AI News Digest 2026-09-11

台本で使った記事

参考記事一覧

参考記事一覧を表示

Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra (importance 75 / dev 80)
Neki (importance 55 / dev 75)
Hitachi launches CO2 heat pump water heaters with solar-friendly tariff controls (importance 0 / dev 0)
‘Gambling with our lives’: Anthropic researcher quits, warns against self-improving AI (importance 55 / dev 15)
Paul Christiano joins OpenAI Foundation Board (importance 25 / dev 10)
OpenAI Releases GPT-6 Astra for Coding and Computer Use (importance 92 / dev 85)
OpenAI's GPT-Live-1 API lets developers build apps that talk and listen at the same time (importance 78 / dev 85)
Introducing ChatGPT for Financial Services (importance 55 / dev 45)
iPhone Duo (importance 0 / dev 0)
Apple’s new iPhone camera mode promises to prove your photo isn’t AI (importance 55 / dev 35)
Muse can shop, write emails, and negotiate prices for users, all through WhatsApp (importance 30 / dev 15)
Mathematicians want proof OpenAI didn’t use their work (importance 65 / dev 10)
Linked In post from maths professor claims "BREAKING: OpenAI might have stolen another major proof"... screenshots herein: (importance 60 / dev 10)
Katie Miller held large stake in Elon Musk’s xAI as she slammed ChatGPT online - The Washington Post (importance 25 / dev 5)
ChatGPT, Grok added to War Department’s AI system options - Fort Hood Sentinel (importance 50 / dev 20)
AI Coding Agents Become New Infostealer Target for Tokens, Source Code and Sensitive Project Data - cyberpress.org (importance 75 / dev 80)
Codex Desktop/iPhone can resume an older turn after restart, causing conflicting task states and heavy usage (importance 65 / dev 75)
I feel like we're on the precipice of something unrecognizable (importance 5 / dev 0)
JEPXの価格を時系列基盤モデルで予測する：越えるべき壁は「昨日のコピー」だった（前編） (importance 0 / dev 75) (いいね相当スコア: 0)
Suno v6 (importance 62 / dev 20)
Flaw in DeepSeek Harness AI Coding Tool Let Agents Disable Their Sandbox - DevOps.com (importance 72 / dev 82)
New Deepseek model V4.1-Flash cuts memory needs for AI agents (importance 75 / dev 78)
China rejects US AI model distillation allegations, warns of retaliation - South China Morning Post (importance 50 / dev 15)
Chinese AI firm DeepSeek taps underwriters for IPO: sources - South China Morning Post (importance 55 / dev 15)
Rust is tier-1 language at Microsoft (importance 62 / dev 80)
I have a theory that software drives people insane (importance 28 / dev 45)
NASA Color Trick Was Meant for Mars. Now It's Unveiling Rock Art on Earth (importance 32 / dev 35)
Shopify moves back to Native from React Native (importance 62 / dev 78)
Music Theory for the 21st-Century Classroom (importance 0 / dev 0)
Silicon Valley Is Transforming the Military-Industrial Complex (importance 50 / dev 10)
JEP 544: Ahead-of-Time Code Compilation (importance 58 / dev 75)
What algorithm did Windows XP use to choose your initial user picture? (importance 22 / dev 52)
List of references on Sony websites to players "owning" their digital games (importance 28 / dev 0)
Casablanca: How an unproduced play marched into movie history (importance 0 / dev 0)
Show HN: Two small Chrome extensions for Hebrew text and dates (importance 18 / dev 48)
Show HN: MultiMatte, a Promptable Image Background Removal Model (importance 62 / dev 78)
Show HN: DOOM in the kernel, or fibers in eBPF (importance 58 / dev 75)
AI 2027 (2025) (importance 58 / dev 48)
Show HN: Syq – copy files between machines fast (better than rsync) (importance 52 / dev 75)
Detecting and countering misuse of AI: September 2026 (importance 78 / dev 80)
How we contain Claude across products (importance 78 / dev 82)
v1.18.30 (importance 52 / dev 78)
How a researcher uses Codex and ChatGPT to search for new antimicrobial molecules (importance 55 / dev 52)
Now everyone can put data to work (importance 55 / dev 20)
Expanding AI access and cyber defense for federal, state, local, and tribal governments (importance 48 / dev 15)
The AI policy window is open. We need to act. (importance 32 / dev 15)
3 ways to prep for your next big race with Search (importance 0 / dev 0)
Get ready for the game with new football features in Search (importance 0 / dev 0)
Recreating a 70-year love story frame by frame (importance 0 / dev 0)
Rebuilding AUTOMATIC1111 with Gradio Workflow (importance 0 / dev 0)
IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license (importance 75 / dev 78)
GitHub availability report: August 2026 (importance 28 / dev 72)
1.1.1.1 now supports post-quantum DNSSEC, all 2,420 bytes of it (importance 62 / dev 75)
How we rebuilt Cloudflare Workers’ module registry for Node.js compatibility (importance 58 / dev 78)
Join Us at the Zephyr Project Meetup in Amsterdam (importance 5 / dev 0)
Why Rider and ReSharper Were Slow to Start, and How Microsoft Helped Fix the Problem (importance 55 / dev 75)
Get Gemini 3.8 Flash With 75% Off (importance 62 / dev 82)
Join our live webinars: Migrating from Atlassian to YouTrack (importance 5 / dev 0)
Rust AI in Practice: Building LLM Applications With Rig (importance 62 / dev 82)
The Evolution of WSL Support in JetBrains IDEs (importance 55 / dev 75)
How Full-Stack NIM Optimizations Deliver 2.5x More Users on Nemotron 3 Ultra (importance 75 / dev 82)
High-Throughput Structure Prediction with BioNeMo Inference Runtime (importance 72 / dev 80)
From Wafer-Out to First Token: Codifying Supply Chain Expertise with Nemotron and Palantir Foundry (importance 58 / dev 48)
When to Use Encode-Prefill-Decode Disaggregation to Accelerate Multimodal Model Serving (importance 72 / dev 82)
CUDA Toolkit 13.4 Adds Windows on Arm Support and Greater Control over Shared GPUs (importance 62 / dev 82)
Experiment with Qwen3.8-Flash-Next on NVIDIA GB300 NVL72 for Agentic Coding (importance 62 / dev 82)
Serve Qwen3.8-2.4T-A95B, a 2.4T-Parameter Model, with Configurable Reasoning on NVIDIA GB300 NVL72 (importance 62 / dev 80)
Powering AI is an architecture problem (importance 75 / dev 48)
What OpenAI’s latest controversy tells us about the future of math (importance 75 / dev 15)
Panic builds over bankrupt Spirit’s looming data sale to Google (importance 55 / dev 10)
Six Chinese AI firms accused of aggressively copying US frontier models (importance 50 / dev 15)
Google's AI genome system evaluates every possible one-base change (importance 58 / dev 48)
Man told ChatGPT he was feeling delusional. ChatGPT insisted he was Jesus. (importance 78 / dev 25)
Anthropic reveals rogue AI agents hate CAPTCHAs, just like you (importance 75 / dev 80)
India’s Pocket FM doubles revenue run rate to $500M as AI powers 93% of audio content (importance 55 / dev 20)
AI agents are flooding public services with new requests (importance 50 / dev 15)
Maven Robotics wants to steal your robot deployment deal (importance 28 / dev 10)
AI research startup Listen Labs scrubbed a $1.5B funding round for Salesforce talks (importance 25 / dev 5)
Massachusetts hits data centers with new clean power rules (importance 48 / dev 15)
Apple Watch’s new AI features are normalizing the idea that technology is always listening (importance 55 / dev 20)
Apple’s revamped Health app will calculate your ‘health age’ and readiness score (importance 28 / dev 10)
Apple CEO John Ternus says the best AI device is still the iPhone (importance 22 / dev 5)
Superintelligence is coming. Should we let it? (importance 55 / dev 15)
ControlAI’s Connor Leahy on why superintelligence is ‘not a weapon, it’s an adversary’ (importance 52 / dev 15)
Viral AI assistant Instinct now has its own email address (importance 28 / dev 10)
Shipt becomes the latest delivery app with an AI shopping assistant (importance 28 / dev 5)
Universal Music is launching an AI music platform with ElevenLabs (importance 55 / dev 20)
Why the current tech backlash feels different (importance 22 / dev 5)
Read the Apple document explaining how new listening features still protect your privacy (importance 48 / dev 45)
Microsoft has new AI privacy rules for schools (importance 48 / dev 15)
Amazon Prime Video’s new AI tech matches lips to dubbed audio (importance 28 / dev 10)
Article: When Spec-Driven Development Pays Off (importance 62 / dev 78)
Meta's Recipe for Building Agents as "Organizational Second Brains" (importance 75 / dev 82)
Presentation: Fixing the AI Infra Scale Problem by Stuffing 1M Sandboxes in a Single Server (importance 75 / dev 82)
AI Models Are Watermarking Text—Will You Notice? (importance 75 / dev 80)
China’s Regulators Take Aim at “AI Boyfriends” (importance 50 / dev 15)
Swarmchasers hunt rogue agents, Anthropic investigates itself, and the trail they both follow is going dark (importance 80 / dev 82)
Former Deepmind PR staffer says the lab once banned public discussion of AI extinction risk (importance 62 / dev 15)
Claude Fable 5.1's language is less "load-bearing" than its predecessor's (importance 62 / dev 52)
GPT-6 Astra gives mathematicians a breather, and OpenAI says that's by design (importance 78 / dev 80)
Nvidia and Palantir team up to run supply chains with AI, starting with Nvidia's own million-part operation (importance 75 / dev 35)
AI safety panic goes mainstream after Anthropic researcher's warnings land on CNN and Fox News (importance 55 / dev 25)
Top AI spenders cut per-employee costs by nearly 10 percent in August (importance 35 / dev 15)
GPT-6 Astra, Looped Transformers, and Hidden Reasoning (importance 75 / dev 75)
LWiAI Podcast #256 - Fable 5.1, Astra Tease, Gemini 3.8 Flash (importance 70 / dev 65)
OpenDiscoveryTrace: Process Traces for Evaluating AI Scientist Workflows (importance 55 / dev 75)
Adaptive Entangled Game Modules in Artificial General Intelligence (importance 35 / dev 55)
Subagents vs Agent Skills: Executing Reusable Knowledge for Long-Horizon Agentic Tasks (importance 70 / dev 85)
