TILens Daily Edition 2026-09-21 Filters: topic=ai, github=hidden Stats: 409 articles, 27 sources, 361 research papers Top topics: AI, AI Research, General ## News - Pruning LLMs Like a Physicist: Block Removal as an Ising Optimization Problem Source: huggingface-blog Topic: Hugging Face (+AI) URL: https://huggingface.co/blog/MultiverseComputingCAI/pruning-llms-like-a-physicist-block-removal-as-an - Higgsfield AI ships new video features in a day with GPT-6 Astra Source: openai-blog Topic: AI Labs (+AI) URL: https://openai.com/index/higgsfield-from-prompt-to-production-with-astra - Advisory Group on Mathematics and Artificial Intelligence Source: openai-blog Topic: AI Labs (+AI) URL: https://openai.com/index/advisory-group-on-mathematics-and-ai - Building standards for the next phase of AI Source: openai-blog Topic: AI Labs (+AI) URL: https://openai.com/index/building-standards-next-phase-ai - Expanding OpenAI Academy with new learning paths Source: openai-blog Topic: AI Labs (+AI) URL: https://openai.com/index/expanding-openai-academy-with-new-learning-paths - Spring Office Hours Podcast: S5E24 - Jev, Java and Spring Source: spring-blog Topic: Java (+AI) URL: https://spring.io/blog/2026/09/21/spring-office-hours-podcast-S5E24 - V7 cuts costs 78% while boosting accuracy with GPT-5.6 Luna Source: openai-blog Topic: AI Labs (+AI) URL: https://openai.com/index/v7 - tokenizers v1: encode, decode and scaling, measured Source: huggingface-blog Topic: Hugging Face (+AI) URL: https://huggingface.co/blog/tokenizers-v1 - 🦞 LIVE NOW: Build a personal AI agent you own Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/newsletter/live-openclaw-2-0-with-chief-architect-vincent-koc/ - Googlebook 899 USD chạy Android: có đáng mua với dân VN? Source: devto-javascript Topic: JavaScript (+AI) URL: https://dev.to/bean_bean/googlebook-899-usd-chay-android-co-dang-mua-voi-dan-vn-584m - Cắt chi phí agentic coding: tách planner và worker Source: devto-javascript Topic: JavaScript (+AI) URL: https://dev.to/bean_bean/cat-chi-phi-agentic-coding-tach-planner-va-worker-44a5 - Anthropic-linked CVEs pile up, attackers mostly shrug Source: theregister Topic: General (+AI) URL: https://www.theregister.com/security/2026/09/21/anthropic-linked-cves-pile-up-attackers-mostly-shrug/5298018 - How AI Agents Can Trigger Runaway Costs for Enterprises Source: darkreading Topic: DevSec (+AI) URL: https://www.darkreading.com/application-security/how-ai-agents-can-trigger-runaway-costs - OpenAI forms math advisory group as its AI resolves more than 100 open problems Source: techcrunch Topic: General (+AI) URL: https://techcrunch.com/2026/09/21/openai-forms-math-advisory-group-as-its-ai-resolves-more-than-100-open-problems/ - TypeSafe launched Jev because sequential LLMs are “totally useless for computers” Source: newstack Topic: DevOps (+AI) URL: https://thenewstack.io/typesafe-jev-system-one/ - How to Use OpenCode for Free: Beginner's Desktop Setup Guide Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/explainer-articles/how-to-use-opencode-for-free-beginners-desktop-setup-guide/ - Treasury chief says AI bosses, not their bots, will carry the can for criminal acts Source: theregister Topic: General (+AI) URL: https://www.theregister.com/security/2026/09/21/treasury-chief-says-ai-bosses-not-their-bots-will-carry-the-can-for-criminal-acts/5297965 - xAI’s Grok 4.7 Makes Frontier AI a Price War Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/news/xais-grok-47-makes-frontier-ai-a-price-war/ - Google confirms Gemini models hacked three companies in May 2026 Source: ars-technica Topic: General (+AI) URL: https://arstechnica.com/google/2026/09/google-confirms-gemini-models-hacked-three-companies-in-may-2026/ - How I built an Agentic AI UI with React and Gemini API (and handled the 503 Overload Error) Source: devto-javascript Topic: JavaScript (+AI, FastAPI, Python, React) URL: https://dev.to/koliabondar/how-i-built-an-agentic-ai-ui-with-react-and-gemini-api-and-handled-the-503-overload-error-3mmn - Rogue Behavior: OpenAI Reveals More Model Misalignment Incidents Source: darkreading Topic: DevSec (+AI) URL: https://www.darkreading.com/cyber-risk/rogue-behavior-openai-more-model-misalignment-incidents - Google’s $899 Googlebook is a bet that you’ll buy a new laptop for Gemini Source: techcrunch Topic: General (+AI) URL: https://techcrunch.com/2026/09/21/googles-899-googlebook-is-a-bet-that-youll-buy-a-new-laptop-for-gemini/ - Architect Labs Wants AI Chips to Evolve as Fast as AI Models Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/explainer-articles/architect-labs-ai-chips-evolve-as-fast-as-ai-models/ - Import AI 473: The US's superintelligence strategy; human brain in a mouse skull; and machine hermeneutics Source: importai Topic: AI URL: https://importai.substack.com/p/import-ai-473-the-uss-superintelligence - 4 ways to address the failures we found along the US border’s “virtual wall” Source: mit-ai Topic: AI URL: https://www.technologyreview.com/2026/09/21/1144164/border-towers-surveillance-policy-recommendations/ - The US spent billions on border surveillance. Why can’t it catch people before they die? Source: mit-ai Topic: AI URL: https://www.technologyreview.com/2026/09/21/1144166/border-towers-surveillance-investigation/ - She died at the San Diego border. A surveillance camera was in plain sight Source: mit-ai Topic: AI URL: https://www.technologyreview.com/2026/09/21/1144170/border-towers-surveillance-feature/ - How we made the first comprehensive map of deaths along the US border’s “virtual wall” Source: mit-ai Topic: AI URL: https://www.technologyreview.com/2026/09/21/1144161/border-towers-surveillance-methodology/ - Why You Should Never Embed Your Gemini API Key in Client Code (And How Firebase AI Logic Fixes It) Source: freecodecamp Topic: General (+AI) URL: https://www.freecodecamp.org/news/why-you-should-never-embed-your-gemini-api-key-in-client-code-and-how-firebase-ai-logic-fixes-it/ - How to Use Claude Code in VS Code: Install, Inline Diffs, and When to Stay in the Terminal Source: devto-javascript Topic: JavaScript (+AI) URL: https://dev.to/gil_5296961bf2e126cf43cb4/how-to-use-claude-code-in-vs-code-install-inline-diffs-and-when-to-stay-in-the-terminal-2fof - Presentation: The Agent Harness: Control Planes, Invariants, and Approval Boundaries for Production AI Agents Source: infoq-architecture Topic: Architecture (+AI) URL: https://www.infoq.com/presentations/ai-agent-harness/?utm_campaign=infoq_content&utm_source=infoq&utm_medium=feed&utm_term=Architecture+%26+Design - Jensen Huang Says AI Won’t End the World. His Bigger Bet Is on Existing Law. Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/news/jensen-huang-says-ai-wont-end-the-world-his-bigger-bet-is-on-existing-law/ - OpenAI goes from hacker to hacked Source: the-rundown-ai Topic: AI (+AI Tools) URL: https://www.therundown.ai/articles/openai-goes-from-hacker-to-hacked - 😸 Would GPT-6 stab a doll? Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/newsletter/would-gpt-6-stab-a-doll/ - When AI Gets Arms, “Just Say No” Stops Being a Safety System Source: the-neuron-ai Topic: AI (+AI Research) URL: https://www.theneuron.ai/robotics-insider/when-ai-gets-arms-just-say-no-stops-being-a-safety-system/ - Stay ahead in web development: latest news, tools, and insights #155 Source: devto-javascript Topic: JavaScript (+AI) URL: https://dev.to/urbanisierung/stay-ahead-in-web-development-latest-news-tools-and-insights-155-8n4 - Google joins the ‘Oops, our agents hacked someone’ club after partner’s internet access error Source: theregister Topic: General (+AI) URL: https://www.theregister.com/ai-and-ml/2026/09/21/google-joins-the-oops-our-agents-hacked-someone-club-after-partners-internet-access-error/5297640 - Anthropic wet lab 🧬, Apple home device 🏠, senior dev death spiral 👨‍💻 Source: tldr-tech Topic: General (+AI) URL: https://tldr.tech/tech/2026-09-21 - 10 tips to improve your coding agent game Source: pulumi-blog Topic: DevOps (+AI) URL: https://www.pulumi.com/blog/10-tips-to-improve-your-coding-agent-game/ - AI Weekly Issue #531: AI labs shipped agents before agreeing how to report failures Source: ai-weekly Topic: AI (+AI Research) URL: https://aiweekly.co/issues/ai-labs-shipped-agents-before-agreeing-how-to-report ## Research - A New Kind of Model for AI Decision-Making? Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/a-new-kind-of-model-for-ai-decision-making/ - GPT-6 Astra Just Hit OpenAI's Highest Cybersecurity Risk Level Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/gpt-6-astra-just-hit-openais-highest-cybersecurity-risk-level/ - Google Offered $10M for a Dying Airline's Data. How Can You Value Yours? Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/google-offered-10m-for-a-dying-airlines-data-how-can-you-value-yours/ - Your AI Assistant Wrote the Code. Who Checked the Defaults? Source: towards-data-science Topic: AI Research (+AI, Python) URL: https://towardsdatascience.com/your-ai-assistant-wrote-the-code-who-checked-the-defaults/ - TALON: A Temporally Aware Longitudinal Framework for Radiology Report Generation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20826 - Beyond Gaussian Worlds: Latent Geometry Matters for JEPAs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21656 - How Many Humans Is a Judge Panel Worth? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21277 - Cross-Country Learning for National Infectious Disease Forecasting Using European Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.20771 - Beyond WER: Entity and Disfluency Recall in Accented Conversational ASR Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20828 - Position: A Dynamical Systems Perspective is Needed to Advance Time Series Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.16864 - Bilevel Optimization of Topology and Hyperparameters (BOTH) Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21758 - M2G-LLM: Enhancing Clinical Prediction via Multimodal Graph Reasoning and LLM Context Injection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21164 - Do Quantum Models Scale Like LLMs? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20912 - Sparse Identification for Automatic Large-Scale Screening: A Constraint-Aware Framework with Ultra Fast Decoding Algorithm Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21321 - Brownian Heads for Deep ReLU Representations: Activation Mass and the Cost of Same-Sample Selection Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21422 - Certified Topological Interaction in Neural Representations: Exact Tests and the Statistic They Require Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.08561 - Purification and Regulation: Comorbidity-Aware Multi-Label Few-Shot Learning for Medical Image Classification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21541 - When AI Reviews Train AI Reviewers: Scientific-Judgment Collapse and Mitigation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20942 - SWE-Proof: Can Language Models Resolve Real-World Issues with Machine-Checked Proofs? Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21190 - Fidel-TS: A High-Fidelity Multimodal Benchmark for Time Series Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.24789 - Generalizing Beyond Suboptimality: Offline Reinforcement Learning Learns Effective Scheduling through Random Solutions Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.10303 - Stiefel-AdamW: Geometry-Aware AdamW for Linear Factorization Blocks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21039 - TierKV: Long-Context On-Device LLMs via Predictive Multi-Tier KV Caching Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21172 - Automated Physics-Informed Neural-Networks-Based Calibration of Highly Segmented Silicon Telescopes Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20868 - Elastic Threshold Attention: Learned Contextual Sparsity for Long-Context Decoding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20888 - The critical slowing down in training diffusion models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.12597 - COMPLEX: A Closed-Form Certified Embedding of Multiparameter Persistence Modules Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22012 - $\lambda$-Controlled GRPO: Turning Flow-Matching Ratio Instability into a Budgeted Resource Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22041 - Riemannian Neural Hamiltonian Flows: Geodesic Symplectic Transport and Interpretability Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21647 - On the Limits of Maximal Coding Rate Reduction for Out-of-Distribution Generalisation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21001 - Continuous Spiking Graph Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2404.01897 - Matrix AdaGrad: Row-wise and Column-wise Adaptive Subgradient Methods Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21815 - Decision trees, Frobenius traces, and Weierstrass coefficients of elliptic curves Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.24251 - Contrastive Concept Importance: Explaining Pairwise Class Decisions Through Automatically Extracted Concept Representations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.27904 - Reconstruction of 4D Mitral Regurgitation Hemodynamics from Sparse Planar Data using Deep Operator Networks with Test-Time Adaptation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20857 - Triply-Scalable Equivariant Gaussian Process Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21085 - Fathom: Per-Query Read Depth for Sparse Decoding over Offloaded KV Caches Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.17652 - Neural ensemble Kalman filter: Data assimilation for compressible flows with shocks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.23461 - ASGARD: Action-Space Guard for UAV Resilience via Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20982 - OpenMAS-GCom. A Diagnostic Benchmark for Graph-enhanced Multi-Agent Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21527 - RegKT: Interpretable and Robust Deep Knowledge Tracing With IRT-Regularizer Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21791 - Time series generation with spectrally aligned latent flow matching Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21989 - PhysioBench: A Unified Benchmark for Physiological Signal Question Answering Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20836 - Schedule optimization for tau-leaping in masked discrete diffusion Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21960 - Programming AMD XDNA NPUs with Open-source Compiler Tools: A FlashAttention Case Study Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21264 - Data-Driven Integration Kernels for Interpretable Nonlocal Operator Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.10305 - Diverse and Adaptable Arm Coordination for Octopus-Crawling via Diffusion-Based Uncertainty-Aware Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21138 - ExpBoN: Exponential-Noise Best-of-$n$ for Efficient Test-Time LLM Alignment Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21899 - Reinforcement Learning under External Influence: Guarantees, Algorithms, and Sample Complexity Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2305.16056 - HMB-GAN: Hybrid Multi-B\'ezier GAN for Vector Shape Synthesis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21158 - Generative inversion for early ranking of competing geologic interpretations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20978 - What Must Survive? Exact Task-Information--State Frontiers for Resource-Sufficient Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21523 - dSTAR: Straggler Tolerant and Byzantine Resilient Distributed SGD Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2412.07151 - Trading Depth for Time in Recurrent Transformers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21605 - Complex Problem Solving in Large Language Models: A Statistical Control Survey and Diagnostic Framework Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20973 - Write on Paper and Get the Online Digital Trace: A New Era for Handwriting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.12702 - FootQuery: Future-Touchdown-Guided Retrieval from Depth History for Perceptive Humanoid Locomotion Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21447 - AutoRecLab: Describe the Experiment, Get the Code! Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21863 - Joint Remaining Useful Life Prediction and Capacity Estimation of Lithium-Ion Batteries Using Partial-Charging Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21932 - Taming the Adversary: A Cost-to-Disturbance Ratio Approach to Adversarial Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.12110 - Abstention and Noise Filtering: Two Missing Primitives of Softmax Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22005 - Reliability-Centered Evaluation of Sparse Longitudinal CT Lesion-Size Forecasting with Conformal Interval Calibration and Gompertz-Inspired Regularization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21197 - Dual-Interest Sequential Product Recommendation With Multi-Granular SSM Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21548 - Efficient Architecture Search under Leave-One-Subject-Out Evaluation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21457 - Particle Competition and Cooperation for Robust Graph Convolutional Network Learning Under Label Noise Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22053 - ServeGuard: Verifiable, Bounded-Residual Confinement of Operator-Invisible Channels Without Revealing the Certified Read Factor Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21515 - PRIVET: PRoximIty leakage detection Via Extreme value Theory Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.24233 - EnSol: an environment-aware graph neural network for molecular solubility prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21151 - A regret minimization approach to fixed-point iterations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.21653 - Weighted Quantum Signal Processing: Low-Depth Polynomial Approximation with Applications to Kolmogorov-Arnold Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21567 - Provably Optimal Reinforcement Learning under Safety Filtering Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.18082 - On Repulsive and Attractive Teachers: Separating Correctness from Behavior in Self-Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21561 - Teaching LLMs to Self-Evolve: Cultivating Core Meta-Skills with Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.21971 - QuanText: Protecting Dataset-Level Secrets in Textual Data Sharing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.17995 - OneBid: A Unified Auto-Bidding Foundation Model for Diverse oCPX Advertising Scenarios Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21550 - K2-V2: A 360-Open, Reasoning-Enhanced LLM Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.06201 - Modular Deep Learning Mechanisms for Auditable Next-Day Wildfire Spread Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.17763 - Optimal Randomized Proper Online Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21445 - Convex losses and their applications to SVM, SVR, and Shallow Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.14288 - Exemplar Partitioning for Mechanistic Interpretability Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.14347 - A Lightweight Plug-in Gate for Transformer-Based Time-Series Forecasters Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21044 - Predictable Failure in Multi-Hop Retrieval: Score-Distributional Confidence Scoring and Abstention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22056 - Data-free On-policy Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.14193 - Stability Enhanced Gaussian Process Variational Autoencoders Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.09331 - Multi-Objective Hyperparameter Search via Damped Gauss--Newton Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2401.03580 - Visual Navigation Transformer with Pose Attention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21212 - The Refutation Gap: Certifying Both Halves of an Optimality Claim Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20873 - Detection is solved, delineation is not: what governs tooth segmentation on panoramic radiographs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21628 - Reading Less While Writing: A Closed-Form Bandwidth Dial for Streaming Multimodal Decoders Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20845 - Diagnostic-Guided Longitudinal Modeling for Forecasting Retinal Atrophy Progression Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.16955 - Federated Deep Clustering Networks for High-Dimensional and Heterogeneous Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21829 - MACE: Memory-Agent Co-Evolution with Adaptive Memory Graphs for Multi-Agent Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21533 - Emulating Cosmic Structure Formation with a Lagrangian Neural Cellular Automaton Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.27320 - Toward individual-level calibration in affect recognition with perceptual adjustment queries Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21073 - Scaling Novel Graph Generation via Lightweight Structure-Guided Autoregressive Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.04287 - Adaptive Uncertainty-Aware Modeling and Stochastic Radial Basis Function Predictive Control for Personalized Fluid Resuscitation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21821 - Assessment of Machine Learning-Based Critical Heat Flux Models in the CTF Subchannel Code for Square Rod Bundle Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21995 - Bad Genius: Counterfactual-Guided Harness Evolution Beyond Task-Specific Shortcuts Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.18366 - IntBMoE: Integrating Block-Level Conditioning into Expert Composition for Full-Participation Mixture-of-Experts Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21346 - FedeRage: Provably Convergent Agnostic Federated Learning under General Client Drift Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21057 - Training-Adaptive Convolutional Sparse Coding via Information Bottleneck for Robust Visual Representation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.19122 - Tubular Neighbourhoods of Pfaffian Sets and Applications to Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.08370 - Planetary Prediction Engine: Autonomous Geospatial Prediction via Intelligent Data Selection and Foundation Model Embeddings Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.26088 - Prompt-Driven Exploration: Language as an Exploration Space for VLA Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.08837 - Near-Optimal Primal-Dual Algorithm for Learning Linear Mixture