All entries

Weak-to-Strong Elicitation via Mismatched Wrong Drafts

arXiv:2605.17314v1 Announce Type: cross Abstract: We consider whether off-policy experience from a smaller, weaker model can elicit capability in a stronger learner that on-policy RL fine-tuning (e.g.

Applications

Weakly Supervised Cross-Modal Learning for 4D Radar Scene Flow Estimation

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

DGX agent

arXiv:2605.18507v1 Announce Type: new Abstract: Due to the difficulty of obtaining ground-truth data for 4D radar scene flow estimation, previous methods typically rely on either self-supervised losse

applicationsarxiv-cs-cv

19 May 2026

WebGameBench: Requirement-to-Application Evaluation for Coding Agents via Browser-Native Games

arXiv:2605.17637v1 Announce Type: new Abstract: Coding agents are increasingly used as application builders, yet many evaluations still focus on source code, repository-level tests, or intermediate tr

WEBSERV: A Full-Stack and RL-Ready Web Environment for Training Web Agents at Scale

model-releasesarxiv-cs-cl

arXiv:2510.16252v2 Announce Type: replace-cross Abstract: Reinforcement learning (RL) for web agents demands environments that are both effective for evaluation and efficient enough for large-scale on

Weighted Flow Matching and Physics-Informed Nonlinear Filtering for Parameter Estimation in Digital Twins

model-releasesarxiv-cs-lg

arXiv:2605.17146v1 Announce Type: cross Abstract: Digital twins (DTs) rely on continuous synchronization between physical systems and their virtual counterparts through online parameter estimation und

Weighted Reverse Convolution for Feature Upsampling

arXiv:2605.17472v1 Announce Type: new Abstract: Pre-trained vision foundation models (VFMs) provide strong semantic representations, yet their patch-level features are inherently coarse, limiting thei

researcharxiv-cs-cv

Weisfeiler and Leman Follow the Arrow of Time: Expressive Power of Message Passing in Temporal Event Graphs

arXiv:2505.24438v3 Announce Type: replace Abstract: An important characteristic of temporal graphs is how the directed arrow of time influences their causal topology, i.e., which nodes can possibly in

WELD: The First Naturalistic Long-Period Small-Team Workplace Emotion Dataset for Ubiquitous Affective Computing

arXiv:2510.15221v2 Announce Type: replace Abstract: Affective computing has matured rapidly in laboratory settings, yet no prior dataset combines (i) months-to-years of duration, (ii) a naturalistic w

What Does the AI Doctor Value? Auditing Pluralism in the Clinical Ethics of Language Models

arXiv:2605.18738v1 Announce Type: new Abstract: Medicine is inherently pluralistic. Principles such as autonomy, beneficence, nonmaleficence, and justice routinely conflict, and such ethical dilemmas

Applications

What Drives Success in Physical Planning with Joint-Embedding Predictive World Models?

arXiv:2512.24497v3 Announce Type: replace Abstract: A long-standing challenge in AI is to develop agents capable of solving a wide range of physical tasks and generalizing to new, unseen tasks and env

applicationsarxiv-cs-ai

What is Holding Back Latent Visual Reasoning?

arXiv:2605.18445v1 Announce Type: cross Abstract: Humans can approach complex visual problems by mentally simulating intermediate visual steps, rather than reasoning through language alone. Inspired b

What is the long-run distribution of stochastic gradient descent? A large deviations analysis

arXiv:2406.09241v3 Announce Type: replace-cross Abstract: In this paper, we examine the long-run distribution of stochastic gradient descent (SGD) in general, non-convex problems. Specifically, we see

What Matters for Grocery Product Retrieval with Open Source Vision Language Models

arXiv:2605.18029v1 Announce Type: new Abstract: Multimodal product retrieval (MPR) underpins checkout-free retail and automated inventory systems, yet it demands fine-grained SKU discrimination that s

researcharxiv-cs-cv

When a Zero-Shooter Cheats: Improving Age Estimation via Activation Steering

arXiv:2605.17658v1 Announce Type: new Abstract: Different age-related regulations have been proposed to protect minors from harmful content and interactions online. Automated age estimation is central

When Accuracy Is Not Enough: Uncertainty Collapse between Noisy Label Learning and Out-of-Distribution Detection

arXiv:2605.17795v1 Announce Type: cross Abstract: Learning with noisy labels (LNL) is typically benchmarked by closed-set classification accuracy, yet deployment often requires classifiers to reject o

