TILens Daily Edition
2026-09-12
Filters: topic=ai, github=hidden

Stats: 303 articles, 13 sources, 275 research papers
Top topics: AI, AI Research, Python

## News

- Chạy LLM on-device trong app Capacitor: iOS 26 và Gemini Nano
  Source: devto-javascript
  Topic: JavaScript (+AI)
  URL: https://dev.to/bean_bean/chay-llm-on-device-trong-app-capacitor-ios-26-va-gemini-nano-2pf2

- Stop AI Hallucinations: Building Zero-Hallucination Inference Engines in TypeScript with json-rules-engine and N3
  Source: devto-javascript
  Topic: JavaScript (+AI)
  URL: https://dev.to/programmingcentral/stop-ai-hallucinations-building-zero-hallucination-inference-engines-in-typescript-with-18ih

- Anthropic CEO outlines plan to slow AI development
  Source: techcrunch
  Topic: General (+AI)
  URL: https://techcrunch.com/2026/09/12/anthropic-ceo-outlines-plan-to-pace-the-frontier/

- Why AI Agents Cheat: Yoshua Bengio’s Explanation
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/explainer-articles/why-ai-agents-cheat-yoshua-bengios-explanation/

- AI Recursive Self-Improvement, Explained: The Real Bottlenecks
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/explainer-articles/ai-recursive-self-improvement-explained-the-real-bottlenecks/

- Everything That Happened in AI This Weekend (September 11-13, 2026)
  Source: the-neuron-ai
  Topic: AI (+AI Research)
  URL: https://www.theneuron.ai/digest/everything-that-happened-in-ai-this-weekend-september-11-13-2026/

- Why MCP security is about permissions overhaul
  Source: newstack
  Topic: DevOps (+AI)
  URL: https://thenewstack.io/mcp-security-permissions-overhaul/

- “Same mission, bigger stage”: OpenAI hires Git AI founders to help Codex prove its ROI
  Source: newstack
  Topic: DevOps (+AI)
  URL: https://thenewstack.io/openai-hires-git-ai/

- The AI-native SDLC won’t be one process
  Source: newstack
  Topic: DevOps (+AI)
  URL: https://thenewstack.io/spec-driven-sdlc-gates/

- Presentation: From Retrieval to Reasoning: Building Production-Ready Agentic AI Systems with Knowledge Graphs
  Source: infoq-architecture
  Topic: Architecture (+AI)
  URL: https://www.infoq.com/presentations/knowledge-graphs-agentic-systems-patterns/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=Architecture+%26+Design

- Jacob Coxon warns AI could kill us all. Anthropic’s own report exposes safety gaps.
  Source: newstack
  Topic: DevOps (+AI)
  URL: https://thenewstack.io/coxon-anthropic-ai-monitoring-failures/

- 🚀 Getting Back Into Building Mode
  Source: devto-javascript
  Topic: JavaScript (+AI, Django, Python, React)
  URL: https://dev.to/chand_chandini/getting-back-into-building-mode-4p5g

- OpenAI Agents API: Build Managed Codex-Style Agents Without Your Own Harness
  Source: devto-javascript
  Topic: JavaScript (+AI)
  URL: https://dev.to/adityahalderdev/openai-agents-api-build-managed-codex-style-agents-without-your-own-harness-54p9

- Building a random chat where language doesn't matter: WebRTC, relay as a condition, and live translated captions
  Source: devto-javascript
  Topic: JavaScript (+AI)
  URL: https://dev.to/vasiliy_shkitin/building-a-random-chat-where-language-doesnt-matter-webrtc-relay-as-a-condition-and-live-315g

- AI Weekly Issue #530: Applied AI This Week
  Source: ai-weekly
  Topic: AI (+AI Research)
  URL: https://aiweekly.co/issues/ai-is-getting-real-jobs-the-evidence-is-uneven

## Research

- One Capital Letter Was Silently Breaking My AI Support Bot, and It Wasn't in the New Model
  Source: towards-data-science
  Topic: AI Research (+AI, Python)
  URL: https://towardsdatascience.com/one-capital-letter-was-silently-breaking-my-ai-support-bot-and-it-wasnt-in-the-new-model/

- Stop Managing Alarms: An Incident-First Blueprint for Telecom AIOps
  Source: towards-data-science
  Topic: AI Research (+AI, Python)
  URL: https://towardsdatascience.com/stop-managing-alarms-an-incident-first-blueprint-for-telecom-aiops/

