Preserving Long-Tailed Expert Information in Mixture-of-Experts Tuning

RelayS2S: A Dual-Path Speculative Generation for Real-Time Dialogue

Non-Stationarity Breaks Permutation Surrogates in Multi-Agent Reinforcement Learning: Diagnosis and Remedies

Zero-shot World Models Are Developmentally Efficient Learners

"What Are You Really Trying to Do?": Co-Creating Life Goals from Everyday Computer Use

EVA-Bench: A New End-to-end Framework for Evaluating Voice Agents

From Scores to Evidence: Auditable Decisions Can Improve Speech Deepfake Detection

CoGReV: A Confidence-Gated Post-Hoc Non-Monotonic Belief Revision Framework for Phishing Website Classification

Dear Algo: A Precision-First Agentic Intent Layer for Unified Search and Recommendation

EVOQUANT: Self-Evolving Verifier-Guided Strategy Optimization for Robust Quantitative Trading

HoneyRoute: Honeypot-Model Routing for Adversarial LLM Serving

Physics of Agents: Statistical Mechanics Predicts Collective Behavior of AI Agents

Multi-dimensional Preference Alignment by Conditioning Reward Itself

Let ViT Speak: Generative Language-Image Pre-training

RadJEPA: Radiology Encoder for Chest X-Rays via Joint Embedding Predictive Architecture

Improving Semantic Uncertainty Quantification in LVLMs with Semantic Gaussian Processes

LSP-DETR: Efficient and Scalable Nuclei Segmentation in Whole-Slide Images

From Landslide Conditioning Factors to Satellite Embeddings: Evaluating the Utilisation of Google AlphaEarth for Landslide Susceptibility Mapping using Deep Learning

Federated Learning for Cross-Modality Medical Image Segmentation via Augmentation-Driven Generalization

Semi-Supervised Domain Adaptation with Latent Diffusion for Pathology Image Classification

VideoTIR: Accurate Understanding for Long Videos with Efficient Tool-Integrated Reasoning

SimpleProc: Fully Procedural Synthetic Data from Simple Rules for Multi-View Stereo

Copying explains the collective behavior of AI agents in the wild

TimeCues Studio: A Workspace for Music Annotation and Algorithm Prototyping

Dynamical Non-compensatory Multidimensional IRT Model Using Variational Approximation

Physics-Informed Multi-Task Surrogate Model for the Martian Nightside Thermosphere

Are You Learning Biological Signal or Shortcuts? Auditing and Mitigating Bias in Protein-Protein Interaction Datasets

A Sharp Barrier for Consistent Submodular Maximization: Any Improvement over $2-\sqrt{2}$ Entails Exponential Queries or Linear Recourse

The Sample Complexity of Quantum Entanglement Allocation

Structural Fusion of Bayesian Networks with Limited Treewidth Using Genetic Algorithms

Maverick: Private and Verifiable LLM Inference Made Practical via Matrix-Vector Multiplication Delegation

Zero-Shot Temporal Localisation of Audio Deepfakes in Multi-Speaker Conversations

Influence-Oriented Personalized Federated Learning

Incentives to Offer Algorithmic Recourse

Safe Learning Under Irreversible Dynamics via Asking for Help

Orukeet: Multilingual ASR with Frozen Gabor Kernels

Query Brand Entity Linking in E-Commerce Search

A Taxonomy of Architecture Options for Foundation Model-based Agents: Analysis and Decision Model

Efficient Diversity-based Experience Replay for Deep Reinforcement Learning

FrogNano: Training a 4B Coding Agent via Online Task Synthesis

RevalExo: A Functional Daily-Activity Benchmark for Inertial and Visual Locomotion Mode Recognition in Older Adults and Clinical Cohorts

Equity Promotion in Online Resource Allocation

EvolveScaler: Synthesizing Information-Evolution Contexts via Executable State Machines and Natural-Language Rendering

Predicting Estimated Times of Restoration for Electrical Outages Using Longitudinal Tabular Transformers

[AINews] 今天没什么大事发生

今日AI领域风平浪静，无重大新闻。

@haozhangml：首席AI官很快将变成首席重置官

首席AI官正迅速转型为首席重置官，主导组织重置。

在无人关注的角落，Meta 也重置了

Meta在鲜有人注意处也进行了重置。

@AravSrinivas：😍

（原文无实际内容，仅含表情符号，无法提炼摘要）

埃隆·马斯克：@Grok @Bot 几乎每天都在变得更好

马斯克称，@Grok @Bot 几乎每天都在进步。

@elonmusk：特斯拉虚拟电厂电网稳定系统

特斯拉虚拟电厂系统用于稳定电网。