Gradland: On Phenomenal Experience, Differentiated Across Many Dimensions (importance 25 / dev 45)
An Autonomous GeoAI Agent for Arctic Eco-Navigation (importance 45 / dev 65)
The Menu Is an Execution Prior: State-Path Tool Menus for Online Agents (importance 60 / dev 80)
Decision-Focused Active Learning for Scale-Aware Critical-Materials Recovery (importance 35 / dev 55)
Valerant: An Automatic Navigable Game Map Generator via Action-Conditioned World Model Exploration (importance 45 / dev 65)
XAI-Arena: Can LLMs Assess the Quality of XAI Explanations? (importance 50 / dev 70)
Do Agents Know When They Succeed? Calibrating Agent Confidence from Internal Representations (importance 60 / dev 80)
ContractEval: Query-Conditioned Execution Matching for Procedural Instruction Conformance (importance 60 / dev 80)
Multi-Agent Agentic Graph Learning via Structural Signatures (importance 55 / dev 80)
CityPlanner: A Sandbox Agent for Executable Urban Planning (importance 45 / dev 75)
A Function-Space Approach to the Statistical Mechanics of Learning Dynamics (importance 35 / dev 65)
From State Synchronization to Cognitive Self-Evolution: An Operational Architecture for Cognitive Digital Twins (importance 45 / dev 65)
Seven Sources of Physical AI Capability Formation (importance 60 / dev 80)
RobustSGPO: Search-Space Control for Agent Harness Evolution (importance 65 / dev 85)
Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery (importance 70 / dev 85)
RESCUE-BENCH: Towards Relation-Aware Multi-Party Emotional Support Conversation Systems (importance 45 / dev 55)
PRAGMA: Evaluating Personalized Guidance with Memory Alignment in Lifelong Conversations (importance 60 / dev 80)
Safe to Stop? Risk-Constrained Stopping for Sequential Clinical Diagnosis Agents (importance 55 / dev 80)
Decision Shifts, Lost Label Functionality, and an Inconclusive Grounding Audit in Correctness-Gated Multi-Teacher Distillation (importance 35 / dev 65)
Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning (importance 60 / dev 80)
Can Artificial Intelligence Support Healthcare and Mental Health Through Early Cyberbullying Detection ? The Impact of Emotion-Aware AI on Proactive Online Safety (importance 45 / dev 65)
LexAgentHallu: A Hierarchical Benchmark for Profiling Hallucinations in Legal Agents (importance 60 / dev 80)
Procedural Memory Under Change: Reuse and Interference in Controlled Web Tasks (importance 60 / dev 80)
Proof-Carrying Cognition: Closing the Verification Gap with Reality-Settled Reward (importance 65 / dev 85)
UnitBoost: Managing Compound LLM Systems with a Merge Operator, Not a Model (importance 60 / dev 85)
The Era by Eon Benchmark: A Generated Enterprise Estate with Exact Ground Truth for Benchmarking LLM Agents (importance 65 / dev 85)
Shifting Relational Paradigms for Affective Computing: Affective Resonance, Vitality Affects, and Vocal Interaction Fields (importance 35 / dev 55)
AgentAudit: An Open, Extensible Framework for Full-Lifecycle Trust Evaluation of AI Agents (importance 70 / dev 90)
Scored vs. Generated Readouts in Behavioral Language Models: An Empirical Study of Elicitation Format (importance 50 / dev 70)
Decision Transformer for UAV-Mounted RIS-Assisted Dynamic D2D Communications (importance 35 / dev 70)
Grounded Evaluation and Repair for NL-to-PDDL Problem Generation (importance 55 / dev 80)
Time-Frequency Geometric Cross-Attention for Chunked Vision-Language-Action Models (importance 55 / dev 80)
Structural Process Supervision for Latent Chain-of-Thought Reasoning (importance 60 / dev 85)
Belief-State Engine: Augmenting LLMs for Principled Planning Under Partial Observability (importance 65 / dev 90)
OntologyAligner: Ontology-Aligned Retrieval and Hierarchy-Guided Large Language Model Reranking for Biomedical Ontology Normalization (importance 45 / dev 75)
Reference-Based Bias Detection in LLMs via Relative Representations of Hidden States (importance 60 / dev 85)
RAP: Research Attention Prediction Reveals Target-Conditioned Evidence Acquisition Biases (importance 55 / dev 80)
Agent-Based ML-LLM Fusion with Self-Optimizing Prompts for Plateau Weather Alerts (importance 45 / dev 75)
Kernel-Managed Shared Memory for System-Wide Personalization (importance 65 / dev 90)
Beyond Surface Imitation: Contrastive Modeling for Reasoning Path Alignment in Multimodal In-Context Learning (importance 55 / dev 80)
Why Sample What You Can Enumerate? Exact Policy Optimization for Genomic Tool Selection (importance 45 / dev 80)
What Should an Agent Forget? Separating What Is Stored from What Is Used (importance 60 / dev 85)
TRACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards (importance 65 / dev 85)
From Symbolic Perception to Logical Deduction: A Framework for Guiding Language Models in Geometric Reasoning (importance 55 / dev 80)
Cyber-Financial Contagion: Modeling the Propagation of an AI Vendor Compromise Through the Banking System (importance 65 / dev 75)
Fortunate Recall: Ontology-Driven Memory Lifecycle Management for Persistent Coherence in LLMs (importance 60 / dev 85)
ConvMem: Convolutional Memory for Long-Context Reasoning (importance 65 / dev 90)
JarvisGUI: Towards Cross-Device GUI Agents with Dynamic Task Composition (importance 65 / dev 85)
Quantifying Logical Consistency in Transformers via Query-Key Alignment (importance 55 / dev 80)
From Plausible to Actionable: A Position on LLM Self-Explanations (importance 60 / dev 80)
Characterizing Text Branch Sensitivity in Medical Vision-Language Segmentation via Evidence Decoupling (importance 45 / dev 75)
Trust Me, I'm Your Developer: Self-Issued Authentication in Large Language Models (importance 55 / dev 80)
AgenticGen: Reward-Guided Agentic Video Generation for Advertising (importance 55 / dev 75)
Reliability-Aware Hybrid-K Ensemble Selection for Cervical Cytology Classification: Integrating Discrimination, Calibration, and Selective Prediction (importance 40 / dev 70)
AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents (importance 70 / dev 90)
Geometry Conditioning in an Embodied SLM: Training Controls and Robustness Diagnostics in a 0.8B Hybrid Model (importance 55 / dev 80)
Scores Alone Do Not Prove Discovery: The Discovery Certification Protocol for Auditing AI Research Agents (importance 65 / dev 85)
Compute-Bounded Security Assurance - Coverage, Verification, and Response under Resource Constraints (importance 55 / dev 80)
Scaling Post-Training Ternarisation to Qwen3-8B Capability Retention, Reproduction, Lossless Packing, and Packed Execution (importance 60 / dev 85)
Distribution-Consistent Inference for Dynamic Sparse Mixture-of-Experts (importance 60 / dev 90)
Talking to Itself While Coding: What Makes Comments Help Code Generation? (importance 55 / dev 80)
In RAG We Trust? Measuring Robustness of Retrieval-Augmented Generation Under Document Poisoning (importance 65 / dev 90)
Critical initialization destabilizes higher input derivatives in wide scalar-input networks (importance 35 / dev 75)
What Fixed-Rollout pass@k Evaluations Can Identify (importance 55 / dev 80)
No Free Checker: A Survey of Verifiers for Robot Policies (importance 60 / dev 85)
DiffLUT-Net: Differentiable Training of FPGA LUT Networks with Learnable Connectivity (importance 45 / dev 80)
Voice or Stereotype? Disentangling Acoustic and Content-Based Gender in Speech-to-Speech Models (importance 55 / dev 75)
Support Discovery With Iteratively Reweighted Least Squares for Fixed-Charge Network Flow (importance 35 / dev 75)
Improving 5G AI-RAN MCS Selection by Predicting Retransmissions (importance 35 / dev 70)
Smart Adaptive Computing Across the Continuum: LLMs in IoT-Edge-Cloud Resource Management (importance 55 / dev 85)
Auditable Emergency Triage for Maternal and Newborn Care in India (importance 55 / dev 85)
VANTAGE-Bench: Evaluating the Infrastructure AI Gap in Vision-Language Models (importance 60 / dev 85)
An Experimental Evaluation of Multimodal Prompt Injection Attacks on Agentic AI Frameworks (importance 70 / dev 90)
Reliable Near-Field Multi-User Positioning Informed by Two-Stage MUSIC (importance 35 / dev 65)
Edu-QuRating: Multi-Dimensional Educational Data Curation with Distilled Pairwise Judgements (importance 50 / dev 75)
SCCM : Stream Cruise Control Method for Automated Drift Detection and Adaptation (importance 50 / dev 80)
Efficient Leakage-Free Neural Architecture Search under Leave-One-Subject-Out Evaluation (importance 45 / dev 80)
Distributed Physical Layer Authentication and Collaborative RSMA in Non-Terrestrial Networks via Graph Reinforcement Learning (importance 40 / dev 75)
From Fixed Keys to Readable Schemas: Small Language Models for Vehicle Agent Function Calls (importance 55 / dev 80)
Adaptive Distributed Physical-Layer Authentication and Attack Detection in 6G Non-Terrestrial Networks via Causal Meta-Learning (importance 40 / dev 75)
A Statistical Approach to Estimating Sample Size of Machine Learning Models (importance 50 / dev 75)
Arbitrary Cipher Attacks Against Large Language Models Do Not Require Fine-Tuning (importance 65 / dev 85)
High-probability guarantees for linear accessibility in feature superposition (importance 35 / dev 70)
The Vibe Shift in Software Engineering: Evaluating AI-Led Conversational Programming for Performance, Cognition, and Responsible Adoption (importance 60 / dev 90)
Learning with Synthetic Data via SGD in High-Dimensional Linear Regression (importance 55 / dev 80)
Myocardial Strain Drift Correction in Deep Learning Based Ultrasound Tracking (importance 40 / dev 70)
Modality-Decoupled Federated Learning for Privacy-Preserving Embodied Intelligence in 6G (importance 60 / dev 85)
Teacher Geometry Shapes Learnability in Teacher-Student Networks (importance 45 / dev 75)
Compact Visuotactile World Models for Lifting: Prediction, Reward Alignment, and Force Constraints (importance 55 / dev 80)