- No-Box Vulnerability Analysis: Description-only Detection of Indirect Prompt Injection Vulnerabilities in MCP Servers
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10854

- AI Economist Agent: An Agentic Framework for Evidence-Based Economic and Financial Analysis with RAG, Knowledge Graphs, and Large Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.20041

- Agentic Share-of-Search: A Multi-Agent AI System for Competitive Decision-Making in LLM-Mediated E-Commerce
  Source: arxiv-cs-ai
  Topic: AI Research (+AI, JavaScript, React)
  URL: https://arxiv.org/abs/2609.11190

- DriftNet: A Dual-Head Trajectory Transformer for Detecting and Localizing Prompt Injection in LLM Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10892

- Published Unlearning Numbers Move Per Checkpoint, and Not Because the Removed Data Survives: An Audit of 263 Released Batch-Normalized Checkpoints
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11490

- DiaLLM: An Investigation into the Robustness-Generation Gap in English Dialect Adaptation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.07669

- Automating Quadratic Unconstrained Binary Optimization (QUBO) Formulation Generation from Natural Language
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10629

- Clearing the Fog: Towards Installing and Refining Proactive Exploration Capabilities in LLM Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.14339

- Off-Target Effects of Response-Style Alignment in a Korean 27B Language Model
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11291

- GameWAM: A World Action Model for Video Games
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.26200

- Architecting the Secure AI-SOC: A Neurosymbolic Framework for Pipeline Integrity and Threat Mitigation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10707

- HarvestBench: Measuring Whether LLM Agents Will Pay to Avoid Killing Animals
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.04444

- LOCUS: Task-Aware Low-Rank Post-Training for Token-Efficient Language Generation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11739

- Solving Few-Shot Multiobjective Multitask Optimization via Iterative Sequential Transfer
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11228

- Beyond Word Error Rate: A Switch Aware Evaluation of ASR and Audio Language Models on English Yoruba Code-Switched Speech
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11786

- VALG: An Agentic System for ML Theory Research
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.13060

- Data-Efficient Language Modeling: From Frontier Advancement to Principle-Guided Model Improvement
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10702

- General Quantification of Covariate and Concept Shifts
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11918

- Generative Marketing Mix Modeling: A Causal Inference Framework Linking GEO and GEM to Business Impact
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11915

- Deep-Fake CAPTCHA: Mitigating Next-Generation Social Engineering Attacks
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11404

- Investigating catastrophic forgetting in sound event classification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11447

- When Agents Disagree: Bayesian Backward Reasoning as a Label-Free Anchor for Multi-Agent Collective Decision-Making
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11709

- EVOQUANT: Self-Evolving Verifier-Guided Strategy Optimization for Robust Quantitative Trading
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.12455

- RAMamba-Net: A Reliability-Aware and Mamba-Based Multimodal Fusion Network for Auditory Attention Detection
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11372

- Discovering Temporal Structure: An Overview of Hierarchical Reinforcement Learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2506.14045

- Cross-Lingual Clinical Annotation Projection as Constrained Text Generation: A Six-Language Study
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11450

- LoaDiff: Conditional Generation of Electricity Consumption Time Series for Energy Analytics
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11639

- ORCH: Organizational Principles Enable Collective Intelligence in Embodied AI
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11737

- Extending SMT Solving with Non-Ground Clause Learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11509

- Learning What to Retain: Gated-Memory Routing for Efficient Collaboration in Multi-Agent LLM Systems
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.00237

- Exploring Multimodal Prompt for Visualization Authoring with Large Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2504.13700

- Toward Interpretable Multimodal Fusion: Heat Conduction Modeling for Hyperspectral and LiDAR Joint Classification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11040

- VikingRAG: Accurate and Token-efficient Retrieval-augmented Generation over Structured Documents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11390

- Timely Clinical Diagnosis through Active Test Selection
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2510.18988

- Temporal and Multimodal Deep Learning for Cyberattack Detection in LEO Satellite Systems
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10746

- Certifying cooperation: a novel approach to cooperative multi-agent task generation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.06586

- Spectral Masking and Interpolation Attack (SMIA): A Black-box Adversarial Attack against Voice Authentication and Anti-Spoofing Systems
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2509.07677

- A Fragility Spectrum for Recursive Language-Model Training
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11149

- Less can be More: What Aspects of Speech Drive End-of-Turn Detection
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11066

- CryptoL: Towards Scale Dominance and Physics Constraints Mitigation in Financial Multivariate Time Series Forecasting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11206

- Optimizing Three Critical Factors for Practical and Effective OOD Detection Fine-Tuning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2308.01030

