On Robustness and Chain-of-Thought Consistency of RL-Finetuned VLMs - 惯性聚合

推荐订阅源

博客园 - Franky

Proofpoint News Feed

Palo Alto Networks Blog

Hackread – Cybersecurity News, Data Breaches, AI and More

The Register - Security

Security @ Cisco Blogs

Simon Willison's Weblog

Recorded Future

大猫的无限游戏

Microsoft Research Blog - Microsoft Research

Tailwind CSS Blog

cs.CV updates on arXiv.org

Cyber Security Advisories - MS-ISAC

Application and Cybersecurity Blog

True Tiger Recordings

有赞技术团队

Cisco Talos Blog

Hacker News - Newest: "LLM"

The GitHub Blog

cs.AI updates on arXiv.org

博客园 - 叶小钗

Hugging Face - Blog

Hacker News: Ask HN

Security Archives - TechRepublic

Future of Privacy Forum

PCI Perspectives

Help Net Security

让小产品的独立变现更简单 - ezindie.com

The Blog of Author Tim Ferriss

Netflix TechBlog - Medium

罗磊的独立博客

Apple Machine Learning Research

Security Latest

美团技术团队

博客园 - 三生石上(FineUI控件)

Schneier on Security

CERT Recently Published Vulnerability Notes

cs.LG updates on arXiv.org

Reinforced Graph of Thoughts: RL-Driven Adaptive Prompting for LLMs ConTact: Contact-First Antibody CDR Design via Explicit Interface Reasoning Predicting Performance of Symbolic and Prompt Programs with Examples Local Covariate Selection for Average Causal Effect Estimation without Pretreatment and Causal Sufficiency Assumptions How Many Different Outputs Can a Transformer Generate? Prototype-Guided Classification Sub-Task Decoupling Framework: Enhancing Generalization and Interpretability for Multivariate Time Series CoFEH: LLM-driven Feature Engineering Empowered by Collaborative Bayesian Hyperparameter Optimization Can Transformers Learn to Verify During Backtracking Search? CausalGuard: Conformal Inference under Graph Uncertainty Representation Gap: Explaining the Unreasonable Effectiveness of Neural Networks from a Geometric Perspective Noise Schedule Design for Diffusion Models: An Optimal Control Perspective Skill Weaving: Efficient LLM Improvement via Modular Skillpacks The Illusion of Reasoning: Exposing Evasive Data Contamination in LLMs via Zero-CoT Truncation Discovering Entity-Conditioned Lag Heterogeneity: A Lag-Gated Neural Audit Framework for Panel Time Series Machine learning prediction of obstructive coronary artery disease using opportunistic coronary calcium and epicardial fat assessments from CT calcium scoring scans Scalable On-Policy Reinforcement Learning via Adaptive Batch Scaling Detecting Atypical Clients in Federated Learning via Representation-Level Divergence Memory-R2: Fair Credit Assignment for Long-Horizon Memory-Augmented LLM Agents Tabular foundation models for robust calibration of near-infrared chemical sensing data LABO: LLM-Accelerated Bayesian Optimization through Broad Exploration and Selective Experimentation Short-Term-to-Long-Term Memory Transfer for Knowledge Graphs under Partial Observability ECPO: Evidence-Coupled Policy Optimization for Evidence-Certified Candidate Ranking Evaluation of Pipelines for Data Integration into Knowledge Graphs Tailoring Teaching to Aptitude: Direction-Adaptive Self-Distillation for LLM Reasoning SCI-Defense: Defending Manipulation Attacks from Generative Engine Optimization Alike Parts: A Feature-Informed Approach to Local and Global Prototype Explanations Three Costs of Amortizing Gaussian Process Inference with Neural Processes Quantitative coronary calcification analysis for prediction of myocardial ischemia using non-contrast CT calcium scoring Aerodynamic force reconstruction using physics-informed Gaussian processes Objective-Induced Bias and Search Dynamics in Multiobjective Unsupervised Feature Selection Support-aware offline policy selection for advertising marketplaces Embedding-Based Federated Learning with Runtime Governance for Iron Deficiency Prediction LCGuard: Latent Communication Guard for Safe KV Sharing in Multi-Agent Systems Graph neural network explanations reveal a topological signature of disease-associated hubs in biological networks Visibility nowcasting in South Korea: a machine learning approach to class imbalance and distribution shift MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems EmoTrack: Robust Depression Tracking from Counseling Transcripts across Session Regimes Provable Joint Decontamination for Benchmarking Multiple Large Language Models TONIC: Token-Centric Semantic Communication for Task-Oriented Wireless Systems Position: The Time for Sampling Is Now! Charting a New Course for Bayesian Deep Learning Expectation Consistency Loss: Rethink Confidence Calibration under Covariate Shift Beyond Euclidean Proximity: Repairing Latent World Models with Horizon-Matched Trajectory Reachability Metrics ChronoMedicalWorld: A Medical World Model for Learning Patient Trajectories from Longitudinal Care Data Dynamic Mixture of Latent Memories for Self-Evolving Agents Protein Thoughts: Interpretable Reasoning with Tree of Thoughts and Embedding-Space Flow Matching for Protein-Protein Interaction Discovery AgForce Enables