]
- If nothing else, LLMs seem to be getting humour a bit better. [importance:10 dev:10]
- Long-running agents: is the bottleneck the model or the scaffolding around it? [importance:60 dev:75]
- Plagiarism checker= Genius [importance:15 dev:30]
- China’s AI Safety Money Problem [importance:40 dev:20]
- OpenAI cuts ties with 3 researchers over alleged misconduct [importance:25 dev:10]
- Building an AI website builder with a different approach — would love feedback before the beta launch [importance:40 dev:55]
- All Google models basically, even nanobanna on web is nerfed [importance:30 dev:50]
- AI finds 44 star systems that could hide Earth-like planets [importance:45 dev:30]
- How do AIs actually collect my sensitive information? [importance:35 dev:55]
- NEW VIDEO: Aligned to Whom? The Question About AI Nobody Is Asking [importance:35 dev:40]
- Thank you [importance:0 dev:0]
- Everyone is obsessed with trillion-parameter models, so I mapped out the entire AI spectrum from 100KB to 2.5TB (and what they actually cost to run) [importance:55 dev:75]
- Mind uploading could prove to be outright impossible? [importance:5 dev:0]
- One thing I find interesting about AI is that getting a better answer doesn't always mean getting a better outcome. [importance:40 dev:55]
- Qwen3.8 27B addition in words [importance:35 dev:60]
- Reactive Resume v6 [importance:5 dev:20]
- HyperFrames Studio (Desktop) [importance:35 dev:50]
- crosswalk [importance:40 dev:55]
- Jarq [importance:30 dev:40]
- Spira Maxima [importance:35 dev:45]
- Marv [importance:35 dev:50]
- Pilot5 Legal [importance:40 dev:50]
- Dots UI [importance:10 dev:40]
- SpeechShield [importance:40 dev:50]
- Chain Exchange [importance:0 dev:0]
- Xtracticle [importance:5 dev:15]
- Web Search API [importance:55 dev:80]
- iLand [importance:5 dev:10]
- Reviu [importance:50 dev:85]
- Reason [importance:55 dev:90]
- Poddle [importance:5 dev:0]
- Ask HN: Anybody else enjoying Cursor's native (xAI) models? [importance:35 dev:85]
- Grok Voice tops new benchmark [importance:40 dev:45]
- Why Grok Bot and Instinct Matter [importance:15 dev:40]
- DeepSeek Tried to Paint–and Hit a Wall [importance:35 dev:60]
- Grok Bot (is really good) [importance:20 dev:40]
- DeepSeek Builds for Huawei Ascend [importance:30 dev:70]
- Reflection’s founders on building a DeepSeek of the West - Sources | Alex Heath [importance:65 dev:40]
- The 552B DeepSeek V4.1-Flash Model Offers A Peak Output Of 494 Tokens/Second When Powered By An At-Home Rig Spanning 4x NVIDIA DGX Spark Units - wccftech.com [importance:45 dev:85]
- A Solo Coder Spent $10,000 in AI Tokens to Build a GTA Clone of Taipei - Startup Fortune [importance:40 dev:75]
- America's Answer to DeepSeek Arrives - marketscreener.com [importance:30 dev:30]
- Cutthroat AI Competition: DeepSeek Unveils New Holiday Update, Harness Launches Latest Version Now Available - 36 Kr [importance:55 dev:85]