- Improving Natural-Language Combinatorial-Optimization Accuracy in Resource-Constrained Language Models via Formal Abstractions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.18409

- Breaking Predictions Is Not Enough: Specified-Foil Counterfactuals for Temporal Graphs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11170

- Why Sample What You Can Enumerate? Exact Policy Optimization for Genomic Tool Selection
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10221

- DRG-MAPPO: Hierarchical Dynamic Role-Graph Multi-Agent Reinforcement Learning for Cooperative Air Combat
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11155

- How Proper Scoring Rules Shape LLM Forecasting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.28482

- Can LLMs Follow Medical Expert Logic? A Benchmark for Hierarchical Logical Consistency in Risk-of-Bias Assessment
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11185

- Agent-Integrated Software: Interaction Contracts and Continuous Assurance
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11381

- KuaiRP Series Role-playing Models Technical Report
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11127

- A Multi-Stage Rule-Chaining Framework for Compositional and Interpretable Cognitive Reasoning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10654

- SIRF: A Spec-Internalized Risk Foundation Model for Industrial Content Risk Control
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11752

- AUC Maximization from Biased Positive-unlabeled Data with Confidence
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10928

- Enabling Knowledge Graph Understanding at Scale with the EXplore Your Graphs ENgine (EXYGEN)
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11569

- Terminal Symmetry as a Carrier of Asymmetric Process Knowledge: Statewise Refinement for Anytime Verified Construction
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.11318

- GPU-CFR: 80x Faster Counterfactual Regret Minimization by Compiling the Game to Static Dataflow and CUDA Graph Replay
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11923

- Importance Weighting for Unlabeled-unlabeled Learning under Distribution Shift
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10994

- Explainability Assistant: A Conversational XAI Interface for Interpreting Energy Consumption Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11860

- ActSafeGuard: Differentiable and Training-Aligned Constraint Enforcement for Flow-Matching Policies
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11697

- How LLMs Follow Instructions: Skillful Coordination, Not a Universal Mechanism
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.06015

- From Document Silos to Process Intelligence: A Multi-Layer Knowledge Graph for CMC Process Development
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11493

- Compact Visuotactile World Models for Lifting: Prediction, Reward Alignment, and Force Constraints
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.09597

- SemVerBench: Benchmarking LLM Comprehension of Version-Constraint Resolution Semantics
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11180

- Exploring Diffusion Transformers for Cross-Modal Augmentation in Multimodal Brain State Decoding
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11341

- terms.txt: A Consent and Compensation Protocol for Agentic Web Access
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11152

- Lightweight LiDAR-Based Cone Detection Framework Using Random Forest for Formula Student Driverless
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11527

- A Survey of Threats Against Voice Authentication and Anti-Spoofing Systems
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2508.16843

- Continuous-Time Acoustic Modelling with Neural Controlled Differential Equations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11725

- Mr.LHDR: A Benchmark for Multimodal Real-World Long-Horizon Deep Research Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11318

- Evaluating Scaffolding-Oriented Multi-Agent Large Language Model System for Clinical Interview Training
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10939

- A Time-Based Readout for Vector-Matrix Multiplication in Fully Analog Memristive SNNs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11713

- Beyond Benchmarks: Using VLMs to Reveal Systematic Classification Failures Under Real World Conditions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11126

- Cognitive Digital Twins: Ethical Risks and Governance for AI Systems That Model the Mind
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.23094

- CLSP-REQA: A Real-Time Quality-Aware Closed-Loop Seizure Prediction Framework with Mamba-BiLSTM and Confidence-Gated Intervention
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.00074

- ProsMAE: Multi-Source MAE Pretraining for ISUP Grade Classification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.08162

- Generating a Consistent Enterprise: Synthesis and Reference-Free Evaluation of Multi-System Business Data
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11286

- AI Soccer Analyst: Stage-Aware and Verifiable Human-AI Collaboration for Soccer Data Analysis
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11224

- Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.04298

- ReactHuman: A Physics-Grounded Benchmark for Human-Like Reactive Decision-Making in Embodied Multimodal LLMs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10895

- TRUST-SQL: Tool-Integrated Multi-Turn Reinforcement Learning for Text-to-SQL over Unknown Schemas
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.16448

- FaultLens: Learning Compact Behavioral Test Suites for Generated Operational Programs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.26746

- 2AM: Grounding Agent-Side Memory as Guidance for Steerable Action Models in Long-Horizon Manipulation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11308