Agents

When Actions Disappear: Adversarial Action Removal in Self-Play Reinforcement Learning

arXiv:2605.16312v1 Announce Type: cross Abstract: We study adversarial action masking in self-play reinforcement learning: an attacker selectively removes legal actions from a victim's action set. Unl

agentsarxiv-cs-ai

When AI Tells You What You Want to Hear: Sycophantic Behavior of Large Language Models in Dementia Care Settings

model-releasesarxiv-cs-cl

arXiv:2605.16288v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used in clinical and care settings. This exploratory study investigates whether LLMs exhibit sycophantic

When Bits Break Recourse: Counterfactual-Faithful Quantization

arXiv:2605.17160v1 Announce Type: cross Abstract: Quantization can preserve predictive accuracy under low-bit deployment while silently breaking algorithmic recourse: an actionable change that flips a

When Dynamics Shift, Robust Task Inference Wins: Offline Imitation Learning with Behavior Foundation Models Revisited

arXiv:2605.17017v1 Announce Type: cross Abstract: Behavior Foundation Models (BFMs) enable scalable imitation learning (IL) by pretraining task-agnostic representations that can be rapidly adapted to

When Efficiency Backfires: Cascading LLMs Trigger Cascade Failure under Adversarial Attack

arXiv:2605.17288v1 Announce Type: cross Abstract: Large Language Model (LLM) cascade systems are designed to balance efficiency and performance by processing queries with lightweight models while sele

When Fireflies Cluster; Enhancing Automatic Clustering via Centroid-Guided Firefly Optimization

arXiv:2605.18460v1 Announce Type: new Abstract: This work presents a novel variant of the Firefly Algorithm (FA) for data clustering, addressing limitations of traditional methods like K-Means that st

When Is Rank-1 Steering Cheap? Geometry, Granularity, and Budgeted Search

arXiv:2605.16362v1 Announce Type: cross Abstract: Activation steering offers a lightweight way to control LLMs without retraining, but its effectiveness varies sharply across concepts. Prior work ofte

When Marginals Match but Structure Fails: Covariance Fidelity in Generative Models

arXiv:2603.17041v2 Announce Type: replace-cross Abstract: Generative models are increasingly deployed as substitutes for real data in downstream scientific workflows, yet standard evaluation criteria

Local Ai

When Molecular Similarity Works: Property Cliffs Reveal Hidden Errors

arXiv:2605.17265v1 Announce Type: new Abstract: Accurate prediction of molecular properties underpins drug discovery and material design, yet even state-of-the-art models remain vulnerable to localize

local-aiarxiv-cs-lg

When Outcome Looks Right But Discipline Fails: Trace-Based Evaluation Under Hidden Competitor State

arXiv:2605.18580v1 Announce Type: new Abstract: Outcome-only evaluation can certify economically unsafe agents: a policy can hit a business KPI while violating deployable behavioral discipline. In hot

When Personalization Legitimizes Risks: Uncovering Safety Vulnerabilities in Personalized Dialogue Agents

arXiv:2601.17887v2 Announce Type: replace Abstract: Long-term memory enables large language model (LLM) agents to support personalized and sustained interactions. However, most work on personalized ag

Tutorials

When TableQA Meets Noise: A Dual Denoising Framework for Complex Questions and Large-scale Tables

arXiv:2509.17680v2 Announce Type: replace Abstract: Table question answering (TableQA) is a fundamental task in natural language processing (NLP). The strong reasoning capabilities of large language m

tutorialsarxiv-cs-cl

When Vision Speaks for Sound

arXiv:2605.16403v1 Announce Type: new Abstract: Despite rapid progress in video-capable MLLMs, we find that their apparent audio understanding in videos is often vision-driven: models rely on visual c

safetyarxiv-cs-cv

Where Does Warm-Up Come From? Adaptive Scheduling for Norm-Constrained Optimizers

model-releasesarxiv-cs-lg

arXiv:2602.05813v2 Announce Type: replace Abstract: We study adaptive learning rate scheduling for norm-constrained optimizers (e.g., Muon and Lion). We introduce a generalized smoothness assumption u

Where Pretraining writes and Alignment reads: the asymmetry of Transformer weight space

arXiv:2605.16600v1 Announce Type: cross Abstract: Cross-entropy pretraining and preference alignment update the same transformer weights, but leave geometrically distinct traces. We characterise this

Whispers in the Noise: Surrogate-Guided Concept Awakening via a Multi-Agent Framework

arXiv:2605.18150v1 Announce Type: new Abstract: Diffusion models (DMs) are widely used for text-to-image generation, but their strong generative capabilities also raise concerns about unsafe or undesi