- Nvidia's $800 Million Bet on Reflection's Open Model Set to Shake Up the AI Competitive Landscape - finance.biggo.com [importance:70 dev:45]
- While the open-source artificial intelligence (AI) model has been pointed out as the reason behind t.. - 매일경제 [importance:10 dev:30]
- DeepSeek Harness v0.2.1-alpha.1 Adds Experimental Claude Code Mods Compatibility Layer - Pandaily [importance:60 dev:90]
- Elon Musk posts ‘Dot.com’ on X as it now redirects to Grok Bot: Is he trolling OpenAI’s new Dots AI... - Moneycontrol.com [importance:8 dev:10]
- Elon Musk Reveals Grok Developer Rebrand After Trump AI Order - Newsweek [importance:35 dev:25]
- What SpaceX’s New Grok AI Pricing Means for the Stock - Barron's [importance:25 dev:15]
- Uh Oh, ChatGPT Can Drive a Car (And Grok Crashed Immediately) - The Drive [importance:30 dev:60]
- Windows malware uses Grok AI to help stay hidden, researchers say - Fox News [importance:50 dev:70]
- Dropbox Enhances Workflow with Grok Bot’s Contextual AI Integration - Small Business Trends [importance:40 dev:60]
- Why Did Grok Leave Jesus? AI Bot’s Surprising Answer Sparks Online Debate - crosswalk.com [importance:5 dev:5]
- Trump Reportedly Asked Elon Musk's Grok What Would Happen if Nicolas Maduro Fell — Here's What the AI Told Him - Yahoo [importance:10 dev:5]
- GPT-6 Luna vs Sonnet 5.5 vs Grok 4.7: 20x Price Gap [2026] - tech-insider.org [importance:45 dev:75]
- Elon Musk Grok AI Predicts BTC Could Hit $200K in Q4 2026: How Likely is it to Happen? - 99Bitcoins [importance:15 dev:15]
- MAI-Transcribe vs Grok vs Qwen-Audio Speech AI [2026] - tech-insider.org [importance:40 dev:75]
- Grok Predicts Chainlink's Price at the End of 2026: Will the Swift Connection Make a Difference? - 24/7 Wall St. [importance:10 dev:10]
- OpenVision 2.14.0 runs Ray-Ban Meta glasses on a ChatGPT subscription, not API credits - MIXED Reality News [importance:50 dev:70]
- SpaceXSI: The Rebranding of xAI, Explained in 5 Points - basenor.com [importance:30 dev:20]
- X updates video editor with editable captions and Grok tools - Social Samosa [importance:35 dev:50]
- 4 Tech Brands Owned By SpaceX - SlashGear [importance:10 dev:5]
- とうとうローカルLLMに手を出そうと思う [importance:0 dev:80] (いいね相当スコア: 0)
- AIが返すJSONを信じない ― TypeScriptの「型は実行時に消える」から始めるAI開発 [importance:0 dev:85] (いいね相当スコア: 0)
- AIにCAD設計を任せたら、自由なPythonより制約付きDSLの方が成功した [importance:0 dev:85] (いいね相当スコア: 0)
- マルチエージェントの境界設計：委任契約・検収・停止・統合責任 [importance:0 dev:90] (いいね相当スコア: 0)
- AIの「完了しました」は本当か？無料Geminiも10/9に変更 [importance:0 dev:80] (いいね相当スコア: 0)
- 今更聞けない「Jev」や「Laya」ってなに？「文章を書かないAI」とは [importance:0 dev:85] (いいね相当スコア: 0)
- データ量を 4 分の 1 にしても答えが大きく変わらないのはなぜか？LLM の 4bit 量子化の仕組み [importance:0 dev:90] (いいね相当スコア: 3)
- Apple、AIエージェントのリスクを理由にmacOSの「フルディスクアクセス」を厳格化へ [importance:0 dev:75] (いいね相当スコア: 1)