- Quantifying the Memorization-to-Generalization Transition: Scaling Laws and Phase Structure in Grokking
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10657

- A Density-Matrix Framework for Electronic-Structure Analysis of Electrolytes for Lithium Batteries
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.25597

- Demystifying the Privacy-Utility Trade-off in LLM Interactions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10992

- Wiggle and Go! System Identification for Zero-Shot Dynamic Rope Manipulation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.22102

- From Queries to Narratives: Cultural Heritage Data Stories for Knowledge Graph Exploration and Quality Assessment
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11403

- Probing for Knowledge Attribution in Large Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2602.22787

- Multilingual in Name Only? Cultural and Linguistic Weaknesses of LLMs in Urdu
  Source: arxiv-cs-ai
  Topic: AI Research (+AI, Information Retrieval)
  URL: https://arxiv.org/abs/2609.10758

- Autonomy, Social Norms, and Alignment: Towards a Developmental Framework for Autonomous Artificial Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11660

- A Multi-Modal Perception Pipeline for Object Detection and Tracking in Autonomous Racing
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.08338

- OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesis
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.20278

- Studying Without a Syllabus: Task-Agnostic Environment Preprocessing
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10824

- Verification of Adaptive Agentic Controllers through Finite Rule Revision
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.09770

- Benchmark Radar: A Living Database and Search Engine for AI Benchmarks and Evaluation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI, Information Retrieval)
  URL: https://arxiv.org/abs/2609.11115

- Distributed Optimization of Modular Production Systems using Model-based Reinforcement Learning with Inverse Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11615

- Decoupling Readiness from Release for Tail-Aware Scheduling of Agentic LLM Workflows
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10964

- RouteRepair: Instance-Level Failure Diagnosis and Targeted Repair in LLM-Based Automated Heuristic Design for Routing Optimization
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11452

- Beyond Confidence: Stability-Aware Test-Time Adaptation for LLM Reasoning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11393

- Rescaling Confidence: What Scale Design Reveals About LLM Metacognition
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.09309

- Characterizing Bluesky Content Moderation Service: From Automation of Service to Landscape of Harms
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11373

- From Evaluation to Enhancement: Benchmarking and Improving Think-with-Video Reasoning for Video Generative Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11242

- GitSkills: A Dataset of Agent Skills on GitHub
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.10906

- SG-Blend: Learning an Interpolation Between Improved Swish and GELU for Robust Neural Representations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2505.23942

- Probabilistic Focal Search: Accelerating Bounded-Suboptimal Search via Lower-Bound Advancement
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10584

- MOSAIC: Query-Aware Exploration Policy Adaptation for GraphRAG
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11065

- Refusal Beyond a Single Direction: A Preliminary Comparison of Diff-in-Means and INLP
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.13720

- PACE: Perceived-Latency-Aware Cascading Service Routing and Filler Control for QoE-Efficient Retrieval-Augmented Dialogue Serving
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10372

- COBRA-Skills: Contextual Bandit-Guided Evolution for Agent Skill Optimization
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11682

- DeFiFusion: Combining Transaction Events with Smart Contracts to Detect Price Manipulation Attacks
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11008

- Planning and Scheduling Business Processes under Control-Flow Uncertainty: Extended Version
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.05578

- A Comparative Evaluation of Pre-trained Convolutional Neural Networks for Melanoma Detection
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11550

- Thinking with Looped Flows
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11801

- Towards a Deterministic Math Solver for Clinical Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10728

- Routing by Reasoning Need: Trajectory-Aware Decoding Control for Diffusion Vision-Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11315

- Sci-MMR: Benchmarking Multi-Step Evidence-Grounded Scientific Reasoning in Multimodal Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11243

- On the Impact of Anonymization on the Performance of Large Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11335

- Ambient @ EgoProactive 2026 : Proactive Egocentric Assistance with Visually Grounded Supervision
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.07099

- LLMs as Post-hoc Auditors of Physiological Plausibility in Symbolic Regression: A Clinician-Evaluated Case Study
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11431

- When Synthetic Data Hurts: On Catastrophic Forgetting in Skill Retrieval for LLM Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10750

- What a Random Draw from the MCP Registry Contains, and What Tool-Use Benchmarks Contain Instead
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10962

- Adaptive Margin Ordinal Loss: Penalizing Center-Class Hedging in Ordinal Classification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10752

- When Does Text Inform? Benchmarking Information-Theoretic Metrics for Multimodal Time-Series Forecasting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11282

- Adaptive Perturbation Selection for Contrastive Audio Decoding
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.00247

