AI News

Observability for Delegated Execution in Agentic AI Systems (arxiv.org)

An 84-Format Numeric Catalog with Bit-Exact Conformance Vectors: A Vendor-Neutral Reference for FP8, BF16, MXFP4, and Microscaling Formats (arxiv.org)

A Graphop Analysis of Graph Neural Networks on Sparse Graphs: Generalization and Universal Approximation (arxiv.org)

Foundation Inference Models for Ordinary Differential Equations (arxiv.org)

On the Superlinear Relationship between SGD Noise Covariance and Loss Landscape Curvature (arxiv.org)

The Label Horizon Paradox: Rethinking Supervision Targets in Financial Forecasting (arxiv.org)

Your Self-Play Algorithm is Secretly an Adversarial Imitator: Understanding LLM Self-Play through the Lens of Imitation Learning (arxiv.org)

When Benign Inputs Lead to Severe Harms: Eliciting Unsafe Unintended Behaviors of Computer-Use Agents (arxiv.org)

Variational Speculative Decoding: Rethinking Draft Training from Token Likelihood to Sequence Acceptance (arxiv.org)

Optimal Fair Aggregation of Crowdsourced Noisy Labels using Demographic Parity Constraints (arxiv.org)

Latent Spherical Flow Policy for Reinforcement Learning with Combinatorial Actions (arxiv.org)

Cosmo3DFlow: Wavelet Flow Matching for Spatial-to-Spectral Compression in Reconstructing the Early Universe (arxiv.org)

Nonparametric LLM Evaluation from Preference Data (arxiv.org)

Alcmean's: Unsupervised community detection using local Laplacian, automatic detection of the number of centers (arxiv.org)

MilliVid: Hierarchical Latents for Long-Range Consistency in Video Generation (arxiv.org)

Families of Control-Cost-Parametrized Inverse-Optimal Universal Stabilizers (arxiv.org)

MEnvAgent: Scalable Polyglot Environment Construction for Verifiable Software Engineering (arxiv.org)

UA-DCM: Uncertainty-aware Causal Decision Making via Effect Bound Decomposition (arxiv.org)

Mechanistic Data Attribution: Tracing the Training Origins of Interpretable LLM Units (arxiv.org)

Comparative evaluation of training strategies using partially labelled datasets for segmentation of white matter hyperintensities and stroke lesions in FLAIR MRI (arxiv.org)

A systematic investigation of molecular encoding methods for drug property predictions across neural network and Transformer encoder-based model (arxiv.org)

Hardening Agent Benchmarks with Adversarial Hacker-Fixer Loops (arxiv.org)

RAM: Reachability Across Morphologies (arxiv.org)

Fourier Neural Operators with rank-1 lattice points and hyperbolic cross (arxiv.org)

CHROMA: Detecting AI-Generated Images through Inter-Channel Color-Space Correlations (arxiv.org)

Structure-Aware Modeling of Multiple-Choice Questions Improves Automatic Difficulty Estimation (arxiv.org)

Meeting SLOs, Slashing Hours: Automated Enterprise LLM Optimization with OptiKIT (arxiv.org)

Generative Reasoning Re-ranker (arxiv.org)

CURE: Curriculum-guided Multi-task Training for Reliable Anatomy Grounded Report Generation (arxiv.org)

XCR-Bench: Benchmarking Cross-Cultural Reasoning in LLMs via Culture-Specific Items and Hall's Triad (arxiv.org)

Multimodal Generative Engine Optimization: Rank Manipulation for Vision-Language Model Rankers (arxiv.org)

CLONE: A 3DGS-Based Closed-Loop Differentiable Optimization Framework for Single-Image Normal Estimation (arxiv.org)

AMix-1: A Pathway to Test-Time Scalable Protein Foundation Model (arxiv.org)

Failure by Interference: Language Models Make Balanced Parentheses Errors When Faulty Mechanisms Overshadow Sound Ones (arxiv.org)

ACTIVE-o3: Empowering MLLMs with Active Perception via Pure Reinforcement Learning (arxiv.org)

From A to B to A: Palindromic Zero-Shot Voice Conversion with Non-Parallel Data (arxiv.org)

IR-SIM: A Lightweight Skill-Native Simulator for Navigation, Learning, and Benchmarking (arxiv.org)

Compositional Approximation Can Strictly Outperform Superpositional Approximation (arxiv.org)

Discovering and decoding latent mean-field structure with variational autoencoders (arxiv.org)

Balancing Real and Synthetic Data for CNN-based Masonry Crack Detection (arxiv.org)

Harmonia: End-to-End RAG Serving Optimization (arxiv.org)

Robust Renal Mass Segmentation on CT: A Validation Study of an AI-Based Framework (arxiv.org)

Deep Tree Tensor Networks (arxiv.org)

FIT-Print: Towards False-claim-resistant Model Ownership Verification via Targeted Fingerprint (arxiv.org)

Complement or substitute? How AI increases the demand for human skills (arxiv.org)

Strategic Integration of Artificial Intelligence in the C-Suite: The Role of the Chief AI Officer (arxiv.org)

Rule-based autocorrection of Piping and Instrumentation Diagrams (P&IDs) on graphs (arxiv.org)

GRPO Does Not Close the Multi-Agent Coordination Gap (arxiv.org)

Latent Structural Categorical Matrix Completion with Application to Quasispecies Analysis (arxiv.org)

Toward autocorrection of chemical process flowsheets using large language models (arxiv.org)