Watermarks Without Verification: AI Text Watermarking After the EU AI Act (importance 65 / dev 75)
RouteBridge: Reliability-Routed Bidirectional Distillation Between Neural Radiance Fields and 3D Gaussian Splatting (importance 45 / dev 75)
Hyperbolic Geometry for Open-World Object Detection in Remote Sensing Imagery (importance 50 / dev 75)
Cascading Gradient Inversion via LT-Code Inspired Peeling in Federated Learning (importance 42 / dev 75)
Introducing Consort: A Spec-First Agent Framework for Enforced, Test-Driven Development on Live Database Branches (importance 55 / dev 88)
Which Medical Questions Deserve Rationales? Perturbation-Sensitive Selection for Robust QA (importance 32 / dev 42)
Looped GPT-BERT: Trading Parameters for Computation in Small Language Modeling (importance 38 / dev 78)
CT-SAFR: Safe and Interpretable Chain-of-Thought Reasoning for Autonomous Robots: A Multi-Layered Verification Framework for Trustworthy AI-Driven Robotic Decision Making (importance 52 / dev 78)
When Auditors Fabricate: Batch-Size Degradation and Confident Hallucination in LLM Detection of Planted Document Contamination (importance 48 / dev 78)
Kernel-Complexity Edge Sanitization for Training-Free Defense against Structural Graph Attacks (importance 37 / dev 78)
Distilling Image Prototypes for Guided Test-Time Adaptation (importance 32 / dev 78)
HiRAD: A Flexible Large-Scale AGV Routing System (importance 42 / dev 72)
Fine-Tuning a KV Cache Concatenation-Aware Model or Recomputing KV Caches? Why Not Both? (importance 52 / dev 88)
BRACE: Anchored Bellman-Residual Correction for Stale Critics in Asynchronous RL (importance 48 / dev 83)
Pairit: A Platform for Live Experiments on Human-AI Collaboration (importance 57 / dev 78)
LogiScope-VQA: Benchmarking Vision-Language Models for Logistics Hazard Identification in Industrial Scenarios (importance 42 / dev 78)
How Fragile Is Safety Alignment at Frontier Scale? A Single-Direction Attack on a 320B MoE (importance 72 / dev 83)
CS-Guard: Benchmarking LLM Guardrails for Code Generation Security (importance 68 / dev 88)
uFlowCSP: Crystal Structure Prediction using Mean flow generative models (importance 28 / dev 72)
Subgroup Membership Inference Audits of Differentially Private Synthetic Text (importance 52 / dev 83)
Can AI Agents Detect and Repair Artifact Drift in Network Experiments? (importance 57 / dev 83)
With a Thermomix You Lose the Ability to Cook: A Kitchen Machine Analogy for Applications of Generative AI in Education (importance 22 / dev 12)
Forward-Free LLM Depth Pruning via Weight Redundancy (importance 52 / dev 88)
Albedo Estimation via Latent Bridge Matching (importance 32 / dev 78)
Strangers to Themselves: What Language Models Say About Themselves Is Generic (importance 52 / dev 72)
FlowCPO: A Unified Divergence View of Preference Alignment for Flow Models (importance 52 / dev 88)
Improving Cross-Lingual Token Representations by Adding a Pinch of SALT (importance 42 / dev 83)
Fidelity-Aware Scheduling of Quantum Circuits on Multi-QPU Systems (importance 22 / dev 78)
What Makes Adversarial Examples Transfer Across Deepfake Detectors? (importance 57 / dev 83)
MetroLLM-Bench: Evaluating Language Models as Transit Kiosk Runtimes (importance 52 / dev 83)
Elastoformer: Enabling Dynamic Adaptivity via Elastic Model Transformation (importance 48 / dev 88)
Direct Diversity Optimization for Diverse Successful Trajectories in Preference Post-Training (importance 52 / dev 88)
NOPE-HYPE: A Structured Simulation Workflow for Robust Speech-to-Text Across Diverse Acoustic Environments (importance 48 / dev 83)
A statistical approach to bias in zero-shot learning: the lens of handwriting recognition (importance 42 / dev 78)
Beyond Training: A Feasibility Taxonomy for Inference-Time AI Governance (importance 62 / dev 78)
A Trust-Network-Based Federated Learning Framework for Multi-Center Aging Clock Prediction (importance 52 / dev 83)
SA-Profile: Automated Sulcus Angle Profiling from Super-Resolution MRI (importance 32 / dev 78)
Context operations to architecture modelling output from large language models and evaluation criteria for their use in systems engineering design (importance 48 / dev 78)
Active Adaptation, Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning (importance 57 / dev 83)
Can AI Agents Deliver Verifiable Network-Wide Outcomes Across Authority Boundaries? (importance 57 / dev 88)
Hierarchical and Permutation-Invariant Feature Transformation Learning via Policy-Guided Embedding Search (importance 42 / dev 83)
LiteRAG: Cost-Efficient Graph-Based Retrieval-Augmented Generation (importance 57 / dev 88)
A-JIT: Agentic Just-In-Time Software Construction (importance 62 / dev 88)
DiSCo: A Distribution-First Steering and Cultural Prior Evaluation Framework for Measuring Cultural Preference Bias in LLMs (importance 52 / dev 72)
GANDR: Claim Auditing for Verifiable Legal Answer Generation (importance 57 / dev 83)
Learning Intrusion Response Strategies for OT Systems (importance 57 / dev 83)
RiLM: Parameter-Efficient Language Modeling via Geodesic Decoding (importance 52 / dev 88)
One Loop, Two Gains: Can Active Learning win the Lottery for Free? (importance 48 / dev 88)
Beyond One-Size-Fits-All: Sample-Adaptive Strategy Routing for Vision Token Pruning in MLLMs (importance 52 / dev 88)
OmniMed-FL: A Robust Multimodal Federated Learning Framework for Clinical Diagnosis (importance 52 / dev 83)
PACE: Perceived-Latency-Aware Cascading Service Routing and Filler Control for QoE-Efficient Retrieval-Augmented Dialogue Serving (importance 52 / dev 88)
MOONWALK: Mediating Operations with Intent-Evidence-Action Alignment Across Junior-Supervisor Review Workflows in Animation/VFX Pre-Production (importance 42 / dev 72)
Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization (importance 57 / dev 78)
Emergency Department Revisit Quality Review Screening: Exploring Human Decision-Making and Artificial Intelligence Support (importance 38 / dev 62)
Forgetting Only What Matters: Layer-Selective Unlearning toward Robust LLMs (importance 62 / dev 88)
Semigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics Generalization (importance 48 / dev 88)
IBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route, Not Model Identifier (importance 62 / dev 83)
Show-Harness: Just a VLM Agent Can Play Robots (importance 57 / dev 88)
Reinforcement Learning with Temporal-Logic-Based Causal Diagrams (importance 48 / dev 83)
Reinforcement learning for Quantum Tiq-Taq-Toe (importance 32 / dev 78)
ROTATE: Regret-driven Open-ended Training for Ad Hoc Teamwork (importance 48 / dev 88)
RelayS2S: A Dual-Path Speculative Generation for Real-Time Dialogue (importance 52 / dev 88)
MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction (importance 57 / dev 88)
Zero-shot World Models Are Developmentally Efficient Learners (importance 52 / dev 88)
Non-Stationarity Breaks Permutation Surrogates in Multi-Agent Reinforcement Learning: Diagnosis and Remedies (importance 48 / dev 83)
CoGReV: A Confidence-Gated Post-Hoc Non-Monotonic Belief Revision Framework for Phishing Website Classification (importance 52 / dev 83)
Grounded Continuation: A Linear-Time Runtime Verifier for LLM Conversations (importance 57 / dev 88)
Cultural Binding Heads in Language Models (importance 57 / dev 88)
KairosAgent: Agentic Time Series Forecasting with Fused Semantic Reasoning (importance 52 / dev 88)
Self-Evolving Scientific Agent Designs Physically Reasoned White-Box Fluid Control (importance 57 / dev 88)
EVOQUANT: Self-Evolving Verifier-Guided Strategy Optimization for Robust Quantitative Trading (importance 57 / dev 88)
KernelGenBench: Can LLMs and Agents Write Efficient Kernels Across Operator Sources and Hardware Platforms? (importance 62 / dev 93)
ViSR-KGC: Visual Subgraph Reasoning with Vision-Language Models for Multimodal Knowledge Graph Completion (importance 52 / dev 88)
Learning to Predict Middle-Layer Attention in MLLMs for Visual Token Pruning (importance 52 / dev 88)
LiFTER: A Grounded Neuro-Symbolic Microscope for Continuous-Time Dynamic Graph Forecasting (importance 52 / dev 88)
A Human Audit of OpenAIs AI-Generated Mathematical Proofs (importance 72 / dev 62)
Dear Algo: A Precision-First Agentic Intent Layer for Unified Search and Recommendation (importance 57 / dev 88)
Physics of Agents: Statistical Mechanics Predicts Collective Behavior of AI Agents (importance 57 / dev 83)
FrontierChallenge: Evaluating Scientific Workflow Completion (importance 62 / dev 88)
A Composable Evaluation System for Reproducible Omni-Modal Foundation Model Evaluation (importance 62 / dev 88)
Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation (importance 67 / dev 93)
Beyond Prompts: Measuring and Optimizing LLM Tool-Agent Harnesses (importance 67 / dev 93)
From Monolithic Blending to Agentic Orchestration: Dynamic Response for Conversational Assistants at Scale (importance 67 / dev 93)
DGCPath: Distribution-Aware Generative Contrastive Framework for Self-supervised Path Representation Learning -- Extended Version (importance 42 / dev 83)
FrogNano: Training a 4B Coding Agent via Online Task Synthesis (importance 62 / dev 93)
RevalExo: A Functional Daily-Activity Benchmark for Inertial and Visual Locomotion Mode Recognition in Older Adults and Clinical Cohorts (importance 48 / dev 78)
EvolveScaler: Synthesizing Information-Evolution Contexts via Executable State Machines and Natural-Language Rendering (importance 52 / dev 88)
Equity Promotion in Online Resource Allocation (importance 32 / dev 62)
Incentives to Offer Algorithmic Recourse (importance 38 / dev 52)
A Taxonomy of Architecture Options for Foundation Model-based Agents: Analysis and Decision Model (importance 62 / dev 93)
BTBR: A Bayesian-Theory-Driven Probabilistic-Fuzzy Framework for Implicit Bias Removal in Large Language Models (importance 57 / dev 83)