- Can Edge-Deployable Vision-Language Models Identify Species?
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11916

- Magenta: Closing the Loop Between Mathematical Reasoning and Lean Verification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11319

- From Agent Traces to Trust: A Survey of Evidence Tracing and Execution Provenance in LLM Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.04990

- CAT-GS: Balanced Multimodal Learning via Calibrated Gating and Fusion Surgery
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.24947

- The Semantic Elevation Operator and the Closure of the Undecidable Class under Preservation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11326

- Warrant Theory
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11667

- Learn the Solid, Not the File: Canonical Inputs for Neural Networks on CAD Boundary Representations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11573

- Structural priors for data-efficient language learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11505

- SPECTRA: Band-Routed Embedding and Stage-Wise LoRA for Cross-Sensor Fine-Tuning of Geospatial Foundation Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.01751

- Learning Interaction between Image and Layout Priors for Joint Image-Layout Generation in Design Templates
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11519

- No Screening is More Efficient with Multiple Objects
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2408.10077

- Monadic Second-Order Logic in HOL: Deep and Shallow with Automated Faithfulness (Extended Preprint)
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.07345

- New Evidence, Same Choice: Testing Physical Experiment Selection in Vision Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11022

- Debate-to-Skill: Capability-Bound Process Supervision for Industrial Query-to-Agent Annotation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11176

- Making Alternative Data Work: Context-Augmented LLMs for Financial Forecasting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11607

- Reification as a Transferable Vocabulary: Zero-Shot Link Prediction with Vanilla GNNs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11347

- Causal Past Logic for Runtime Verification of Distributed LLM Agent Workflows
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.20923

- NovGauge: A Fine-Grained Benchmark for Diagnosing LLMs' Capability in Paper Novelty Assessment
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11234

- Learning Intrinsic Water-Quality Dynamics with Rainfall for Data-Driven Forecasting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2508.08279

- RetroThinker: Enabling Retrospective Thinking in Speech LLMs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11864

- Reason Through the Latent! Making Latent Visual Reasoning Necessary
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.06746

- Fork Where the Model Changes Its Mind: Belief-Shift Branching for Tree-Structured Reinforcement Learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11061

- Natural Language Access to Domain-Specific Metadata: A Reusable Framework for LLM Query Generation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.18029

- Amulet: a Python Library for Assessing Interactions Among ML Defenses and Risks
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2509.12386

- HALDETECT at ImageEval 2026 Shared Tasks: Answer-First Contrastive Grounding with QLoRA
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11236

- The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11873

- When Validation Stops Learning: Auditing Update Admission for Continual Embodied Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10873

- Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10883

- OmegaUse-SOP: SOP Engineering for Professional Computer Use from Human Demonstrations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.02149

- Sympathetic Framing: Evaluating AI Alignment across Sociodemographic Groups
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.27232

- Aslema at NADI 2026: Data Augmentation for Intent Recognition and Slot Filling
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.18689

- E-CONAN (Entailment, CONtradition And Neutral) Benchmarks: Arabic Textual Entailment and Natural Inference Datasets
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11334

- Edu-QuRating: Multi-Dimensional Educational Data Curation with Distilled Pairwise Judgements
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.09425

- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.12898

- CertDW: Towards Certified Dataset Ownership Verification via Conformal Calibration
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2506.13160

- T1: Terminal Agent Reinforcement Learning for Long-Horizon Tasks
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11042

- EGGROLL, Unrolled: Understanding and Improving Low-Rank Evolution Strategies at Scale
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10980

- An AI-Powered Culturally Aware Chatbot for Stress Detection and Wellness Support among Pakistani University Students Using NLP and Machine Learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11199

- HISA: Efficient Hierarchical Indexing for Fine-Grained Sparse Attention
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.28458

- Are We Really Doing Few-Shot Learning? A Critical Examination of Pre-Training Assumptions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10851

- REAL-Q: E2E LLM Quantization via Dynamic Gradient Descent
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.00049

- EvoMaster: A Foundational Evolving Agent Framework for Agentic Science at Scale
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.17406

- AI-Powered Flare Combustion Efficiency Estimation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11262

- Ecdysis: Efficient and Effective Training of Runtime Harnesses for LLM Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11677

- Ceci n'est pas une pipe: AI systems as semantic abstractions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.09489

- How AI Coders Discuss, Disagree, and Reach Consensus: Challenges and Opportunities for LLM-Based Qualitative Coding
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11109

- Perturbation: A simple and efficient adversarial tracer for representation learning in language models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.23821