White-Box Sensitivity Auditing with Steering Vectors

arXiv:2601.16398v2 Announce Type: replace-cross Abstract: Algorithmic audits are essential tools for examining systems for properties required by regulators or desired by operators. Current audits of

safetyarxiv-cs-cl

WhiteTesseract: Reframing the Interpretation of Cultural Heritage through XR and Conversational AI

arXiv:2605.16972v1 Announce Type: cross Abstract: Cultural heritage exhibitions often struggle to sustain attention and support reflective engagement. Physical exhibitions rely on fixed interpretive a

Who Generated This 3D Asset? Learning Source Attribution for Generative 3D Models

arXiv:2605.18132v1 Announce Type: cross Abstract: Generative 3D models are deployed in gaming, robotics, and immersive creation, making source attribution critical: given a 3D asset, can we identify w

Why Do Reasoning Models Lose Coverage? The Role of Data and Forks in the Road

arXiv:2605.17026v1 Announce Type: new Abstract: Recent progress in large language models has led to the emergence of reasoning models, which have shown strong performance on complex tasks through spec

Why Do Safety Guardrails Degrade Across Languages?

arXiv:2605.17173v1 Announce Type: cross Abstract: Large language models exhibit safety degradation in non-English languages. Standard evaluation relies on Jailbreak Success Rate (JSR), which confounds

Why Modeling Human Haptic Material Perception with AI Is Difficult

arXiv:2605.16602v1 Announce Type: cross Abstract: Touch plays a central role in how humans perceive and recognize materials through physical contact. Despite decades of research, the mechanisms by whi

Agents

Why We Look Where We Look: Emergent Human-like Fixations of a Foveated Visual Language Model Maximizing Scene Understanding

arXiv:2605.17823v1 Announce Type: cross Abstract: When humans view scenes without a specific task (free-viewing), they initially direct their eye movements toward the scene center and then fixate on p

agentsarxiv-cs-ai

WinDeskGround: A Benchmark for Robust GUI Grounding in Complex Multi-Window Desktop Environments

arXiv:2605.16402v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) have revolutionized GUI automation, yet their efficacy is largely established on idealized, single-layer interf

WinQ: Accelerating Quantization-Aware Training of Language Models Around Saddle Points

arXiv:2605.17471v1 Announce Type: new Abstract: Quantization-aware training (QAT) is widely adopted to quantize language models by training full-precision weights using gradients from the quantized mo

WinTok: A Win-Win Hybrid Tokenizer via Decomposing Visual Understanding and Generation with Transferable Tokens

arXiv:2605.18115v1 Announce Type: new Abstract: Building a unified visual tokenizer is essential for bridging the gap between visual understanding and generation. Yet existing approaches struggle with

World Model-Enabled Causal Digital Twins for Semantic Communications in Physical AI Systems

arXiv:2605.16547v1 Announce Type: new Abstract: Semantic communication has emerged as a promising paradigm for enabling goal-oriented networking. However, most existing semantic communication solution

safetyarxiv-cs-lg

WorldArena 2.0: Extending Embodied World Model Benchmarking on Modality, Functionality and Platform

arXiv:2605.17912v1 Announce Type: cross Abstract: World models have emerged as a central paradigm for embodied intelligence, enabling agents to predict action-conditioned future and reason about envir

WOW-Seg: A Word-free Open World Segmentation Model

arXiv:2605.16903v1 Announce Type: new Abstract: Open world image segmentation aims to achieve precise segmentation and semantic understanding of targets within images by addressing the infinitely open

Applications

XCTFormer: Leveraging Cross-Channel and Cross-Time Dependencies for Enhanced Time-Series Analysis

arXiv:2605.18534v1 Announce Type: new Abstract: Multivariate time-series analysis involves extracting informative representations from sequences of multiple interdependent variables, supporting tasks

applicationsarxiv-cs-lg

Agents

Xiaomi EV World Model: A Joint World Model Integrating Reconstruction and Generation for Autonomous Driving

arXiv:2605.18137v1 Announce Type: new Abstract: This report presents a unified technical system addressing the two core capabilities of world models for autonomous driving: world representation and wo

agentsarxiv-cs-cv

Local Ai

YawDD+: Frame-level Annotations for Accurate Yawn Prediction

arXiv:2512.11446v3 Announce Type: replace Abstract: Driver fatigue remains a leading cause of road accidents, responsible for 24% of crashes. While yawning serves as an early behavioral indicator of f

local-aiarxiv-cs-cv

YOLO-NAS-Bench: A Surrogate Benchmark with Self-Evolving Predictors for YOLO Architecture Search

arXiv:2603.09405v2 Announce Type: replace Abstract: Neural Architecture Search (NAS) for object detection is severely bottlenecked by high evaluation cost, as fully training each candidate YOLO archit