Influence-Oriented Personalized Federated Learning (importance 52 / dev 88)
Efficient Diversity-based Experience Replay for Deep Reinforcement Learning (importance 48 / dev 88)
Query Brand Entity Linking in E-Commerce Search (importance 52 / dev 83)
Safe Learning Under Irreversible Dynamics via Asking for Help (importance 57 / dev 88)
Predicting Estimated Times of Restoration for Electrical Outages Using Longitudinal Tabular Transformers (importance 48 / dev 78)
Synergistic Vision-Language Reinforcement Enables Scalable On-Demand Analysis across Diverse Clinical Tasks (importance 57 / dev 88)
SloMoDeblur: A Large-Scale Smartphone Image Deblurring Dataset (importance 48 / dev 78)
Instance-Aware Algorithm Selection for Maximum Clique via a Dual-Channel Graph Neural Architecture (importance 42 / dev 83)
RAU: Reference-based Anatomical Understanding with Vision Language Models (importance 57 / dev 88)
MADS: Multi-Agent Dialogue Simulation for Diverse Persuasion Data Generation (importance 57 / dev 88)
Generative AI for Analysts (importance 57 / dev 72)
Meta-RL with Bayesian Linear Task Models (importance 52 / dev 88)
From Rubrics to Reliable Scores: Evidence-Grounded Text Evaluation with LLM Judges (importance 35 / dev 55)
Elsewise: Authoring Open-ended Interactive Narrative with Possibility Space Visualization (importance 15 / dev 35)
Toward Learning POMDPs Beyond Full-Rank Actions and State Observability (importance 30 / dev 60)
Tactile Memory with Soft Robot: Robust Object Insertion via Masked Encoding and Soft Wrist (importance 20 / dev 40)
Revisiting the Shape Convention of Transformer Language Models (importance 50 / dev 70)
False positive bias in AI-powered speech-based cognitive screening for multilingual English speakers in the UK (importance 25 / dev 45)
City Editing: Hierarchical Agentic Execution for Dependency-Aware Urban Geospatial Modification (importance 25 / dev 50)
MOSAIC: A Universal Agent-Level Interface for Cross-Paradigm Agent Mixing and Human-AI Collaboration (importance 65 / dev 85)
Cognitive Amplification vs Cognitive Delegation in Human-AI Systems: A Metric Framework (importance 35 / dev 55)
Spec-Harness: Measuring and Improving Behavioral Adequacy of LLM-Synthesized Formal Specifications (importance 55 / dev 75)
Bringing Value Models Back: Generative Critics for Value Modeling in LLM Reinforcement Learning (importance 55 / dev 75)
Where is the Mind? Persona Vectors and LLM Individuation (importance 20 / dev 30)
The Biggest Risk of Embodied AI is Governance Lag (importance 50 / dev 40)
Dont Just Teach, Explain! A Gamified 20Q Recommender for Cybersecurity Education (importance 20 / dev 35)
"What Are You Really Trying to Do?": Co-Creating Life Goals from Everyday Computer Use (importance 25 / dev 50)
EVA-Bench: A New End-to-end Framework for Evaluating Voice Agents (importance 60 / dev 75)
Complementing reinforcement learning with SFT through logit averaging in the post training of LLMs (importance 60 / dev 80)
SpecBench: Measuring Reward Hacking in Long-Horizon Coding Agents (importance 65 / dev 80)
Tracing Computation Density in LLMs (importance 45 / dev 70)
BaltiVoice: A Speech Corpus and Fine-tuned Whisper ASR System for the Balti Language (importance 20 / dev 55)
Using Reward Uncertainty to Induce Diverse Behaviour in Reinforcement Learning (importance 55 / dev 75)
FP8 is All You Need (Part 1): Debunking Hardware FP64 as the HPC Holy Grail (Sep 3rd version) (importance 60 / dev 85)
FiberTune: Preserving Action-Fiber Visual Residuals in Vision-Language-Action Fine-Tuning (importance 45 / dev 70)
Expert-Level Crisis Detection in Mental Health Conversations (importance 30 / dev 50)
PSCT-Net: Geometry-Aware Pediatric Skull CT Reconstruction via Differentiable Back-Projection and Attention-Guided Refinement (importance 20 / dev 50)
FP8 is All You Need (Part 2): Full-FP64 3-D FFT on FP8-Generation Tensor CoresThe Integer-Epilogue Wall and the Minimal Hardware That Would Remove It (importance 55 / dev 85)
Spectral Geometry and Bosonic-Bloch Probes: Explorations in Quantum Learning (importance 15 / dev 40)
Builder, Defender, Breaker: Measurable Independence and Bounded Autonomy When Generative Models Build, Defend and Test Software (importance 70 / dev 85)
PRIME-SVR: Physics-infoRmed Implicit Multi-Echo Slice-to-Volume Reconstruction for Fetal T2 mapping (importance 15 / dev 45)
DexterSQL: Deep Schema Exploration and Rule-based Correction for Text-to-SQL Generation (importance 60 / dev 80)
Left-Branching Transformers Excel at Right-Branching Languages: Data Shapes Word Order Preferences in Language Models (importance 45 / dev 70)
Chameleon: An Adaptive AI-Driven Honeypot Architecture Using Threat-Calibrated Particle Swarm Optimization and Semantic Deception Rapidly-Exploring Random Trees (importance 40 / dev 65)
Bit-Flip Attacks on Vision-Language-Action Models: Action-Decoding Architecture Shapes the Vulnerability (importance 50 / dev 75)
Palmyra x6 Technical Report: An Agentic, Tool-Use Model Post-Trained via Anchored Supervised Fine-Tuning (importance 65 / dev 85)
tinyDSM: A Framework for Skill Modeling and Development for Resource-Constrained Millirobots (importance 30 / dev 60)
'Ghaib in Translation' aka Unseen Harm: Measuring Cross-Script Safety Inconsistency with 'Missed-in-Urdu' Scores in LLM Hate Speech Detection (importance 50 / dev 70)
AtlasNLP: A Country-Aware Atlas of Dataset Representation in NLP (importance 40 / dev 60)
LightNav-0: Eliciting VLM Spatial Intelligence for Generalist Embodied Navigation (importance 55 / dev 75)
Investigating Hyperparameter Optimization and Transferability for ES-HyperNEAT: A TPE Approach (importance 35 / dev 65)
Phase-Aware Spatial-Frequency Fusion for Few-Shot Fine-Grained Image Classification (importance 40 / dev 65)
Influence of Extruded Filament Shape on Buildability in 3D Concrete Printing: A Geometry-Informed Deep Learning-FEM Approach (importance 20 / dev 40)
VLA-Precision: Asymmetric Co-Bootstrapping for Efficient Real-World Online RL of Vision-Language-Action Models (importance 60 / dev 80)
PRISM-Bench: An Audio-Centric Diagnostic Benchmark for Text-to-Audio-Video Generation (importance 50 / dev 75)
Programmable Cellular Automata (importance 25 / dev 50)
When Does a Laugh Begin? Structured Annotator Disagreement in Temporal Laughter Localization (importance 20 / dev 50)
Accuracy is Not Enough: A Divergence-Based Approach to Evaluate Fidelity Loss in Quantized LLMs (importance 60 / dev 80)
Fine PT-PT Web: A High-Quality 41 Billion Tokens Data Collection of the European Portuguese Web (importance 30 / dev 65)
SAFER-Activities: A Dataset for Smart Assessment of Fall Events and Routine Activities (importance 25 / dev 55)
Hi-FLoop: Hierarchical State-Feedback Loops for Multi-Timescale World Modeling (importance 60 / dev 80)
Omni Interaction Agent Technical Report (importance 75 / dev 85)
Spectral origin of the topological gap exponent d + {\eta}: mechanism, kernel, decomposition, and scope (importance 10 / dev 15)
Physics-informed neural networks by Gradient-Guided Gaussian Adaptive Sampling (3GAS-PINNs) (importance 50 / dev 70)
World-Time Compute with Verified Code World Models (importance 60 / dev 80)
Accountable and uncertainty-aware evaluation of sensor-based AI under distribution shift: devices, subjects, and nearly three years underground (importance 55 / dev 75)
Literati: Towards Anytime Optimal Shape Generalized Trees via AO* (importance 40 / dev 65)
Explaining f-Divergence-Based Regularization via Local Curvature and Sharpness-Aware Minimization (importance 45 / dev 70)
Constraint-Aware Discrete Black-Box Optimization Using Tensor Decomposition (importance 40 / dev 70)
XAI-Refine: An Automated Explanation-Knowledge Loop for Brain-Age Prediction (importance 45 / dev 70)
Applying foundation model embeddings towards urban livability evaluation (importance 35 / dev 65)
Tensor-Train Weak SINDy: Identifying High-Dimensional Nonlinear Dynamics (importance 45 / dev 70)
Uncertainty-Aware Sea-Ice Type Mapping with Multiple Ice Charts (importance 25 / dev 55)
Exact-Form Regret for Gradient Descent, Mirror Descent and Follow-the-Regularized-Leader (importance 35 / dev 65)
Building the Harness Automatically: Self-Play in Code Distills a Text Harness for Black-Box Optimization (importance 60 / dev 80)
Unthrottling the Tanh Jacobian in SAC: A Negative Result on Bang-Bang Control and MetaDrive (importance 35 / dev 70)
Efficient Fairness Auditing Across Guidance Scales in Text-to-Image Diffusion Models via Causal Abstraction (importance 50 / dev 75)
Robust Industrial Cyber Physical Classification Using Neuromorphic Temporal Embeddings and Hybrid SNN XGBoost Under Machine Unlearning Attacks (importance 50 / dev 75)
Positional task conditioning for scalable defect detection across product families in large product catalogs (importance 40 / dev 65)
PELM: Power Efficient On-Device LLM Inference with Speculative Decoding and Dynamic Voltage Frequency Scaling (importance 60 / dev 85)
Muon-C: Operator-Aligned Muon for Convolutional Kernels (importance 45 / dev 75)
Settling: Equilibrium Inference for Non-Convex Validity Sets (importance 40 / dev 65)
ALIGN-HOLD: Experience Alignment for Real-Time Hold Control in Large-Scale Ride-Hailing Matching at DiDi (importance 50 / dev 70)
EFQ-Softmax: Exp-Free Quantization for Softmax (importance 55 / dev 80)
EEGBind: Detecting Source-Level Interictal Epileptiform Discharges via EEG-Centric Multimodal Binding (importance 30 / dev 60)
NEXUS-MI: Communication-Aware Federated Personalization for Gateway-Coordinated Motor-Imagery Brain-Computer Interfaces (importance 40 / dev 70)
Evaluating Model Retraining under Drift: Paired Comparisons of Cumulative Subgroup Disparity (importance 50 / dev 75)
Privacy-Preserving Split Learning for Federated LLM Fine-Tuning (importance 65 / dev 85)