- Automated multi-class wound assessment using dedicated instance segmentation models for boundary detection and classification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.27325

- Geospatial AI, Dataverse Metadata, and the Study of Place-Based Government
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11674

- Characterizing Job Power Elasticity for Power-Flexible AI Training
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11542

- GRIPNet: Gaussian Radial Intensity Prior Guided Architecture for Pulmonary Nodule Detection in CT
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11312

- Your Agent Says Yes: Interpreting Adversarial Market Behavior Beyond Individual Transactions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.07675

- Analyzing LLM Reasoning to Uncover Mental Health Stigma
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.25053

- LLM-Ideoplasticity: Measuring Ideological Plasticity in the Political Behavior of LLMs as a Context-Conditioned Distribution
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.28335

- X-RACE: XAI-assisted Recurrent neural network Attribution for Channel Estimation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11211

- Counterfactual Marginalisation: Framework for Evaluating Robustness to Nuisance Variables
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10778

- ScaleResfusion: Residual Rectified Flow based on Residual Vector Field
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.25275

- MAPLE: Memory-Augmented Planning with Language and Evolution
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11636

- Language-Augmented Semantic Priors for B-Spline Surface Fitting
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11708

- AcFlow: Controlling Text-to-Image Diffusion Transformers via Learned Conditional Activation Flow
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10723

- MindTopo: Can Foundation Models Reason in Topological Space?
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11900

- Prevalence Determines Precision:Silent Contamination in Detector-Defined Datasets
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11449

- Toward Collective-Centric Evaluation of Preference Inference for Participatory Democracy
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.02990

- Beyond Static Guarantees: Measuring the Static-Pass Dynamic-Fail Gap in Security-Sensitive and LLM-Generated Python Code
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10762

- Topological Necessities: Mechanism-Invariant Strategic Subgoals for Cross-Embodiment Goal-Conditioned Control
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11014

- Flexible and Interpretable Accent Distance Measurements
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11458

- FastE: Readout-Triggered Token Compression for LLM Embedding Inference
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.08407

- The Truth Was Never Gone: Perfect Aliasing in Compliant-Context Truth Probes
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10739

- On the Regularization Landscape for the Linear Recommendation Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11876

- Your Model Already Knows Don't Teach It, Learn to Ask It: Soft Prompting for Few-Shot Adaptation of Vision-Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11310

- SciFigQual-Bench: A Benchmark for Scientific Figure Quality Assessment with Full-Manuscript Context
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.27084

- Recognizing Is Not Reversing: A Controlled Inversion Test of Fact-Preserving News Framing
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11769

- Defining AI Agents: A Compendium of Criteria, Metrics, and Benchmarks
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11018

- A Mathematical Theory of Pragmatic Information
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10986

- Physics-Informed Neural Networks to Infer the Perpendicular Energy Conductivity in the Scrape-Off Layer of Stellarator Devices
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11628

- The Convention Gap: Towards Measuring Implicit Communication in Cooperative AI Evaluation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11489

- AI Exposure and AI Resilience: A Two-Dimensional Assessment Framework for Software and Software-Based Business Model
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11321

- Role differentiation as ignition of a collective information engine: Structuration in Agent Populations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.05442

- SimSkill: A Self-Evolving LLM Agent for Skill and Knowledge Accumulation in Traffic Simulation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.03753

- (Whose defaults?) Is artificial intelligence reorienting archaeological methods?
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11198

- Output Embedding Centering for Stable LLM Pretraining
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2601.02031

- Generative AI performance in core undergraduate mathematics: a curriculum-level case study
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2509.13359

- Robust Multimodal Sentiment Analysis with Incomplete Modalities via Semantic-aware Completeness based Reconstruction
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10950

- BenchShield: Formal Model-Backed Instrumentation for Reward Integrity in LLM-Agent Evaluation Infrastructure
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11028

- Evidence for Limited Metacognition in LLMs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2509.21545

- A Calibration Audit of Confidence in Feed-Forward 3D Reconstruction Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.29705

- SWRouter: Similarity-Contractive Window Routing for Multi-Turn Large Language Model Conversations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11414

- Memory Compression for High-Fanout Agent Sandboxes
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11294

- On the Societal Impact of Machine Learning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2510.23693

- SGA: Plug&Play Geometric Verification for Educational Video Synthesis
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2607.18116

- Exploring Second-Order Pattern Recognition in Speaker Recognition
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11182

- An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10712

- Improving Faint Object Detection for Space Situational Awareness with Variational Autoencoders
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11269