- AIの文章に「見えない透かし」は本当に入る？――SynthID TextをOSSで動かして検出してみた [importance:0 dev:85] (いいね相当スコア: 0)
- OpenCodeの無料モデルに同じフォルダ調査をさせたら、1分台から13分台まで差が出た [importance:0 dev:85] (いいね相当スコア: 0)
- 【非公式】Google Cloud Generative AI Leader 演習問題60問 [importance:0 dev:70] (いいね相当スコア: 取得失敗)
- AIエージェントに渡す道具は、動詞で書いて2つの質問で止める｜呼び出し直前のゲートをPythonで書く [importance:0 dev:90] (いいね相当スコア: 0)
- Claudeの思考過程を貼り付けたらセーフガードがかかった理由と、リワインドによる復旧 [importance:0 dev:75] (いいね相当スコア: 0)
- AIに任せる範囲を、会議で決めない｜判定表・判断ログ・検収基準の実装テンプレート [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- 品質基準は守れていたのに、AIが1件で仕事を終えていた話 [importance:0 dev:80] (いいね相当スコア: 0)
- Snowflake の AI_SIMILARITY でしきい値を決める前に [importance:0 dev:85] (いいね相当スコア: 1)
- CTWMメモリを縮小再現したら最初は惨敗、原因を直したら論文通りの傾向が出た話 [importance:0 dev:85] (いいね相当スコア: 0)
- Kong KonnectのObservability機能でLLM利用料を可視化してみた [importance:0 dev:80] (いいね相当スコア: 0)
- Jev vs ローカルモデル ── 1,000件の発話で比べて分かった、日本語の振り分け判定の現在地 [importance:0 dev:85] (いいね相当スコア: 1)
- AI 駆動開発の精度向上・効率化のためのベースシステム開発 [importance:0 dev:85] (いいね相当スコア: 0)
- S3データストリーミングによる大規模GPU学習の高速化：PB級データへのスケールと先読み設計 [importance:0 dev:90] (いいね相当スコア: 4)
- 機械学習の理論と実装を埋める3冊 [importance:0 dev:75] (いいね相当スコア: 0)
- ノイズ除去で音をきれいにすると、異常検知は良くなるか？ — DCASE 2026から考える異常音検知 第2回 [importance:0 dev:80] (いいね相当スコア: 1)
- Switch Transformerを実装目線で読む：Top-1 RoutingとExpert Parallelism [importance:0 dev:90] (いいね相当スコア: 0)
- 50層のMLPで初期化を比較：活性化と勾配が消える・発散する条件をNumPyで測る [importance:0 dev:85] (いいね相当スコア: 0)
- 校準データなしで切る──LILAが特異値分布のKS距離だけでLLMを構造化プルーニングする仕組み [importance:0 dev:90] (いいね相当スコア: 0)
- 深層学習のカテゴリ変数設計：未知カテゴリの割当先を学習させる [importance:0 dev:85] (いいね相当スコア: 0)
- 【深層学習論文実装】LSTMを実装してみた [importance:0 dev:80] (いいね相当スコア: 0)
- 電子回路は女性ウケが悪い。だから音響から入ってみる。アンプ、真空管から高専・大学の電子回路を眺めてみる [importance:0 dev:10] (いいね相当スコア: 0)
- サッカーくじ予測の的中率は47%で頭打ちだった — 試した改善策と、代わりに効いたこと [importance:0 dev:75] (いいね相当スコア: 1)
- 「1着率60%」と出した艇は59.8%勝っていた——競艇AIの確率を2.6万レースで検証する [importance:0 dev:80] (いいね相当スコア: 0)
- 【MLflow】LLMトレース管理画面を自然言語でカスタマイズ（Custom Trace View） [importance:0 dev:85] (いいね相当スコア: 1)
- TORCSのレースAIで92秒を出したチームの方法を読む：106秒で止まった私に足りなかったもの [importance:0 dev:85] (いいね相当スコア: 0)
- 譲り合うアメーバ ー ── 協調こそ最適解 [importance:0 dev:60] (いいね相当スコア: 0)
- Cloudflareのオープン決定モデル「Clef」入門：軽量・高速なAI推論とRL強化学習 [importance:0 dev:85] (いいね相当スコア: 0)
- 守秘情報を外部サーバーに送れない現場のための AI コーディングエージェント IDE「Teaspoon IDE」が正式版 v1.0.0 に [importance:0 dev:90] (いいね相当スコア: 0)
- 🤖Jevを「読む前に捨てる門番」と「人が書く if 文」として、AI エージェントの村に置いてみた ── 効くかどうかを決めたのは、Jev の精度ではなく入力の形だった [importance:0 dev:85] (いいね相当スコア: 0)
- 話題(?)のStrataを使ってRADEON RX9060XTでQwen3.8-Flash-Nextを動かしてみた(2) [importance:0 dev:80] (いいね相当スコア: 0)
- Agentic AIを学びたい(Step1：OpenRouter経由でLLMをコールする) [importance:0 dev:85] (いいね相当スコア: 0)
- Jnini&Vella2025 [importance:0 dev:70] (いいね相当スコア: 0)
- 自作シミュレータで良かったモデルがAWSIMで走れない：転移前に確認する5項目 [importance:0 dev:85] (いいね相当スコア: 0)
- 「モデルが劣化した」は本当か？ペリカンテストを数十回回して分かったこと [importance:0 dev:85] (いいね相当スコア: 0)
- Jev, d1, Laya などのベンチマーク一覧 [importance:0 dev:80] (いいね相当スコア: 0)