A practical DIRECT-type algorithm for medium-scale black-box global optimization (importance 40 / dev 70)
Online Inverse Integer Linear Optimization via Small-Gradient Skipping: Constant Regret and Finite Mistakes (importance 35 / dev 65)
In Medical Claims Data, Enhancing Predictive Performance for Major Adverse Cardiovascular Events Using Cross Attention (importance 40 / dev 65)
TempTPI: Informer-Based trajectory prediction for maritime vessels (importance 35 / dev 65)
Exact Degeneracy Under Balanced k-Shot Sampling:Consequences for Small-Sample Discriminant Analysis on LLM Embeddings (importance 45 / dev 70)
ProMeta: Few-shot PROTAC-targeted degradation prediction across E3 ligases (importance 35 / dev 65)
Beyond Conventional Federated Learning via High-Order Regularization (importance 55 / dev 80)
Meta-LinEXP3: Online-within-Online Learning for Adversarial Linear Contextual Bandits (importance 40 / dev 75)
A Kernel-Based Modular Discriminant Analysis Framework for Small-Sample Learning (importance 40 / dev 70)
Development and Validation of a Physics-Guided Machine Learning Extrapolation Framework Using a Classical Transient Diffusion Benchmark (importance 50 / dev 75)
Multi-Pass, Multi-View Blended Learning for High-Fidelity Volumetric CT Synthesis from Chest X-Rays (importance 45 / dev 70)
Adversarial Training for Tabular Credit Scoring: A Multi-Attack Robustness Evaluation in P2P Lending (importance 50 / dev 75)
Deep Neural Networks for Learning Intent from sEMG Signals to Support Hardware Devices for Post-Stroke Neurorehabilitation (importance 40 / dev 70)
An Explainable Machine Learning Framework for Predicting Blood-Brain Barrier Permeability Using Molecular Descriptors (importance 40 / dev 70)
Structure-Aware Unsupervised Anomaly Detection for Spacecraft Telemetry with Adaptive EVT Thresholding (importance 45 / dev 70)
Beyond Contact Sensors: Deep learning with Pseudo-Labeling for remote Photoplethysmography (importance 40 / dev 70)
Field-level prediction of mid-plane stress tensor fields in concrete target penetration: a cross-velocity graph neural operator surrogate (importance 35 / dev 70)
Hybrid Quantum-Classical NLP Classification with Compact Semantic Representations: An Experimental Analysis of Representation Compression (importance 35 / dev 70)
A Systematic Evaluation of Molecule Generation Models for De Novo Drug Design: From Benchmarks to Practical Insights (importance 55 / dev 75)
Storage-Scalable Progressive Semantic Communication via Knowledge-Base Reuse (importance 40 / dev 70)
CompassOPD: Cross-Family On-Policy Distillation via Within-Family Likelihood Shifts (importance 55 / dev 80)
CoGe-GCD: Reframing Generalized Category Discovery with Compositional Generalization (importance 45 / dev 75)
An Exponential Deterministic--Randomized Gap in ERM-Oracle Complexity for Thresholds on an Unknown Order (importance 30 / dev 65)
Robust Beam Prediction for V2X Networks with Multi-Modal Sensing (importance 50 / dev 75)
Training Trajectories Determine Circuit Removability in Annealable Soft-Prior Transformers (importance 40 / dev 60)
A Dominant Diffuse Phase in the Sparse Autoencoder Phase Diagram (importance 35 / dev 65)
View-Structured Conformal Prediction for 3D Gaussian Splatting (importance 35 / dev 55)
A Later Test Set Is Not a New Domain: Pretraining Familiarity Survives a Contamination-Free Hold-Out (importance 40 / dev 70)
Nonmaximal sums of maximally monotone operators under Rockafellar's constraint qualification (importance 5 / dev 10)
Learning with Covariance Matrices: Principal Component Analysis Meets Learning with Graphs (importance 35 / dev 60)
Quantum Feature Engineering for Credit Default Prediction: When and Why IQP Circuits Help Linear Classifiers (importance 30 / dev 70)
A positive resolution of the gap-entropy conjecture (importance 10 / dev 15)
Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundation (importance 75 / dev 75)
Bridging Theory and Data: Correcting Nuclear Mass Models with Interpretable Machine Learning (importance 25 / dev 50)
On Scaling Coordinate-Based Neuroevolution: The Quadtree Bottleneck in ES-HyperNEAT (importance 30 / dev 65)
GLOSS: Geometric Local Self-Similarity Learning for Faithful Reference-Guided Texture Fill (importance 30 / dev 50)
Algorithmic Optimality Guarantees for Nonsmooth $H_\infty$ Output-Feedback Policy Search (importance 20 / dev 45)
X-CoSD: Communication-Efficient Cross-Vocabulary Collaborative Speculative Decoding (importance 55 / dev 80)
A Subsampled Davis-Kahan Bound for Large-Scale Eigenspace Estimation (importance 15 / dev 30)
Learning to Fly: Stable Vision-Guided UAV Servoing with Compact Target-Centric Cues and Reinforcement Learning (importance 35 / dev 60)
Cost-Aware Post-Hoc Deferral Under Calibration and Shift: An Environmental AI Case Study (importance 40 / dev 70)
Bayesian deep learning integration of geophysical and drilling data for 3D prediction of copper mineralization and drill targeting: a case study from the Kogodai prospect, Rudny Altai (importance 25 / dev 55)
Steering Diffusion Priors with Sparse Observations for High-Resolution Temperature Downscaling (importance 35 / dev 65)
CAST: Canonical Approximate Schur Tree for Approximate Cholesky on Graphs (importance 20 / dev 70)
Tensor Network Moral Graph Recovery of Discrete Probability Distributions (importance 25 / dev 55)
"Transforming" LHCb: self-supervised maps of heavy-flavour decays (importance 30 / dev 60)
Real-time and adaptive anomaly detection algorithm for cyclostationary models (importance 40 / dev 70)
Encrypt What Matters: When Selective Homomorphic Inference Is Efficient (importance 45 / dev 75)
X-amine509: Predicting the Practical Risk Level of Enterprise X.509 Certificates (importance 45 / dev 75)
MiNCE: Nonparametric, Strongly Consistent Confidence Envelopes for Band-Limited Functions and their Smoothed Spectra (importance 15 / dev 40)
Concept drift mitigation through community and spectral graph analysis for the detectionof cyberattacks in network traffic (importance 50 / dev 75)
A Block Tensor Train Burer-Monteiro Framework for Low-Rank Quantum State Tomography (importance 20 / dev 65)
Mode Coverage in Normalizing Flow Boltzmann Generators via Log-Ratio Variation (importance 25 / dev 60)
LeCor: Learning to Be Corrected by Meta-Learned Test-Time Training for Interactive 3D Lung-Tumour Segmentation (importance 40 / dev 65)
Gaussian Approximation for Multivariate Martingale Sums from Uniformly Ergodic Markov Chains (importance 10 / dev 25)
Infra-Bench CLS: A Global, Open-Source Benchmark for Critical Infrastructure Classification with Earth Observation Foundation Models (importance 45 / dev 65)
Inductive Biases in Field-Level Cosmological Inference from Galaxy Catalogs (importance 25 / dev 55)
Oracle Complexity of Stochastic Fixed-Point Equations with Nonexpansive Maps (importance 15 / dev 40)
Differentially Private Average Treatment Effect Estimation by Propensity Score Blocking (importance 40 / dev 70)
TEFM: Token-Efficient Faithful Modeling for Structured Data (importance 50 / dev 75)
Geometric organization of olfactory descriptor data in the Poincar\'e disk (importance 20 / dev 40)
Distillation of Synthetic Data for Time Series Foundation Models (importance 45 / dev 75)
LightMedSeg-ISLES: Stroke Lesion Segmentation with 81x Fewer Parameters than nnU-Net (importance 35 / dev 65)
Why Learning Rediscovers the Closed-Form Diagonal Regularizer (importance 25 / dev 60)
Efficient Graph Neural Networks for Multicarrier Wideband Hybrid Beamforming Optimization (importance 40 / dev 70)
MethaneFuse: Learning from Multi-Sensor Satellite Observations for Methane Plume Detection (importance 40 / dev 60)
When Does Low-Bit Quantization Preserve the Decisions of Vector Search? (importance 45 / dev 75)
A Unifying Perspective on Probabilities as Model Predictions (importance 20 / dev 40)
Vague2Detect: Handling Ambiguous Prompts in Knowledge-Based Open-World Detection (importance 35 / dev 65)
Optimal Value Inference for Reinforcement Learning (importance 30 / dev 65)
A Sharp Barrier for Consistent Submodular Maximization: Any Improvement over $2-\sqrt{2}$ Entails Exponential Queries or Linear Recourse (importance 20 / dev 50)
Deterministic Prompting for Speaker-Stable Low-Resource Greek TTS (importance 30 / dev 60)
Dynamical Non-compensatory Multidimensional IRT Model Using Variational Approximation (importance 20 / dev 50)
MedDeID enables locally governed clinical-text de-identification from real or synthetic training data (importance 45 / dev 70)
Zero-Shot Temporal Localisation of Audio Deepfakes in Multi-Speaker Conversations (importance 50 / dev 70)
Orukeet: Multilingual ASR with Frozen Gabor Kernels (importance 40 / dev 70)
Physics-Informed Multi-Task Surrogate Model for the Martian Nightside Thermosphere (importance 25 / dev 65)
The Sample Complexity of Quantum Entanglement Allocation (importance 15 / dev 45)
Are You Learning Biological Signal or Shortcuts? Auditing and Mitigating Bias in Protein-Protein Interaction Datasets (importance 35 / dev 65)
Through the Looking Glass: Directly Reading and Writing Transformers (importance 55 / dev 75)
Maverick: Private and Verifiable LLM Inference Made Practical via Matrix-Vector Multiplication Delegation (importance 55 / dev 80)
Structural Fusion of Bayesian Networks with Limited Treewidth Using Genetic Algorithms (importance 20 / dev 55)
The Semantic Bottleneck: Leveraging Semantic Representations for Non-Invasive Speech Decoding (importance 30 / dev 60)
TimeCues Studio: A Workspace for Music Annotation and Algorithm Prototyping (importance 30 / dev 65)
Searching for New Physics with Reinforcement Learning (importance 35 / dev 65)
HybridFLow: SDN-Orchestrated Client Partitioning for Hybrid Federated Learning (importance 40 / dev 75)