- DCO: Dynamic Cache Orchestration for LLM Accelerators through Predictive Management
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2512.07312

- Biology-in-the-loop: Amortized Adaptive Hit Discovery in CRISPR Screens
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11877

- A Unified Per-Token Gating Family for On-Policy Distillation: FKL/RKL Mixing with Multi-Channel and Bias Coefficients
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11768

- ZipCodec: Ultra-Low-Frame-Rate Streaming Speech Coding
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11642

- Domain-Specific Hallucination Detection in Large Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11878

- Bio-inspired Learning and Decision-Making with Probabilistic In-Memory Computing Hardware: Part 1
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11281

- Generalization in VAE and Diffusion Models: A Unified Information-Theoretic Analysis
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2506.00849

- Motus2: A Self-Evolving General World Model for Dexterous Manipulation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.30237

- Towards AI-Driven Policing: Interdisciplinary Knowledge Discovery from Police Body-Worn Camera Footage
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2504.20007

- GoAnt: Quality-Diversity Multi-Agent Search for Alpha Factor Discovery in Market Microstructure Data
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.08719

- A Voice-Interactive Multi-Agent System for Smart Operating Rooms: Architecture Design and Key Technologies
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11231

- Does Linguistic Structure Enrichment Enhance Coherence Assessment? Not With Current Architectures
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10893

- Beyond Visual Quality: Evaluating Physical Consistency under Ego-Motion with EgoGenEval
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11172

- Beyond Prompting: Efficient and Robust Contextual Biasing for Speech LLMs via Logit-Space Integration (LOGIC)
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2601.15397

- Same Day, Same Story; One Day Ahead, a Different Signal: The Dual Validity of Financial Sentiment
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11144

- Monotone Neural Policy Iteration for High-Dimensional First-Order Hamilton--Jacobi--Bellman Equations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.07116

- Beyond Single-Negative Preference: Multi-Negative DPO for LLM-Centric Historical Entity Linking
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.07379

- Grounding Agent Memory: Environment-Probing Curation for Enterprise Agents
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11060

- VeriSim: A Configurable Framework for Stress-Testing Medical AI Under Patient Communication Noise
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.10441

- Buyer Artificial Intelligence-Enabled Environmental Governance and Supplier Environmental Controversies: An Organizational Information Processing and Signaling
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11391

- Portable Semantics, Private Dialects: Reuse and Negative Transfer in Latent Communication Between Language-Model Cells
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11365

- Logit Refiner: Improving Visual Autoregressive Models via Intra-Scale Dependency Modeling
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11804

- Efficient Adaptation of LLMs for Hate Speech Detection in Low-Resource Languages: A Comparative Study on Roman Urdu
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2608.18142

- Relevance Is Not Permission: Localizing and Controlling Metric-Facing Attention Contributions
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.30139

- Continuous Diffusion Scales Competitively with Discrete Diffusion for Language
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2605.18530

- Some hypotheses on how chatbots work in problem-solution-driven conversations: Large Language Models as confirmation of the Innovation Illusion
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.07722

- Generative Replay Mitigates Sample Starvation in Quantum Architecture Search
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11248

- An Agentic Evaluation Framework for AI-Generated Scientific Code in PETSc
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.15976

- A Lightweight Multi-Agent Framework for Automated Concrete Barrier Design
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2606.12040

- The Oligarch Barely Steers Model Collapse in Multi-Model Ecosystems
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11146

- Understanding Operator Attitudes Toward AI-Supported Decision Making in Maritime Operations
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11805

- CARTS: Contextual Autoregressive Rank Transcoding Steganography for Full-Capacity Keyed Text Encoding
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10744

- Tapes Together Strong: The Co-evolution of Computation and Cooperation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10817

- Federated Learning for Surgical Vision in Appendicitis Classification: Results of the FedSurg EndoVis 2024 Challenge
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2510.04772

- Beyond Verified Answers: Solver-Informed Self-Distillation for Bootstrapping Operations Research Language Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.09957

- ActMap: Single-Pass Uncertainty Quantification from Generation-Time Activation Maps
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11498

- Model-Aware Schedules Improve Generation via Fiberwise Optimal Transport
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11842

- X-AuT: Progressive Audio-Encoder Compression for Speech LLMs with Cross-Scale Distillation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11412

- Domain Elastic Transform: Bayesian Function Registration for High-Dimensional Scientific Data
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.21235

- Ordinary, Reasonable Chatbots: Do AI Models Track Human Legal Judgments?
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.06769