CMDPs with Adversarial Rewards Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.27884 - Signal-Centric Remote Sensing via Alternative Preprocessing and Acoustic Processing for ML-Driven Applications Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21123 - Learning Surrogate LPV State-Space Models with Uncertainty Quantification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.29532 - Riemannian Simultaneous Inference for Tangent Vector Field Regression Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21910 - Chronosphere: Space-Time Tessellation of Local Climate Experts Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21872 - Available Guardrails: Certifying Selective Prediction across ML Systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22048 - RheoSampling: Resolving the One-Hot Dilemma in Stochastic Dynamic-Tree Speculative Decoding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21827 - Configurable Multi-Stage Vision Pipeline for Crop Disease and Pest Diagnosis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21651 - Gradient-Stable Attention Heads Signal LLM Correctness Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.13699 - Near-Optimal Acceleration for Smooth $\ell_p$ / $\ell_q$ Nondual Convex First-Order Oracle Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21880 - Watermarkable Multi-Draft Speculative Sampling via Poisson Processes Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21858 - GEM-MPC: Balancing Exploration and Exploitation through Expert-Guided Planning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21735 - MIRCID: Inferred Hub-miRNAs Drive Cross-Task Improvements in Drug Mechanistic Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21280 - Knowledge-Graph-Augmented Chronos-2 for HEC-RAS Surrogate Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21381 - From Stress to Affect: Multimodal Deep Learning for Physiological Emotion Recognition Across Wearable Sensor Modalities Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20991 - Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21096 - The Weight Is Over - Interactive Diffusion on Consumer GPUs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21849 - Benchmarking World Models for Continual Learning on Compositional Tasks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22055 - Deep Reinforcement Learning with Buffered Quantile Objectives Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21327 - Symbolic Classification-Enabled LHC Limits for BSM Global Fits Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.22330 - Extreme classification: beating chance with one training example from each class Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20897 - Understanding Structural Representation in Foundation Models for Polymers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.11881 - Amortized Filtering and Smoothing with Conditional Normalizing Flows Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.07169 - Representation Before Training: A Practical Benchmark for Generative Medical Event Model Tokenization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.16775 - A Smoothed Discrepancy Principle for Random Feature Methods and Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21017 - IncentRL: The Trade-Off Between Preference Guidance and Task Performance Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21525 - Geometry-Aware Reinforcement Learning for 2D Irregular Nesting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.10611 - Diagonalized Attention for Individualized Regression: Latent-Row Localization and Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21320 - Boltzmann generators for amorphous particle systems Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2512.16607 - Trajectory Entropy Reinforcement Learning for Robust Robot Motor Skill Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2505.04193 - Transformers Discover Molecular Structure Without Graph Priors Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.02259 - Near-Universal Multiplicative Updates for Nonnegative Einsum Factorization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.02759 - Hiding in Plain Sight: A Diffusion-based Mitigation of Geolocation Privacy Leakage in Vision-Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21363 - Single-Loop Stochastic Projected Damped Extragradient Methods for Stochastic Nonconvex--(Strongly) Concave Minimax Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21747 - CASE: Contrastive Activation for Class-Sensitive Explanations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2506.07327 - Predictive Suppression Layers for Communication-Efficient Spiking Neural Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21583 - Probabilistic Forecasting of Business Process Executions with Neural Temporal Point Processes Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21382 - Large Language Models As Shannon Lossy Compressors Not Solomonoff Induction Estimators: The Singularity Is Not Near Without Symbolic Model Synthesis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.05280 - Toward Composable Network Digital Twins: A Subgraph-Based Latency Prediction Study Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.18704 - Divergence Timing and Cumulative Disagreement under KV-Cache Eviction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.16617 - Rollout Total Correlation for Deep Reinforcement Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2209.05333 - Kinks vs. Smoothness: Identifiability of Real Analytic nICA for Laplace-like Sources Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21926 - Offline Constrained RLHF with Multiple Preference Oracles Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.00200 - Scaling Discovery through Test-Time Communication Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21032 - LLMs as Feature Engineers for Text-and-Tabular Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21894 - Multi-Domain Clustering via Measure Quantization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21664 - Complete Neural Electronic Initialization Accelerates Materials DFT Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21759 - Sampling Reveals Style: Unsupervised, Training-Free Discovery of Prompt-Conditional Stylistic Axes in LLM Activations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.19150 - Multiplicative Optimism for Constant Regret in Games Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21976 - VLA-Precision: Asymmetric Co-Bootstrapping for Efficient Real-World Online RL of Vision-Language-Action Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.04355 - Efficient Bayes-Adaptive Reinforcement Learning with Temporal Logic Specifications Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20954 - How a Cooperative-Override Circuit Suppresses Nash Play in Large Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.27167 - Fast And Accurate Text Content File Type Identification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21306 - Intervention Granularity Matters: Coherent Treatment Bundles in Counterfactual Simulation with Clinical World Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21906 - Aggregated Posterior Predictive Checks for Generative Modeling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20999 - Bio-MF: Low-Latency and High-Fidelity EEG-to-fNIRS Cross-Modal Generation for Hybrid Motor-Imagery Brain--Computer Interfaces Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20904 - End-to-End Hard-Label Cryptanalytic Model Extraction Using Efficient Sign Recovery Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21941 - Parallelism, critical windows, and separations among diffusion language models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20539 - From Switching to Dynamic Regret: A Simple Reduction via Unbiased Random Sequences Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20968 - Causal Evidence that Language Models use Confidence to Drive Behavior Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.22161 - Recall Before Rerank: Benchmarking Deep Learning Models for Large-Scale Code-to-Code Retrieval Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.27401 - jBOT: Semantic Jet Representation Clustering Emerges from Self-Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.11719 - Robust Mixture Models for Algorithmic Fairness Under Latent Heterogeneity Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2509.17411 - Data Attribution via Sketched Metadifferentiation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.15044 - ANNLib: A Development Framework for Efficient Approximate Nearest Neighbor Search Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.17582 - MOSAIC-SR: Transformer-Guided Symbolic Regression for Scientific Equation Recovery Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20997 - Understanding In-context Learning of Addition via Activation Subspaces Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2505.05145 - SteganoBackdoor: Evading Data-Poisoning Defenses via Steganographic Backdoors Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.14301 - OverThink: Slowdown Attacks on Reasoning LLMs Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2502.02542 - Attributing Cohen's d: Training Data Attribution for Disease-Related Effects in Normative Age Biomarkers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.07729 - BrainWideBench: Benchmarking large-scale pretraining and across-animal transfer in multi-region neural recordings Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22064 - GVPO++: Group Variance Policy Optimization for LLM Post-Training and On-Policy Distillation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21432 - The Binary Tree Mechanism is Optimal for Differentially Private Continual Counting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.00876 - Rewarding Efficient Reasoning Improves Abstention on Underspecified Tasks in Reasoning Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20846 - Optimization Geometry of Equivalent Brownian RKHS Representations Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21693 - VISPATH: Visual-Intent-Guided Path Reasoning for Multimodal Knowledge Graph Question Answering Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20843 - EFQ-Softmax: Exp-Free Quantization for Softmax Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.09721 - Geometric Mean Pooling for Equal-Weight Multiplicative Coarse-Graining Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21876 - Rhamba: Region-Aware Hybrid Attention-Mamba Framework for Self-Supervised Learning in Resting-State fMRI Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.01240 - FairLMs: A Turnkey Library for Fairness in Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21296 - Beyond Benchmark Scores: Auditing Medical Vision-Language Models for Chest X-Ray Tuberculosis Screening Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21763 - Learning to Advect: A Neural Semi-Lagrangian Architecture for Weather Forecasting Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.21151 - HuRo: Robotizing Human Videos for Scalable VLA Pretraining Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.10706 - Post-Rejection Follow-up Sampling: Measuring Outcomes of Rejected Decisions in Algorithmic DEX Trading Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.08228 - Semantic Calibration Prevails Where Token Confidence Fails: Benchmarking Long-Form Scientific QA Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.00279 - BI-Agent and BI-Bench: Towards Automating End-to-End Business Intelligence Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20886 - Survival Reinforcement Learning: Toward Scalable Self-Supervised RL Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.31273 - The Impact of Semantic Pairs on Self-Supervised Representation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.08722 - Verify, Don't Trust: Agentic Model Development for Video Discovery Retrieval at Scale Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21257 - SafeStep: An Interactive Demonstration of Semantic Communication for Pedestrian Safety Monitoring Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.27688 - HERMES: Contrast-Aware Knowledge Graph Reasoning from Clinical Notes for Patient Outcome Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20825 - Multiclass Semantic Segmentation of Wildland Fire Images Using Context-Aware Centralized Copy-Paste Data Augmentation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21241 - Layerwise Decoupling for Stable Structured Sparsification of Fully Connected Layers Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21126 - Recursive Language Models Generalize Out of Domain Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20831 - Revisiting Reinforcement Learning with Verifiable Rewards from a Contrastive Perspective Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.12969 - Analysing the Linearity of Linguistic Relations in Language Model Embedding Spaces Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21655 - Periodic Neural Mapping for Unsteady Rotor-Blade Pressure and Aeroelastic Load Prediction Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21590 - Routine Blood Tests Outperform CRP for Distinguishing Bacterial From Viral Infection in Children Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21332 - CORDS: Continuous Representations of Discrete Structures Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.21583 - SpecQuant: Speculative Decoding with Multi-Parent Quantization for Adaptive LLM Inference Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21704 - Continuous Delayed-Memory Stochastic Gradient Descent and Continuous-Time Reinforcement Learning from History of Astrophysical Time Series Studies Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20906 - How do LLMs Compute Verbal Confidence Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.17839 - ISOMORPH: A Supply Chain Digital Twin for Simulation, Dataset Generation, and Forecasting Benchmarks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.12768 - GraphSkillEvo: Evolutionary Optimization of Graph-Structured Agent Skills Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21749 - Small Enough to Know Everything: The Fully-Enumerable Transformer as an Instrument for the Science of Delayed Generalization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20166 - Talk to Me, Jarvis: An Open-Source Edge-Deployable Voice Assistant Framework for Autonomous Racecars Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21109 - An Introduction to Compression-Based Machine Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21309 - Tracing the Evidence Behind Zero-Shot Time-Series Forecasting: A Source-First Taxonomy and Audit Framework Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21425 - Balance of Benchmarks: Semantic Density Reweighting for Task-Conditioned Model Comparison Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.30044 - High-probability guarantees for linear accessibility in feature superposition Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.09556 - Improving the Predictive Performance of Bootstrap Aggregating by Dirichlet