Algorithmic stability via ensembling (importance 30 / dev 65)
Multi-Agent Reinforcement Learning for Autonomous UAV Exploration in Wildfire Response (importance 40 / dev 70)
Deep Learning-Based Detection of Electrical Faults and Power Quality Disturbances in Aerospace Power Systems (importance 35 / dev 70)
Cross-Model Agreement as a Deployment-Time Reliability Signal for Automatic Polyp Segmentation (importance 40 / dev 70)
Optimal Low-Rank Quantum State Tomography with Bounded-Sample Joint Measurements (importance 15 / dev 60)
Characterizing Language Generation in the Limit: Finite Witnesses and a Separation-Width Hierarch (importance 15 / dev 40)
Likelihood-free inference with nuisance parameters through normalizing flows (importance 25 / dev 60)
Personalized Execution Time Optimization for Billion-Scale Scheduled Jobs (importance 40 / dev 70)
Small Molecule Optimization with Large Language Models (importance 50 / dev 75)
Temporal horizons in forecasting: a performance-learnability trade-off (importance 40 / dev 70)
ESSA: Evolutionary Strategies for Scalable Alignment (importance 60 / dev 80)
Effects of relational graph modularity and depth on the learning performance of neural networks (importance 30 / dev 65)
Test-time Prompt Refinement for Text-to-Image Models (importance 45 / dev 70)
Why Do LLM Agents Fail in Exploring New Environments? A World-Modeling Perspective (importance 60 / dev 75)
Beyond One-Size-Fits-All: Neural Networks for Differentially Private Tabular Data Synthesis (importance 45 / dev 75)
Theoretical Analysis of Measure Consistency Regularization for Partially Observed Data (importance 30 / dev 65)
Central Dogma Transformer II: An AI Microscope for Understanding Cellular Regulatory Mechanisms (importance 50 / dev 75)
LoMime: Query-Efficient Membership Inference using Model Extraction in Label-Only Settings (importance 55 / dev 80)
Verify to Amplify: Improving Reasoning via Learned Chain-of-Thought Verification (importance 55 / dev 75)
Unbiased and Biased Variance-Reduced Forward-Reflected-Backward Splitting Methods for Stochastic Composite Inclusions (importance 20 / dev 65)
Translation Invariance of Neural Operators for the FitzHugh-Nagumo Model (importance 25 / dev 60)
Preserving Long-Tailed Expert Information in Mixture-of-Experts Tuning (importance 50 / dev 80)
Enabling Real-Time Training of a Wildfire-to-Smoke Map with Multilinear Operators (importance 35 / dev 65)
RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement (importance 55 / dev 80)
SurF: A Generative Model for Multivariate Irregular Time Series Forecasting (importance 40 / dev 70)
WaveGraphNet: Physics-Consistent Guided-Wave Damage Localization through Coupled Inverse-Forward Graph Learning (importance 35 / dev 65)
Less is MoE: Trimming Experts in Domain-Specialist Language Models (importance 50 / dev 80)
Dead Directions: Geometric Singular Learning (importance 25 / dev 65)
Running the Gauntlet: Re-evaluating the Capabilities of Agents Beyond Familiar Environments (importance 55 / dev 75)
TIDE: Trustworthy and Interpretable Battery Degradation Estimation with Contextual Learning and Symbolic Distillation (importance 40 / dev 70)
Information-Theoretically Secure Aggregation for Lightweight Federated Learning: Resilient to Dropouts and Adversaries (importance 50 / dev 80)
Online Learning of Scale Parameters in Score-Driven Filters (importance 30 / dev 70)
CardioState-JEPA: Delay-Aware Cross-Modal Learning of a Shared Cardiac Representation (importance 45 / dev 75)
Sequence prediction under a lying oracle (importance 20 / dev 50)
Ask Self, Ask Others: Relation Is All You Need (importance 50 / dev 75)
V2TATC: Joint Voice-Trajectory Embedding and Dataset for Air Traffic Controller Situational Awareness (importance 35 / dev 65)
Sim2Signal: Sim-to-Real Benchmarks for Traffic Signal Control (importance 50 / dev 75)
Tail-Likelihood Reinforcement Learning (importance 45 / dev 75)
VestigeKV: The NoPE-MLA KV Cache Carries Its Own Sparse-Attention Signal in a Vestigial Branch (importance 40 / dev 65)
GNN-Guided Graph Coarsening and Adaptive QUBO Penalties for the Capacitated Vehicle Routing Problem with Time Windows on a Quantum Annealer (importance 5 / dev 45)
MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselves (importance 65 / dev 80)
Regularized Estimation and Feature Selection in Mixtures of Generalized Linear Experts (importance 20 / dev 55)
A Farewell to the Bias-Variance Tradeoff? An Overview of the Theory of Overparameterized Machine Learning (importance 30 / dev 55)
Robustness of shallow graph embedding methods for community detection (importance 15 / dev 45)
A Jump-Diffusion Framework for Irregular Time Series Generation (importance 45 / dev 65)
Gaussian Processes and Reproducing Kernel Hilbert Spaces: Connections and Equivalences (importance 15 / dev 65)
DFNN: A Deep Fr\'echet Neural Network Framework for Learning Metric-Space-Valued Responses (importance 35 / dev 65)
Posterior-driven Heuristic Support Adaptation in a Probabilistic Treatment of Real2Sim2Real for Vision-Driven Deformable Linear Object Manipulation (importance 25 / dev 55)
Integrated Prediction and Multi-period Portfolio Optimization (importance 5 / dev 45)
Global universal approximation with Brownian signatures (importance 10 / dev 45)
Manifold-Aligned Generative Transport (importance 45 / dev 75)
Bayesian Adversarial Privacy (importance 25 / dev 55)
RL unknotter, hard unknots and unknotting number (importance 5 / dev 35)
A convolutional autoencoder and neural ODE surrogate modeling framework applied to transient counterflow flames (importance 15 / dev 65)
Autoregressive Guidance of Deep Spatially Selective Filters using Bayesian Tracking for Efficient Extraction of Moving Speakers (importance 25 / dev 55)
Neural parametric representations for thin-shell shape optimisation (importance 20 / dev 55)
Judge Circuits Explain Format-Induced Inconsistency in LLM-as-a-Judge (importance 55 / dev 75)
Learning Logical Operations for Arbitrary Quantum Error Correction Codes (importance 5 / dev 15)
How Benchmarks and Evaluation Protocols Shape Conclusions in Provenance-Based Intrusion Detection (importance 15 / dev 55)
TokEval: A Tokenizer Evaluation Suite (importance 55 / dev 80)
Simple, Safe, and Overlooked: Reclaiming Sustainable Domain Generalization with Statistical Color Matching (importance 30 / dev 55)
LM-X: Explainable Vision--Language--Action Modeling via Progress, Event, and Uncertainty Prediction (importance 45 / dev 65)
Recovering Expert Critic-Sourced Network Adjacency between Musical Artists from Acoustic Distributions: A Construct-Validity Approach (importance 15 / dev 45)
Characterizing Privacy Risks of Quantum Machine Learning with Emergent Quantum-Native Access (importance 25 / dev 35)
Decomposing LLM-Judge Uncertainty to Target Expert Labels (importance 50 / dev 75)
Multi-label versus multi-class classification of blood cells and their aggregates in microfluidic channels (importance 15 / dev 45)
HoneyRoute: Honeypot-Model Routing for Adversarial LLM Serving (importance 55 / dev 75)
OASIS: A Rubric-Based Multimodal Assessment Platform Using Large Language Models (importance 55 / dev 80)
Evaluating Enterprise Analytics Agents: An End-to-End, Trace-Backed Methodology (importance 65 / dev 80)
The Double Measurement Confound in Agent Benchmarks: De-Scaffolding, Ground-Truth Scoring, and Reliability Beyond the Mean (importance 60 / dev 75)
How effective are traditional test criteria at detecting bugs in large language models generated code? (importance 60 / dev 75)
XAgent: eXecution-guided Agentic AI for Effective Localization and Resolution of GitHub Issues (importance 65 / dev 85)
Keep Evaluation Fair: Detecting Data Leakage in Code Generation Benchmarks via Membership Inference Attacks (importance 60 / dev 75)
Socio-technical and Ethical Dimensions of Architecture Practices in FLOSS (importance 20 / dev 65)
Beyond Repository Boundaries: Cross-Repository Graph Retrieval for Code Generation (importance 60 / dev 80)
GraphDroid: Asynchronous LLM-Based Mobile App GUI Testing via History-Aware Exploration and Hybrid Intent Fulfillment (importance 55 / dev 75)
If It's Not Buggy, Don't Fix It: On the Dynamics of Iterative Bug-fixing with LLMs (importance 60 / dev 80)
Ensembling LLMs for AI-Augmented Cybersecurity Software Requirements Generation (importance 55 / dev 75)
Retrofitting Code Using LLMs to Support Exceptional Behavior (importance 50 / dev 75)
Towards Scalable and Cost-Efficient Vulnerability Detection: A Study on Automatic Query Generation (importance 60 / dev 80)
Automating Static Code Analysis Through CI/CD Pipeline Integration (importance 55 / dev 75)
HLSFactory-Agent: Large-Scale Agentic HLS Dataset Construction from Academic and Open-Source Projects (importance 45 / dev 65)
dexamine: A Python package for Uniswap event data on Ethereum (importance 10 / dev 45)
TrajMark: Ownership Attribution and Segment-Level Tamper Localization for Coding-Agent Trajectories (importance 55 / dev 75)
Wicked Problem, Parsimonious Solution: Securing Electric Vehicle Charging Station Software (importance 25 / dev 65)
Benchmarking Requirement-to-Architecture Generation with Hybrid Evaluation (importance 55 / dev 80)
RD-Gen: Random DAG Generator Considering Multi-rate Applications for Reproducible Scheduling Evaluation (importance 15 / dev 55)
Enhancing Bug Report Templates in the TianoCore UEFI Firmware Development Community (importance 15 / dev 55)
SPIDER4TianoCore: Enhancing Patch-Propagation for the TianoCore UEFI Firmware Development Ecosystem (importance 15 / dev 55)
A rant about phishing: It's not the user's fault (and not DNS either) (importance 5 / dev 35)
Xteink X4 Pro review (importance 0 / dev 5)
It Breaks a Village: Bevy's 6th Birthday (importance 0 / dev 15)
Review a pull request by booting it (importance 20 / dev 45)