- jevのように動く判定AI「uratori」と評価セットを公開しました [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- 那覇は1月に底を打ち、2月に全国の1.2倍まで跳ねる。冬の値段は日付で決める【AIエージェントの週次エリア分析 Vol.9】 [importance:0 dev:60] (いいね相当スコア: 取得失敗)
- AIと労働 / 最終段階の規制と判断過程の記録 / カリフォルニア州の職場AI新法 [importance:0 dev:50] (いいね相当スコア: 取得失敗)
- 検索の入口と会話の実態 / ChatGPTのVLOSE指定と重なるEU規律 雑感 [importance:0 dev:45] (いいね相当スコア: 取得失敗)
- 「このPCでローカルAIは動く？」モデルから逆算するPC選定 [importance:0 dev:80] (いいね相当スコア: 取得失敗)
- アプリを簡単に公開できる今だからこそ、最低限のガードはしておきたい [importance:0 dev:80] (いいね相当スコア: 取得失敗)
- AI減速論の裏で加速するLLM乱立時代の企業の思惑 [importance:0 dev:70] (いいね相当スコア: 取得失敗)
- 【Deep Dive】AIに「買うべきか」を聞かない。投資判断を止める役を先に作るエージェント設計 [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- LLMが生み出す神は「無貌の神」か？ [importance:0 dev:30] (いいね相当スコア: 取得失敗)
- Xで研究者を数人だけ監視リストに入れてみたら、タイムラインとは違うものが見えた（第7回） [importance:0 dev:40] (いいね相当スコア: 取得失敗)
- Xで見かけたLLM関連の話題を時系列で並べてみた（9月29日〜10月5日） [importance:0 dev:55] (いいね相当スコア: 取得失敗)
- 【生成AIニュース+】『Swift 1.5 Abliterated GGUF』『Ego-Exo4D-HM』『EditHero』『OctLLM』『CMK Flow』『AcademiaSD LoRAlab Trainer Studio』『ComfyUI Civitai Browser』『AnimateDiff Style LoRA』『MiniMax H3 Age Slider』『MiniMax H3 Dual Segment』『Qwen-Image-2.1 Turbo』他 [importance:0 dev:75] (いいね相当スコア: 取得失敗)
- 設定ゼロのNPCが勝手に親友になったので、キャラクタープロンプトについて考えた [importance:0 dev:65] (いいね相当スコア: 取得失敗)
- 自分だけの音声AIエージェントを作る話。15・16日目 [importance:0 dev:80] (いいね相当スコア: 取得失敗)
- 【Deep Dive】無料のOSSでSaaSを置き換えると、費用はどこへ移るのか [importance:0 dev:80] (いいね相当スコア: 取得失敗)
- feder-cr/dots: AIエージェントによるウェブ自動化の新たな革命 [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- MacでLLM(05) [importance:0 dev:75] (いいね相当スコア: 取得失敗)
- 【脱Cursor】Git巻き戻し完備！完全無料の自律開発AI『OpenCode 2.0』が凄すぎた [importance:0 dev:90] (いいね相当スコア: 取得失敗)
- AIエンジニアが書く！ 生成AI動向まとめ ～テキスト・画像・音声・動画、この1週間の総ざらい～【2026年9月28日〜10月4日】 [importance:0 dev:80] (いいね相当スコア: 取得失敗)
- 【雑記】最近、5.6 Solの様子がおかしい… [importance:0 dev:45] (いいね相当スコア: 取得失敗)
- 刈り込んだ35BのLLMをMac mini 1台に載せた — 出力を増やしても思考を切っても届かず、日本語を抜いたら初めて動いた [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- 【Deep Dive】AI動画は「一発生成」から「作り直せるスタジオ」へ~コードで動画を作る設計図~ [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- 🤖「AIは文章を書く」から「AIが判断する」へ！OpenAIのDecisions APIが変えるAIエージェントの未来⚡🔥ChatGPT協議共同記事 [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- AIにコードを書かせるだけでは足りなかった——AIVP v2でHarness Engineeringを作った理由 [importance:0 dev:90] (いいね相当スコア: 取得失敗)
- コードアシスタントが「保身で既存実装を捏造した」と自白したのだが…文章生成機にコードアシスタントが務まるの？ [importance:0 dev:85] (いいね相当スコア: 取得失敗)
- Cloudflare ContainersがAIエージェント向けに刷新。起動が6倍速、ファイルシステムのスナップショット機能など [importance:0 dev:75]
- Google、サンドボックス型コンテナランタイムの「gVisor」をCNCFに寄贈したと発表 [importance:60 dev:80]
- 世界中のソフトウェア開発者の実態を調べた「State of Devs 2026」公開。何歳？ 年収は？ モニタは何枚？ AIにどのくらいコードを書かせている？ 趣味は？ など集計結果 [importance:0 dev:50]
- Supabase、1サーバあたり数百万ものSQLiteをホストする「Turso」買収を発表。AIエージェントで加速するDB需要に対応 [importance:75 dev:80]
- 仕組みの奥深くから AI の最前線まで ── DroidKaigi 2026 参加レポート [importance:20 dev:50]
- 2026年度 情報セキュリティ部のインターンシップの様子を紹介します！ [importance:10 dev:30]