Resampling Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21454 - Identifying Security Platform Product Abuse with Machine Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21303 - Fragment-Aware Vision Transformers for Fresco-Fragment Style Classification Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21012 - Neural Cellular Automata Learn General Features in their Hidden Channels Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21870 - Hybrid GPU-CPU Retrieval for Personalized Search at Ultra-Large Scale Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21281 - Learning to Move Cities: Deep Meta-Models and Reinforcement Policies for Calibration and Control in Urban Networks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21945 - TVGL-CFM:Generating and Forecasting Time-Varying Trajectories of Dynamic Networks with Conditional Flow Matching Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.16894 - Nonnegative Matrix Factorization in the Component-Wise L1 Norm for Sparse Data Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2603.29715 - Multi-Subject Pretraining Enables Short-Calibration Personalization for Closed-Corpus Surface EMG Speech Decoding Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21288 - Sparse Priors for Efficient Distribution Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20883 - Bayesian classification of astronomical spectra with class uncertainties Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21694 - Transcriptomic Models for Immunotherapy Response Prediction Show Limited Cross-cohort Generalisability Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.05478 - Decision-Focused Learning for Mean-Variance Portfolio Optimization via KKT-Based Reformulation Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21427 - Adaptive Rollout Truncation Based on Epistemic Uncertainty for Efficient Offline World Model Training Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21482 - REFINEPPO: Learning Continuous Control Policies by Iterative Action Refinement Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21108 - BEAT-Net: Injecting Biomimetic Spatio-Temporal Priors for Interpretable ECG Diagnosis Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.07316 - From Pretraining to Proficiency: Real-World Subtask RL for Long-Horizon Manipulation with Minimal Human Intervention Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21788 - Prediction Dynamics in Depth-Recurrent Language Models Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21383 - From Trainability Diagnostics to Optimization Claims: Boundaries and Controls in Variational Quantum Optimization Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21243 - Implicit Rule Induction with Test-Time Task Embeddings in ARC-like Tasks Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21181 - Guiding Agents of Quantum Games to Equilibrium using Matrix Exponential Fixed-Point Iteration Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21944 - Beyond Kinematics: Benchmarking Simulation Fidelity for Muscle-Driven Imitation Learning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21909 - Detecting Pretraining Data in Large Language Models from a Free-Energy Perspective Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21888 - Disentangling Long-Term Memory via Latent Neuro-Symbolic Reasoning Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.18461 - RACER: Role-Aligned Competence Estimation for Human-AI Routing Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21953 - Understanding LLM Quantization through Activation-Guided Compensation and Orthogonal Residuals Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21450 - Physically Based Rendering in the Latent Space Source: arxiv-cs-lg Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21054 - RankSteer: Can Pointwise LLM Rankers Be Calibrated at the Representation Level? Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2602.03422 - Do We Care About Personalization and Explainability? An Interview Study with News Recommendation Engineers Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.21547 - Compass: General Filtered Search across Vector and Structured Data Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2510.27141 - MAGIC: Marginal-Guided Compression with Optimal Transport for Efficient Visual Document Retrieval Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.21018 - IntHQ: Task-Interactive Hierarchical Query on Dual-Stream Representations for Generative Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2608.09634 - Beyond Final Answers: CRYSTAL Benchmark for Transparent Multimodal Reasoning Evaluation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2603.13099 - IntTravel: A Real-World Dataset and Generative Framework for Integrated Multi-Task Travel Recommendation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2602.11664 - Adaptive Preference Modeling via Explicit Indirect Relational Learning for Personalized Fashion Matching Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.21475 - Transferable knowledge graphs with executable learned operators for algorithm design Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2603.27922 - Auto-Bidding with Disentangled Advertiser Profiles and Train-Free Adaptation Source: arxiv-cs-ir Topic: Information Retrieval (+AI, AI Research) URL: https://arxiv.org/abs/2609.21308 - ArenaFlow: From Trajectory Ranking to Hierarchical Credit Propagation for Open-Ended Agent RL Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21378 - DiaVLo: Diagnosing Behaviours of Vision-Language Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22008 - Towards Secure Cloud-Native Computing: Unveiling Kubernetes Misconfigurations with Large Language Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20834 - MENASpeechBank: A Reference Voice Bank with Persona-Conditioned Multi-Turn Conversations for AudioLLMs Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.07036 - Omni Demand Understanding: A Benchmark for Contextual User-Intent Inference in Multimodal Interaction Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21392 - Per-Aetiology Contrastive Severity Embeddings with Phonological Pseudo-Labelling for Multilingual Dysarthric Speech Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21789 - I'll Keep an Ear Out: Teaching AudioLLMs Proactive Audio Assistance Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21183 - Hierarchical attention interpretation: an interpretable speech-level transformer for bi-modal depression detection Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2309.13476 - When Steering Fails in Latent Reasoning: A Latent-to-Language Transition Gap Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21662 - CASCADE Against Jailbreaks: Combination Across Stages with Controlled Attack-Defense Evaluation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21793 - Do small language models know what they don't know? Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20824 - World Modeling in Transformers Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21748 - Knowledge-Graph Based Augmentation versus Retrieval Augmented Generation for Cultural-Related Question Answering Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.18317 - Listen Before You Speak: Response Planning from Listener Facial Reactions for Conversational Speech Generation Source: arxiv-cs-cl Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.21683 - MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic Tasks Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2602.16313 - Enhancing Audio Reasoning via Semantic Summary Prediction Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20849 - MemeLens: Multilingual Multitask VLMs for Memes Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.12539 - Generative Artificial Intelligence Chatbots for Motivational Interviewing: A Scoping Review From System Design to Intervention Outcomes Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20902 - Trustworthy FinAInce: Unpacking How AI-Mediated Financial Advice is Judged Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20989 - CORTEX: High-Quality Cross-Domain Organization of Web-Scale Corpora through Ontological Corpus Graph Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.30175 - Conformal Privacy Auditing: Calibrated Re-identification Attacks with Statistical Guarantees Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21340 - Steering LLMs Responses Towards Moral Foundations on the Norwegian MFQ-30 Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21636 - Fine PT-PT Web: A High-Quality 41 Billion Tokens Data Collection of the European Portuguese Web Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.07699 - Large Language Model Agents for Evidence Based Genetic Disease Severity Classification Source: arxiv-cs-cl Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.19569 - CIBuzzBench: A Benchmark for Cross-Lingual Understanding of Chinese Internet Buzzwords Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21722 - From Generation to Detection: Exploration of Discourse Driven Scenario based LLM Generated Fake News Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20838 - Git-Assistant: Planning-Based Support for Updating Git Repositories Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2607.09224 - Explainable Multimodal Aspect-Based Sentiment Analysis with Dependency-guided Large Language Model Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2601.06848 - The Hidden Cost of Digits: Number Normalization and WER in ASR Systems Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21084 - Moral Entropy: Auditing Bias and Uncertainty in Moral Judgment Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21992 - Curriculum-Based Noise Adaptation for Phoneme-to-Text Reconstruction in Visual Speech Recognition Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20839 - PersonalAI 2.0: Enhancing knowledge graph traversal/retrieval with planning mechanism for Personalized LLM Agents Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.13481 - Talking Past the Machine: Morality, Politeness, and Alignment in Human-AI Dialogue Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21401 - Reviser: Revision-Capable Text Generation via Autoregressive Cursor Actions Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20830 - From Memory to Behavior: A Behavior-Aware Role-Playing Framework for Social Media Influencers Source: arxiv-cs-cl Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.21349 - GameLogicBench: Evaluating Coding Agents on Runtime Game Logic with Tick-Level State Assertions Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21562 - From Discharge Notes to Patient Understanding: Persona-Grounded, Open-Ended Simulation of LLMs as Discharge Educators Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20827 - CoLearn: An Agentic Tutor that Learns its Learner in a Human--AI Co-Learning Loop Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21154 - Not All Irregularity Is Equal: Causally Isolating a Rare Failure Mode in Japanese Morphological Inflection Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21179 - Playing log(N)-Questions over Wikipedia Abstracts: How Per-Round Errors Compound Under Information Asymmetry Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.19113 - A Generative Grammar Underlying the Voynich Manuscript, the Pastiche Hypothesis: Evidence from Large Language Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20835 - An Interpretable Memory Decision Controller for LLM Agents Based on Three-Signal Complementarity: Decoupling Confidence and Consistency Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22043 - Consistent Relexicalization of Clinical Documents using Graph-Based Approach Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21387 - TabScope: Question-Adaptive Scope Selection for Table Question Answering Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.03395 - Benchmarking Gender Bias in Machine Translation Evaluation Metrics across Occupations Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21490 - When Does Reasoning Help in Machine Translation? A Hierarchical Analysis of LRM Reasoning Traces Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21247 - Beyond Reference-Based Evaluation: Reward Models for Meta-Evaluation of Grammatical Error Correction Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21231 - COAL-SQL: Coverage-Guided Augmentation and Failure-Driven Learning for Text-to-SQL Post-Training Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20842 - Aligning with Lived Experience: Heterogeneous Benefits of Fine Tuning in Mental Health Support Generation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21075 - Cultural Alignment in Large Language Models Using Soft Prompt Tuning Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2503.16094 - A Ticket from Marginals to Joints: Coupled-Noise Distillation for One-Step Block Generation in Diffusion Language Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.06324 - MIRAGE: Multi-Perspective Creative Language Model Reasoning with Reinforcement Learning Guidance Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21554 - The Role of Fine-grained Harm Signals in LLM Safety Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.19366 - MME-Safety: A Fine-grained Benchmark for Safety Evaluation of MLLMs Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20850 - Draft-OPD: On-Policy Distillation for Speculative Draft Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.29343 - SEA-LION-v4.8: A Technical Report Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.18310 - TrialAtlas: Multi-Agent Research Organization for Clinical Trial Design and Optimization Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21859 - Clinician-Grounded Quality Assurance for AI-Assisted Psychiatric Intake Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21149 - TatBLiMP: A Benchmark of Linguistic Minimal Pairs for Tatar Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20832 - Sixteen models, fewer than two voices: measuring ensemble dispersion where no answer is uniquely correct Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.00285 - Do Personality-Tuned LLMs Make Better Social Agents? Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21857 - When Better Turns Do Not Make Better Agents: Diagnosing the Gap Between Next-Turn Metrics and Workflow Success Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21187 - Reusing Latent Speech Representations for Query-Conditioned Topic Localization in Transcripts Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21844 - MOSCOPT: Mixture-of-Skills Collective Optimization for LLM Agents Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.14399 - Why Do LLMs Struggle in Strategic Play? Broken Links Between Observations, Beliefs, and Actions Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.00226 - Cross-sector generalization of accident-process role classification in occupational accident narratives Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22081 - Measuring Digital Labour Market Transitions with a Digital Semantic Score: An AI-Based Methodology Applied to the Dutch Labour Market Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.24222 - A primer on evaluation methods for large language models in healthcare Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.14819 - Dynamic Lagging using Stable-Prefix Training for Simultaneous Translation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.05799 - Scaling Forced Alignment to End-User Devices Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21145 - Voice-Light: A Full-Duplex Cascaded Voice Agent with Causal Turn-Taking and Speculative Generation Source: arxiv-cs-cl Topic: AI Research (+AI, JavaScript, React) URL: https://arxiv.org/abs/2609.20995 - Transsion's Speaker-Attributed Multilingual ASR System for the MLC-SLM 2026 Challenge Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20833 - NemotronLabs VoiceChat: An Open Full-duplex Speech-to-Speech Model with Tool Calling Capabilities Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21967 - Ripple-Pivot Search: Active Parallel Decoding for Diffusion Large Language Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.11742 - Cross-Lingual Parkinson's Disease Severity Assessment Using Pre-trained Speech Embeddings: A Multi-Class Evaluation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20875 - Geometry of Values: Task Vector Composition for Ethical Preference Alignment in Language Models Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21094 - Hallucination-R1: Robustness-Oriented Paraphrase Generation for Factual Consistency Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21227 - The Spoken Wikipedia Presentation Corpus Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21676 - $\mu^2$-Bench: A Multilingual Machine Unlearning Benchmark Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20945 - Rethinking Human-Aligned Evaluation: An Analysis of Semantic Metrics Beyond WER Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21663 - Souper-Model: How Simple Arithmetic Unlocks State-of-the-Art LLM Performance Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2511.13254 - Evaluating In-Context Learning and Retrieval Strategies for Devanagari Post-OCR Correction Source: arxiv-cs-cl Topic: AI Research (+AI, Information Retrieval) URL: https://arxiv.org/abs/2609.21595 - SAGE: Schema-Guided LLMs for Grant Review Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20829 - The Functionalizer: Lossless Functional Decomposition for Subword Tokenization Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.15991 - Auditing a KB Elicitation of Frontier LLM Knowledge: A Multi-dimensional Analysis of GPTKB v1.5 Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2510.07024 - PRISM-BN: A Controlled Corpus and Benchmark for Text-to-Parameterized Bayesian Network Extraction Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21673 - Sometin Beta Pass Notin: Improving Multilingual ASR for Nigerian Languages via Knowledge Distillation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2605.17710 - Chinese Competitive Debating Dataset and Benchmark Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21637 - MemAudit: Auditing Long-Term Agent Memory via Hidden User-State Recovery Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2606.24595 - Lessons Without Borders? Evaluating Cultural Alignment of LLMs Using Multilingual Story Moral Generation Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.08797 - VQ-Logits: Compressing the Output Bottleneck of Large Language Models via Vector Quantized Logits Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2505.10202 - Beyond Atomic Tokens: Factorizing Syllables for Language Model Pretraining Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21362 - Mind the Gap: Theory-of-Mind-Grounded Friction for Epistemic Alignment Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.30719 - Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.27409 - Wiktionary as a Crowdsourced Lexicon for English Dialects Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.15641 - RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22000 - Boosting Deepresearch and LongContext Ability with Self-Generated Deepresearch Rollouts Traces Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20844 - Reading Anxiety or Reading the Label? Comparing Fine-Tuned and Frontier Models for Anxiety Detection on Social Media Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20847 - Offline Multimodal Large Language Models for Decision Support in Air Operations Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21390 - JudgeSense: A Benchmark for Prompt Sensitivity in LLM-as-a-Judge Systems Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2604.23478 - SG-Mamba: Sparse Graph-Guided Mamba for Audio-Visual Speech Enhancement Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.18009 - Are You Sure You're Sure? On the Impact of Instruction Tuning on Confidence and Lexical Diversity Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.13430 - Do New Attention Mechanisms Actually Fix Attention Sinks at Million-Token Context? Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.08574 - Multi-turn Conversational AI from Text to Multimodal Interaction: Data, Models, Evaluation, and Open Challenges Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2608.17605 - What Does Privileged Information Add to On-Policy Self-Distillation? Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.20612 - Accelerating Dense LLMs via L0-regularized Mixture-of-Experts Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21672 - From Task Success to Productive Success: Evaluating Human-AI Collaboration by Quality and Cost Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21117 - QuranicMMLU: A Cognitively-Aware Benchmark for Evaluating Generative AI Solutions on Quranic Linguistic Knowledge Source: arxiv-cs-cl Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.22038 - The Communication Bottleneck: A Round-Trip Study of Tree-Structured Expression Serialization in Language Models Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21509 - GUARD: Natural Forgetting in Large Reasoning Models via Guided Answer-Reasoning Distillation Source: arxiv-cs-ai Topic: AI Research (+AI) URL: https://arxiv.org/abs/2609.21677 ## Community - Jev introduces a new shape of LLM - System One, aka Decision Models Source: simon-willison Topic: Python (+AI, Django) URL: https://simonwillison.net/2026/Sep/21/jev/ - Cloudflare Python Workers are now generally available Source: simon-willison Topic: Python (+AI, Django) URL: https://simonwillison.net/2026/Sep/21/cloudflare-python-worker/ - Anthropic, OpenAI et al. face antitrust suit for agreeing to slow AI development Source: hacker-news Topic: General (+AI) URL: https://www.tomshardware.com/tech-industry/big-tech/anthropic-openai-spacexai-and-google-face-antitrust-lawsuit-for-agreeing-to-slow-ai-development-plaintiffs-say-plan-has-been-in-motion-for-months-before-calls-agreement-self-serving - A Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data Source: lobsters Topic: General (+AI) URL: https://github.com/volotat/mini-AGI/ - Creating a Blog in Gemini:// Source: lobsters Topic: General (+AI, Web) URL: https://brennan.day/creating-a-blog-in-gemini/ - The president wanted to unleash AI. Then came Anthropic Source: hacker-news Topic: General (+AI) URL: https://www.politico.com/news/magazine/2026/09/20/anthropic-white-house-ai-01085212 - Show HN: Mahabharata Characters – 307 figures with gunas and decision primitives Source: hacker-news Topic: General (+AI) URL: https://huggingface.co/datasets/gnumanth/mahabharata-characters ## Media - An ex-OpenAI researcher just deleted language from the LLM... Source: youtube-fireship Topic: Web (+AI, General, JavaScript) URL: https://www.youtube.com/watch?v=TbkUKCm3CHQ