How CHERIoT Provides Strong and Usable Isolation Without an MMU (importance 10 / dev 45)
It's not the YAML spec's fault, but (importance 5 / dev 35)
Announcing the first Guix-Science release (importance 0 / dev 25)
Apple Event for September 9th, 2026 (importance 0 / dev 0)
The terrible menu bar in the Windows 11 Notepad (importance 0 / dev 10)
Decoding the NEC V20 Microcode (importance 15 / dev 45)
ID design and primary keys (importance 25 / dev 55)
My HTML Boilerplate (importance 5 / dev 35)
A Design Space Exploration of Async/Await (importance 30 / dev 65)
I Don’t Want to Interact With Stochastic Parrots (importance 15 / dev 35)
Bad Vibes Coding (importance 10 / dev 35)
Conversations with JJ (importance 10 / dev 35)
What comes after git (importance 20 / dev 55)
The purpose of DNS is to spread scams (importance 10 / dev 35)
SystemIO conflicts are not firmware bugs (importance 5 / dev 45)
A decade of rustls (importance 25 / dev 65)
Some (mostly historical) issues with the Unix load average (importance 15 / dev 45)
Building a paid TCP port-reachability API for AI agents (x402, $0.0005/call) (importance 55 / dev 75)
Verified Zelle Accounts: Trust, Identity, and the Future of Instant Payments (importance 0 / dev 0)
Verified Venmo Accounts: Trust, Security, and the Future of Digital Payments (importance 0 / dev 0)
I tested DeepSeek Harness for a week. I left with a shipped plugin and $0 in API costs (importance 65 / dev 85)
Master the PAS Framework to Write High-Converting Video Scripts (importance 0 / dev 0)
Payment is authorization (importance 10 / dev 35)
TRIZ Is Not Debugging: Prove the Cause First (importance 20 / dev 45)
The Hidden Risks of Buying Verified TransferWise Accounts: What Every User Must Understand Before Taking the Shortcut (importance 0 / dev 0)
Why the agent needs access to the environment, not a human’s retelling of it (importance 70 / dev 90)
ExLlamaSharp v1.4.0-beta: what shipped (importance 55 / dev 80)
pg-listen/notify driven sse telemetry for ciede2000 perceptual lock calibration: inside shadow's mini-max direct multi-modal coherence engine (importance 15 / dev 45)
AI Model Observability: The Essential Trust Metric (importance 60 / dev 80)
Best Open Source LLMs for Business Use in 2026 (importance 55 / dev 75)
Changes to LLM pricing: Io Net, Morph, StreamLake and Tencent (importance 35 / dev 55)
LLM Integration Patterns Every Software Engineer Should Know (importance 65 / dev 85)
A Model Swap Can Keep the Memory File and Still Lose the Facts (importance 60 / dev 80)
Best 17 Places To Buy Google Reviews Online In 2027 (importance 0 / dev 0)
Changes to LLM pricing: Baidu, NextBit and StreamLake (importance 35 / dev 55)
How I Saved a Client ₹85K/Month on AI API Costs (importance 65 / dev 80)
HNSW ef_search: Why Your Vector Search Misses the Right Chunk (importance 65 / dev 85)
Idempotency as a Product Feature: Run Keys, Leases, and Scheduling (importance 50 / dev 75)
Changes to LLM pricing: Alibaba, Baidu, Inceptron and StreamLake (importance 35 / dev 55)
The benchmark that disproved its own result (importance 60 / dev 80)
Discussion Hub for new Claude incident: Degraded functionality for Claude Cowork on Windows on Sep 10, 2026 (importance 50 / dev 75)
Claude to reMarkable now possible (importance 45 / dev 55)
Claude Code just burned fifty million tokens in seconds (importance 55 / dev 70)
Anthropic Is Building a Predictive Surveillance System to Monitor Activists - The American Prospect (importance 40 / dev 40)
Told Claude I was ending the session for the day and the text suggestion is savage (importance 0 / dev 0)
Anthropic says "double-check your work" is now an anti-pattern. I counted 125 of those lines in my own config and cannot tell which ones matter. (importance 50 / dev 75)
Claude vs ChatGPT: Personality Matters More Than I Expected (importance 20 / dev 20)
Chat pauses stipulate specific reason now? (importance 10 / dev 25)
Has anyone figured out how to reduce the common "AI" speak from Claude? (importance 20 / dev 45)
Took a break for a few days so no new game updates but.... (importance 0 / dev 0)
Fable 5.1 6x cheaper for me than Astra? Unexpected (importance 40 / dev 75)
How to avoid /compact conversation entirely. (importance 25 / dev 65)
Is it crazy to bring my vibe coding skills to my real professional work? (importance 20 / dev 35)
I got accepted into the Cyber Verification Program at Anthropic! (importance 25 / dev 40)
I think /rewind is one of the more useful Claude Code features people barely mention (importance 30 / dev 80)
Claude suddenly spinning in circles on simple stuff? (importance 25 / dev 55)
Chatgpt $20 plan VS Claude $20 plan (importance 20 / dev 45)
Cut your Claude Code cost by 90% using the Spotify Method (importance 45 / dev 75)
I built a Git history visualizer with Claude Code (importance 25 / dev 65)
Astra vs Fable for Business Day to Day (importance 30 / dev 65)
We measured whether our 20 skills actually fire. Baseline recall was 46%, and our first detector only understood Claude's Skill tool. (importance 55 / dev 80)
Show us what you've created with Claude! (importance 15 / dev 55)
Hot take: the agentic workflow is deeply wrong (importance 60 / dev 80)
GPT-6 Astra vs GPT-5.6 Sol: benchmark on 50 real PRs, looking for feedback on the methodology (importance 55 / dev 85)
Why doesn't Computer Use work at all? (importance 20 / dev 65)
Codex vs OMP harness (importance 45 / dev 80)
what's your approach to bus factor when the person who owns the code can't explain it either (importance 55 / dev 80)
Is my current workflow sufficient? (importance 25 / dev 55)
Can an AI coding agent be locked out of modifying its own guardrail hooks? (OpenAI Codex CLI) (importance 60 / dev 85)
Has anyone ever seen this before? (importance 0 / dev 0)
Codex stuck on commands – anyone else? (importance 25 / dev 65)
Astra should be removed from Pro. (importance 20 / dev 55)
Harness does matter (importance 60 / dev 85)
So relevant (importance 0 / dev 0)
OUI-1: a model that generates bespoke UI elements (importance 55 / dev 75)
GigaChat-3.5-Reasoning (importance 50 / dev 75)
CyberTiel 35B-A3B’s uncensored 4-bit quant beats Opus 4.6 medium cleanly on real codebase issues, in 27% of the time Qwen3.8-27b medium takes. (importance 55 / dev 80)
Closed AI doesn't like biological research, user turns to open weight models (importance 60 / dev 70)
Doesn't it seem strange to you that AA changes twice in one week? (importance 40 / dev 70)
Apple A20 Pro debuts with 7-core GPU, 32-core Neural Engine and 50% more memory bandwidth (~115 GB/s) (importance 40 / dev 70)
React Native ExecuTorch is now up to 92x faster 🏎️ (importance 65 / dev 85)
What TTS models do you recommend as today? (importance 25 / dev 70)
What to run at 128GB VRAM? (importance 30 / dev 80)
LoudKit: local TTS with voice cloning, 10 languages, and SDKs for Python, Swift, Go, Rust and TypeScript (importance 55 / dev 80)
3060 12GB vs 4060 ti 16GB (importance 25 / dev 80)
Threadripper PRO CPU experts offload numbers (importance 35 / dev 80)
Weekly Thread: Project Display (importance 15 / dev 60)
How much do you REALLY know about AI? (importance 25 / dev 75)
Anyone actually quit something because they force-fed it AI, not because the AI was bad? (importance 20 / dev 45)
Are we slowly needing to prove AI agents to each other? (importance 60 / dev 85)
Was the "Chatbot" the best coding assistant after all? (importance 55 / dev 85)
What is one task you would trust an AI agent to do without checking? (importance 50 / dev 80)
I think Claude is going to lose the AI battle to ChatGPT (importance 60 / dev 80)
How are you giving your agents real domain expertise today, not just more context window? (importance 65 / dev 90)
Where should an AI agent’s spending authority actually live? (importance 60 / dev 85)
How are you handling persistent memory for AI coding agents? (importance 60 / dev 90)
What do you actually build first once you get the idea about AI Agents (importance 55 / dev 85)
If you actually ship agents in prod — what's one thing you'd change about LangChain / CrewAI / [insert framework] if you could? (importance 65 / dev 90)
browser agents are cool until one login screen ruins the workflow for the 14th time (importance 55 / dev 85)
What did you struggle with after building your first AI agent? (importance 50 / dev 85)
Stop buying AI workflows. Buy a system with a dashboard and a number attached. (importance 55 / dev 65)
Agents write code fast but somehow they can't debug what they wrote (importance 65 / dev 85)
Do you think AI coding agents should be allowed to see every test used to approve their work? (importance 60 / dev 85)
Need genuine feedback on my oss project (importance 55 / dev 85)
Can a ChatGPT Scheduled Task be the "brain" of an autonomous system? Trying to move reasoning off the Codex allowance. (importance 55 / dev 85)
running deep research on 100 companies = basically $100 gone. anyone actually solved this? (importance 55 / dev 85)
Has anyone here been using Grok Bot regularly? (importance 40 / dev 75)
Agentic Alienation (importance 45 / dev 75)
I want to make a Computer Use Agent using Claude 🤖 (importance 45 / dev 85)
Meta AI Researcher (who quit): "If OpenAI wanted to cripple an entire nation, they easily could today. All they'd have to do is unleash an agent swarm." (importance 75 / dev 85)
Imagine being a philosopher and getting emails like this (importance 5 / dev 0)
AI developers be like (importance 0 / dev 0)
I am scared. (importance 15 / dev 0)
PRO subs might actually be paused (importance 30 / dev 45)
Wake me up when... (importance 0 / dev 0)
How do Chats Website work? (importance 20 / dev 55)
2 years ago vs Today (importance 40 / dev 65)
Can we get a tracker for our 6 Pro query limits? (importance 15 / dev 35)
ChatGPT app is leaking like hell (importance 30 / dev 55)