Antigen-conditioned Generative Antibody Design Models Can Model, But Can't Bind: Structured Grounding in Text-to-Optimization When to Switch, Not Just What: Transition Quality Prediction in Clash Royale Leveraging Self-Paced Curriculum Learning for Enhanced Modality Balance in Multimodal Conversational Emotion Recognition $\textit{BlockFormer}$ : Transformer-based inference from interaction maps Ex-GraphRAG: Interpretable Evidence Routing for Graph-Augmented LLMs One-Way Policy Optimization for Self-Evolving LLMs PEARL: Unbiased Percentile Estimation via Contrastive Learning for Industrial-Scale Livestream Recommendation IKNO: Infinite-order Kernel Neural Operators Holomorphic Neural ODEs with Kolmogorov-Arnold Networks for Interpretable Discovery of Complex Dynamics I-SAFE: Wasserstein Coherence Metrics for Structural Auditing of Scientific AI Models How Sparsity Allocation Shapes Label-Free Post-Pruning Recoverability ARC-STAR: Auditable Post-Hoc Correction for PDE Foundation Models Compiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost Engineering Hybrid Physics-Informed Neural Networks for Next-Generation Electricity Systems: A State-of-the-Art Review Toward Understanding Adversarial Distillation: Why Robust Teachers Fail What are the Right Symmetries for Formal Theorem Proving? Bandit Convex Optimization with Gradient Prediction Adaptivity The Attribution Impossibility: No Feature Ranking Is Faithful, Stable, and Complete Under Collinearity Measuring Cross-Modal Synergy: A Benchmark for VLM Explainability Double descent for least-squares interpolation on contaminated data: A simulation study Cross-domain benchmarks reveal when coordinated AI agents improve scientific inference from partial evidence TBP-mHC: full expressivity for manifold-constrained hyper connections through transportation polytopes On the Sample Complexity of Discounted Reinforcement Learning with Optimized Certainty Equivalents Harnesses for Inference-Time Alignment over Execution Trajectories PeakFocus: Bridging Peak Localization and Intensity Regression via a Unified Multi-Scale Framework for Electricity Load Forecasting When Are Teacher Tokens Reliable? Position-Weighted On-Policy Self-Distillation for Reasoning Thermodynamic Irreversibility of Training Algorithms AutoMCU: Feasibility-First MCU Neural Network Customization via LLM-based Multi-Agent Systems Manifold-Guided Attention Steering Frequency-Domain Regularized Adversarial Alignment for Transferable Attacks against Closed-Source MLLMs Equilibrium Propagation and Hamiltonian Inference in the Diffusive Fitzhugh-Nagumo Model OPPO: Bayesian Value Recursion for Token-Level Credit Assignment in LLM Reasoning One LR Doesn't Fit All: Heavy-Tail Guided Layerwise Learning Rates for LLMs Symbolic Density Estimation for Discrete Distributions On-Policy Consistency Training Improves LLM Safety with Minimal Capability Degradation Reasoning through Verifiable Forecast Actions: Consistency-Grounded RL for Financial LLMs Same Architecture, Different Capacity: Optimizer-Induced Spectral Scaling Laws Algebraic Machine Learning for Small-to-Medium Datasets Is Competitive against Strong Standard Baselines CASE-NET: Deep Spatio-Temporal Representation Learning via Causal Attention and Channel Recalibration for Multivariate Time Series Classification stable-worldmodel: A Platform for Reproducible World Modeling Research and Evaluation An Improved Adaptive PID Optimizer with Enhanced Convergence and Stability for Deep Learning Correcting Class Imbalance in Prior-Data Fitted Networks for Tabular Classification Calibration, Uncertainty Communication, and Deployment Readiness in CKD Risk Prediction: A Framework Evaluation Study Provable Robustness against Backdoor Attacks via the Primal-Dual Perspective on Differential Privacy Beyond Single Slot: Joint Optimization for Multi-Slot Guaranteed Display Advertising Optimal Guarantees for Auditing Rényi Differentially Private Machine Learning MMD-Balls as Credal Sets: A PAC-Bayesian Framework for Epistemic Uncertainty in Test-Time Adaptation From Sequential Nodes to GPU Batches: Parallel Branch and Bound for Optimal $k$-Sparse GLMs Dropout Universality: Scaling Laws and Optimal Scheduling at the Edge-of-Chaos Temporal Contrastive Transformer for Financial Crime Detection: Self-Supervised Sequence Embeddings via Predictive Contrastive Coding DualOptim+: Bridging Shared and Decoupled Optimizer States for Better Machine Unlearning in Large Language Models A Reproducible Log-Driven AutoML Framework for Interpretable Pipeline Optimization in Healthcare Risk Prediction Can Breath Biomarkers Causally Influence Blood Glucose? Investigating VOC-Mediated Modulation in Diabetes Beyond Scalar Objectives: Expert-Feedback-Driven Autonomous Experimentation for Scientific Discovery at the Nanoscale

On Robustness and Chain-of-Thought Consistency of RL-Finetuned VLMs

Rosie Zhao, · 2026-05-23 · via cs.LG updates on arXiv.org

此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。原文来自 — 版权归原作者所有。