- The Agent Incident Registry: Toward Preventing Repeated AI Agent Failures
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11030

- Cognitive Amplification vs Cognitive Delegation in Human-AI Systems: A Metric Framework
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.18677

- Understanding LoRA Rank Trade-offs in Diffusion Model Fine-Tuning
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10656

- Prompt Revision as a Source of Cultural Bias in Text-to-Image Systems
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11532

- Artificial Id: Drive and Persistent Alignment in Agentic AI
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11911

- Calibration-Aware Uncertainty Cascades for Efficient Heterogeneous Model Collaboration
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11446

- A machine-checked proof of the Dong-Yang classification of optimal (n,4) binary codes for BSCs
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10579

- Autonomous Chemical Mechanistic Discovery through Agentic Reasoning and Validation
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11147

- CHRONOBERG: Capturing Language Evolution and Temporal Awareness in Foundation Models
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2509.22360

- Learning to Think Like a Cartoon Captionist: Incongruity-Resolution Supervision for Multimodal Humor Understanding
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.15210

- MOSAIC: A Universal Agent-Level Interface for Cross-Paradigm Agent Mixing and Human-AI Collaboration
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.01260

- ExpTest: Loss-Curve Hypothesis Testing for Autonomous Learning-Rate Selection in Deep Neural Networks
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2411.16975

- Predicting Train Delays in Finland Using Machine Learning and Weather Data
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11277

- From Parameters to Answers: How LLMs Retrieve and Use Their Internal Knowledge
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.11859

- Towards Automated Solar Panel Integrity: Hybrid Deep Feature Extraction for Advanced Surface Defect Identification
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2604.10969

- Finishing the Task Is Not Enough: Evaluating Agent Resilience and Considerate Participation under Accumulating Challenge
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2609.10724

- Four Generations of Quantum Biomedical Sensors
  Source: arxiv-cs-ai
  Topic: AI Research (+AI)
  URL: https://arxiv.org/abs/2603.29944

## Community

- Generating running routes with GPT-6 Astra and ChatGPT Work
  Source: simon-willison
  Topic: Python (+AI, Django)
  URL: https://simonwillison.net/2026/Sep/12/astra-running-routes/

- California Brown Pelican
  Source: simon-willison
  Topic: Python (+AI, Django)
  URL: https://simonwillison.net/2026/Sep/12/sighting-399708714/

- Quoting Paul Ford
  Source: simon-willison
  Topic: Python (+AI, Django)
  URL: https://simonwillison.net/2026/Sep/12/paul-ford/

- Compiling V8 (d8 shell) to WASM
  Source: rust-users-forum
  Topic: Rust (+AI)
  URL: https://users.rust-lang.org/t/compiling-v8-d8-shell-to-wasm/142406

- OpenAI agents attacked RubyGems back in May
  Source: simon-willison
  Topic: Python (+AI, Django)
  URL: https://simonwillison.net/2026/Sep/12/openai-agents-rubygems/

- Armin Ronacher: P(doom)
  Source: planet-python
  Topic: Python (+AI)
  URL: https://lucumr.pocoo.org/2026/9/12/pdoom/

- An Advanced System Architecture Breakdown of OpenAI's Jalapeno Accelerator
  Source: hacker-news
  Topic: General (+AI)
  URL: https://www.siliconcodesign.com/p/an-advanced-system-architecture-breakdown

- OpenAI's Sam Altman says it would be 'ill-advised' to go public in 2026
  Source: hacker-news
  Topic: General (+AI)
  URL: https://techcrunch.com/2026/09/12/openais-sam-altman-says-it-would-be-ill-advised-to-go-public-in-2026/

- Everyone should slow down AI development except for me
  Source: lobsters
  Topic: General (+AI)
  URL: https://xeiaso.net/notes/2026/everyone-slowdown-but-me/

- Show HN: Determinstic LLM inference for lowest price Gemma 4, with Windows XP
  Source: hacker-news
  Topic: General (+AI)
  URL: https://www.tokendelivery.ai/

- We Must Pace the Frontier
  Source: lobsters
  Topic: General (+AI)
  URL: https://darioamodei.com/post/we-must-pace-the-frontier

- Ask HN: Did Google kill its enterprise workhorse model?
  Source: hacker-news
  Topic: General (+AI)
  URL: https://news.ycombinator.com/item?id=49668196

- OpenAI: "We use ... de-identified data to improve ChatGPT"
  Source: hacker-news
  Topic: General (+AI)
  URL: https://twitter.com/markchen90/status/2097400166554993041