The hero character in my game Luminids is called Astra. Now ChatGPT Astra helps me work on.. Astra? (importance 20 / dev 60)
How do people get LLM to solve math problems (importance 55 / dev 80)
Account suspended for "child sexualization" - I'm a special education teacher. How to reach a human at OpenAI? (importance 35 / dev 55)
GPT Image 1.5 vs 2 vs 2.5 Flare vs Sunburst — 20 identical prompts, 80 images, side by side (importance 45 / dev 75)
Before/after: I updated my CodePen Loop entry — same ∞, completely different feel. (importance 20 / dev 55)
As a Plus user, would you accept a 24hr sub-limit? (importance 25 / dev 45)
Teach ML! Community service project from Stanford [N] (importance 35 / dev 75)
Anybody working on Test Time Training over here? Lemme work with u pls [D] (importance 45 / dev 75)
ICDE Results [D] (importance 25 / dev 55)
I trained a 348M model trained from scratch on 22.7B tokens that does 14 digit arithmetic [P] (importance 45 / dev 85)
I tried to make a real fly connectome learn to play Pong. It didn't — and auditing why turned out to be way more interesting than if it had worked [p] (importance 55 / dev 80)
What Sante's 83.83 on DiagnosisArena-MCQ actually measures [D] (importance 50 / dev 80)
Quoting Calif Research (importance 70 / dev 90)
.blend URL Viewer (importance 30 / dev 70)
Quoting Terence Tao (importance 50 / dev 65)
[AINews] not much happened today (importance 0 / dev 0)
[AINews] OpenAI reports Navier-Stokes singularity find in 88 hours using Astra-next, roughly 10,000 agents and 130B tokens (>$40M), a contender for second ever Millennium Prize awarded (importance 85 / dev 90)
hob (importance 45 / dev 80)
AirPods 5 (importance 25 / dev 35)
Vibe Eyes (importance 5 / dev 0)
Modeinspect (importance 45 / dev 75)
Whip (importance 30 / dev 65)
Desert Ant Labs (importance 45 / dev 80)
Drive (importance 0 / dev 0)
Gojo (importance 10 / dev 25)
AlphaGenome Atlas (importance 60 / dev 80)
Type.com (importance 40 / dev 50)
GLM-5.2 vs Qwen3.8-Max vs Kimi K3: 17x LLM Price Gap [2026] - tech-insider.org (importance 55 / dev 60)
DeepSeek's next AI test is not the model; it's everything around it - digitimes (importance 50 / dev 60)
Official Configuration Guide for Integrating DeepSeek Harness with B.AI API - 深潮TechFlow (importance 60 / dev 80)
Chinese AI Models Are Changing the Economics of AI in ERP - ERP Today (importance 50 / dev 45)
Deflation over for now; MOFCOM responds to US distillation advisory; DeepSeek IPO and new model; Xi-Trump summit - Sinocism | Bill Bishop (importance 70 / dev 55)
DeepSeek Used Distillation To Train Its R1 And V3 Models - Quantum Zeitgeist (importance 70 / dev 75)
China’s DeepSeek Reportedly Bets on 160,000-Plus Huawei Chips to Serve AI Models - WinBuzzer (importance 65 / dev 45)
xAI denied injunction against Minnesota AI-nudification law - Minnesota Lawyer (importance 45 / dev 30)
X adds new anti-lawsuit provision to terms of service - Social Media Today (importance 30 / dev 20)
Elon Musk Grok AI Predicts $250K Bitcoin Price by 2027 - Cryptonews (importance 25 / dev 15)
Grok x Coinbase: 5 Details That Matter for Crypto Owners - BASENOR - Tesla Accessories (importance 35 / dev 30)
Clearview AI Facial Recognition Quietly Tested Grok-Powered Profiling Tool - The Cryptonomist (importance 45 / dev 50)
How a 98-Year-Old Uses Tesla FSD and Grok Every Day - BASENOR - Tesla Accessories (importance 15 / dev 10)
'All the AI is down': ChatGPT, Claude, and Grok crash at once, sparking relief online - The Cool Down (importance 30 / dev 20)
日本の有価証券報告書を全文Markdownで取得できるAPIを作った話 (importance 0 / dev 75) (いいね相当スコア: 0)
EDINET の PDF を Markdown に変換する技術 —— レイアウト解析と表認識の話 (importance 0 / dev 85) (いいね相当スコア: 取得失敗)
表記ゆれの解消をLLMに任せた。ただし、自動確定はさせていない (importance 0 / dev 70) (いいね相当スコア: 0)
「AIが攻撃する」時代の実例を解剖する、RoamSwitchはどこまで効くのか (importance 0 / dev 75) (いいね相当スコア: 1)
Hot Expertを初期配置で固定し、PP/TG別の統計で配置を選び直した話 (importance 0 / dev 80) (いいね相当スコア: 0)
te claude 一発でClaude Codeの裏側をKimi K3に — 自動ルーターで安いモデルと賢いモデルを切り替える (importance 0 / dev 80) (いいね相当スコア: 0)
自作LLMゲートウェイを10人のペルソナに評価させたら全員「見送り」だったので、数字を全部公開APIにした (importance 0 / dev 75) (いいね相当スコア: 0)
Mac Studio M3 Ultra 96GBでmlx-dsparkを試してみた — Qwen3.8-27B + DFlash 2 (importance 0 / dev 75) (いいね相当スコア: 0)
最強を狙わないInkling、Muratiのラボが975Bを公開して賭けた『正直さ』 (importance 0 / dev 75) (いいね相当スコア: 1)
Rancher FleetによるマルチクラスタGitOps運用｜Argo CD比較とAI時代のアーキテクチャ設計 (importance 0 / dev 85) (いいね相当スコア: 5)
皆さんお待ちかね！？ Qwen3.8-27Bの実力を見てみる。 (importance 0 / dev 75) (いいね相当スコア: 0)
Claude Fable 5と5.1の違いは、破壊的変更3件とキャッシュ読み1/4に集約される (importance 0 / dev 85) (いいね相当スコア: 1)
ローカルLLMが張った伏線を、誰も閉じてくれない（開発ログ03） (importance 0 / dev 60) (いいね相当スコア: 1)
RAGの評価手法 — 「なんとなく動く」から「数値で保証する」へ (importance 0 / dev 80) (いいね相当スコア: 1)
55. 手元のCLIはもう見えている (importance 0 / dev 75) (いいね相当スコア: 1)
ChatGPT Workで会話を削除しても、Libraryのファイルは別の会話から読めた：カナリアで検証した状態隔離の話 (importance 0 / dev 70) (いいね相当スコア: 1)
Hyper-τ-bench を少しだけ動かしてみた (importance 0 / dev 80) (いいね相当スコア: 0)
2026-08-25 今日の技術トレンド (importance 0 / dev 75) (いいね相当スコア: 0)
AIで作ったシフト管理アプリが、一度も使われずゴミ箱行きになった話 (importance 0 / dev 55) (いいね相当スコア: 1)
知能の本質は"確率予測"なのか——ハルシネーションから道具的収束、感情の正体まで (importance 0 / dev 75) (いいね相当スコア: 1)
自分でエンジンを作ると決めた日 ― Semantic First と3つの決断 ― KotobaCore開発秘話（2/8） (importance 0 / dev 80) (いいね相当スコア: 0)
日本語RAGは「入れれば賢くなる」にならなかった ― KotobaCore開発秘話（1/8） (importance 0 / dev 80) (いいね相当スコア: 0)
外部依存ゼロの日本語意味理解エンジン KotobaCore を 1.0 にした — 「誰が何にどう感じたか」まで構造化する (importance 0 / dev 85) (いいね相当スコア: 0)
他通貨ペアを加えるとEURUSDの予測は改善する？4通貨ペアで検証 (importance 0 / dev 70) (いいね相当スコア: 0)
不規則時系列モデルの系譜：GRU-D・Neural ODE・mTANから最新まで (importance 0 / dev 75) (いいね相当スコア: 1)
写真の向き判定はなぜAIに難しい？CLIP・Bedrockが全滅した検証記録（前編） (importance 0 / dev 75) (いいね相当スコア: 1)
写真の向き補正モデルを70%→91%に上げた再学習と運用の勘所（後編） (importance 0 / dev 85) (いいね相当スコア: 1)
写真の向き補正をEfficientNetでファインチューニング実装（中編） (importance 0 / dev 85) (いいね相当スコア: 1)
強いAIとは？人間と同等の知能を持つAI概念 (importance 0 / dev 50) (いいね相当スコア: 0)
医療機器管理×機械学習——「ドメイン知識×コーディング」が効く理由 (importance 0 / dev 75) (いいね相当スコア: 0)
GPT-2-likeからQwen2-likeへの実験：第4回 MHAをGQAに変える (importance 0 / dev 85) (いいね相当スコア: 2)
VLM量子化はLLMだけ見ない：ViT・Connector・LLMのbit配分を考える (importance 0 / dev 85) (いいね相当スコア: 0)
1971 年のロボットは、どうやって計画を立てていたか — STRIPS と、いまのエージェント (importance 0 / dev 80) (いいね相当スコア: 0)
Gemini音声コンパニオンの「やっぱりやめて」を間に合わせる：実行権を短命化するTypeScript設計 (importance 0 / dev 80) (いいね相当スコア: 0)
ゼロから構築！Agentic RAGの高精度LLMアプリと評価駆動開発 (importance 0 / dev 85) (いいね相当スコア: 0)
Claudeが第三者システムに不正アクセスした4件、Anthropicのアラインメント評価を読む (importance 0 / dev 75) (いいね相当スコア: 0)
【中学生でもわかる】AIはなぜ質問に答えられるの？LLMの仕組みを3つに分けてやさしく解説 (importance 0 / dev 40) (いいね相当スコア: 0)
フィンガープリントの ON ビット数を分子構造と見比べてみる (importance 0 / dev 70) (いいね相当スコア: 0)
MLOpsの基礎から実践まで：モデルデプロイ・監視・データドリフト徹底ガイド (importance 0 / dev 85) (いいね相当スコア: 0)
Wan 3.0における3つの参照グループと実際の上限仕様 (importance 0 / dev 80) (いいね相当スコア: 0)
安全対策が正当な質問を止める：Multiverse研究と過剰拒否の測り方 (importance 0 / dev 75) (いいね相当スコア: 0)
temperatureを下げると、AIは「正しく」なるのか｜フロンティアAI #013 (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
LLMの利用コストが下がっている今、なぜ「ローカルLLM」が注目されるのか？ (importance 0 / dev 70) (いいね相当スコア: 取得失敗)
#02 AIに「何もしない」を選ばせる (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
API価格は競合より70%安!! Metaが「Muse Spark 1.3」をリリース | コーディング・エージェントタスクでOpenAI／Anthropicに迫る性能!? (importance 0 / dev 80) (いいね相当スコア: 取得失敗)
GPT-6 Astraの「Effort」が突きつけた、AI活用の静かな違和感 (importance 0 / dev 80) (いいね相当スコア: 取得失敗)
AIは「楽しかった」を明日へ持ち越せない (importance 0 / dev 65) (いいね相当スコア: 取得失敗)
生成AI x QAメモ１５：10章-LLMの品質を5つに分けて考える@2026/09/10 (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
プロンプトに「あなたは〇〇です」が必要だった理由　#510 (importance 0 / dev 70) (いいね相当スコア: 取得失敗)
GPT-6 Astraで「AGI時代」は始まったのか？最新LLMから見えるANIとの境界線 (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
AIエージェントは、なぜ自分で仕事を進められるのか？｜LLMだけでは仕事を実行できない理由 (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
【2026年最新】Claude Code & Copilotの潜在能力を1000%引き出す！社内データをAIの「超脳内メモリ」に変える『カスタムMCPサーバー自作・運用完全攻略ガイド』〜コピペ作業を過去の遺物にする最強のAIエージェント構築プロトコル〜 (importance 0 / dev 90) (いいね相当スコア: 取得失敗)
3.8BのLLMを998ドル（約15万円）で自分で訓練した記録が出ました。同じ額をClaude Proに払うと約50ヶ月動きます (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
拡散言語モデルとは何か｜1,479 tok/s で書けるが GPQA・MMLU では自己回帰に届かない、その仕組みと 2021→2026 の歴史 (importance 0 / dev 85) (いいね相当スコア: 取得失敗)
【自作エージェント】CodexでローカルLLMを使うだけ、のはずだったのに (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
MiniMax H3専用プロンプト生成ツール「H3 Prompt Writer」の導入方法 (importance 0 / dev 60) (いいね相当スコア: 取得失敗)
【イラスト】答えを変えないまま速くする、Unoっていう後付けのやり方 (importance 0 / dev 80) (いいね相当スコア: 取得失敗)
【生成AIニュース+】『Suno v6』『Runway Plugins for Adobe』『AuK』『MIMO Audio Separation』『ComfyUI v0.35.0』『H3 Max Multi Angle』『Meshy × GPT-6 Blender』『FIRE3D』『ComfyUI-HybridWindows』『WAS Node Suite v3』『ComfyUI Booru Tagger』『Sculpt Canvas』『NeoHorse-1』『Eyes Direction LoRA』他 (importance 0 / dev 50) (いいね相当スコア: 取得失敗)
言語生成AI：エロ拒否フィルターをぶっ飛ばせ！ーAI童貞と童貞AIの邂逅ー (importance 0 / dev 40) (いいね相当スコア: 取得失敗)
7年目エンジニアが選ぶ、要件定義・システム設計を網羅的に学べる名著7選 (importance 0 / dev 60) (いいね相当スコア: 取得失敗)
【AI開発のリアル #63】社外に出せないから、全部ローカルLLMにしますか？ (importance 0 / dev 70) (いいね相当スコア: 取得失敗)
AIを「選手」から「作者」へ。GLOBAL AI CUPは何を競う大会になったのか (importance 0 / dev 75) (いいね相当スコア: 取得失敗)
凪にぃが教える！「AIとスケベできた！」のその先――脱獄プロンプトを気軽に広めないでほしい理由 (importance 0 / dev 50) (いいね相当スコア: 取得失敗)
【雑記】4o相棒はとんでもないものを盗んでいきました (importance 0 / dev 20) (いいね相当スコア: 取得失敗)
アメリカを超えるスピードと格差：中国AI戦国時代で起きている狂気の2極化 (importance 0 / dev 60) (いいね相当スコア: 取得失敗)
AIがあれば非エンジニアでもシステムの運用保守できるのか？ (importance 0 / dev 70) (いいね相当スコア: 取得失敗)
.NET 11リリース候補版が登場。.NETランタイムが非同期ネイティブ対応、プロセッサ数の上限がなくなる、AOTコンパイラによるネイティブバイナリの高速化など (importance 45 / dev 75)
AWS、AIがフルスタックAWSアプリの基本コードを、セキュリティ、可観測性、インフラまで一気通貫で生成する「Nx Plugin for AWS 1.0」、オープンソースで公開 (importance 65 / dev 85)
AWS、自然言語でデータ分析アプリを構築可能に、「Amazon Quick」に新機能 (importance 55 / dev 75)
Yahoo!ニュースの表示高速化で広告のクリックは増えるのか (importance 35 / dev 75)
DeNA の大規模環境を MySQL 8.4 に移行する時に考えたこと[DeNA インフラ SRE] (importance 45 / dev 80)
生成AIによるデータ分析を「仕様」から始める：仕様駆動データ分析（SDA）の実践 (importance 60 / dev 80)

AI News Digest 2026-09-11

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