All entries

Transferable Evidence Reconstruction for Longitudinal Glucose Representations

arXiv:2609.28199v1 Announce Type: new Abstract: Long physiological recordings contain many routine measurements, while predictive information is often concentrated in rare events, sustained burden, an

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Turning Safety into Competence: Minimally Exploitable Robot Policies via Safety-Filtered Reinforcement Learning

DGX agent

arXiv:2609.27312v1 Announce Type: cross Abstract: Robots deployed for competitive tasks must outmaneuver their opponents without sacrificing safety. Existing approaches, including safe reinforcement l

safetyarxiv-cs-ai

24 Sept 2026

TwinCheck: Evidence-Grounded Negative-Twin Verification for Stateful Tool Agents

arXiv:2609.26911v1 Announce Type: new Abstract: A single locally plausible tool call can derail an otherwise successful agent trajectory. Suspicion alone does not justify intervention, because the rep

Two Global Crops Suffice: Locating Semantic Emergence in DINO-Style Self-Supervised Learning

arXiv:2609.28187v1 Announce Type: new Abstract: Self-supervised vision transformers trained with DINO-style objectives exhibit striking emergent semantic representation quality across visual tasks, ye

safetyarxiv-cs-cv

Type-II Error Bounds for Test Supermartingales from Lower-Tail Hypotheses

arXiv:2609.27766v1 Announce Type: cross Abstract: In safe hypothesis testing with test supermartingals, Ville's inequality provides anytime-valid type-I error guarantees for every significance level a

Uncheatable Eval: Dynamic Compression-Based Evaluation of Language Models

arXiv:2609.27510v1 Announce Type: cross Abstract: Modern large language models are pretrained on massive datasets, making it difficult to prevent benchmark data from entering their training sets and u

UniDataAgent: An Ontology-Grounded Agent for Enterprise Question-to-Report Automation

arXiv:2609.27257v1 Announce Type: new Abstract: Enterprise data agents must preserve organization specific semantics, not just translate questions into queries. We present ChinaUnicom DataAgent (UniDa

agentsarxiv-cs-cl

Applications

Untangling the Geometry and Speed for RF Sensing Spectrograms

arXiv:2609.26960v1 Announce Type: cross Abstract: A fundamental challenge in RF sensing is that Doppler signatures observed by a link entangle the target's motion with the sensing geometry, resulting

applicationsarxiv-cs-lg

Using Vision Language Foundation Models to Generate Plant Simulation Configurations via In-Context Learning

arXiv:2603.08930v2 Announce Type: replace-cross Abstract: This paper introduces a benchmark for evaluating whether vision-language models (VLMs) can generate plant simulation configurations from image

UVU: Improving Multimodal Understanding via Vision-Language Unified Autoregressive Paradigm

arXiv:2609.27915v1 Announce Type: new Abstract: Despite remarkable advancements in multimodal large language models (MLLMs), their fine-grained visual understanding is constrained by a primary relianc

researcharxiv-cs-cv

Validation and Simulation Catch Different Errors: Four Levels of Evaluation for LLM-Generated Circuits

arXiv:2609.26830v1 Announce Type: cross Abstract: Simulation success is not equivalent to structural correctness for LLM-generated circuits. We define and measure four evaluation levels -- schema vali

Value Mirror Descent for Reinforcement Learning

arXiv:2604.06039v2 Announce Type: replace-cross Abstract: Value iteration-type methods have been extensively studied for computing a nearly optimal value function in reinforcement learning (RL). Under

safetyarxiv-cs-lg

Variance Reduction for Independent Metropolis

arXiv:2406.17699v3 Announce Type: replace-cross Abstract: Assume that we would like to estimate the expected value of a function F with respect to an intractable density pi, which is specified up to s

Variational Bayesian Flow Network for Graph Generation

arXiv:2601.22524v2 Announce Type: replace Abstract: Graph generation aims to sample discrete node and edge attributes while satisfying coupled structural constraints. Diffusion models for graphs often

VCMM: Variance-Calibrated Momentum for Multimodal Learning

arXiv:2609.27577v1 Announce Type: new Abstract: Multimodal joint training often suffers from modality imbalance, where a dominant modality suppresses the optimization of others. Existing methods mainl

safetyarxiv-cs-lg

Verifiable Hidden Dynamics Play: Generating Agentic RL Environments from Solved Mechanisms

arXiv:2609.27321v1 Announce Type: new Abstract: Language-model agents increasingly face long-horizon tasks with evolving state, interdependent decisions, and delayed outcomes. Scaling their training r

agentsarxiv-cs-ai

Applications

VGM-VS: Rethinking Visual Geometry Model for High-Precision Visual Servoing

arXiv:2609.28312v1 Announce Type: cross Abstract: We present VGM-VS, a visual servoing method built on a pretrained feed-forward visual geometry model. Given the current view and a reference image cap

applicationsarxiv-cs-cv

Visibility-Guided Structured Measure Flow for Class-Conditioned 3D Gaussian Generation

model-releasesarxiv-cs-cv

arXiv:2609.27778v1 Announce Type: new Abstract: 3D Gaussian Splatting (3DGS) has made real-time, high-fidelity 3D rendering practical, yet turning this explicit representation into a native generative

Vision-Based Control of a Tether-Suspended Aerial Radiation Sensing Payload

arXiv:2609.27219v1 Announce Type: new Abstract: Aerial radiation surveys achieve higher sensitivity when the radiation detector is held close to the ground. Detector sensitivity falls off roughly with

researcharxiv-cs-ro

Visual Tripwires: Anticipating Failure in Deep Vision Systems

model-releasesarxiv-cs-cv

arXiv:2609.28099v1 Announce Type: new Abstract: Deep vision systems remain vulnerable to corruption, occlusion, and distribution shift despite strong benchmark performance. Existing reliability method

VIVAS: Vitalizing Visual Perception in VLM Pre-training via Vision-language Unified Autoregressive Supervision

arXiv:2609.27948v1 Announce Type: new Abstract: While Vision-Language Models (VLMs) demonstrate strong capabilities, they continue to suffer from a critical limitation: insufficient fine-grained visua

researcharxiv-cs-cv

VLM2GeoVec: Toward Universal Multimodal Embeddings for Remote Sensing

model-releasesarxiv-cs-cv

arXiv:2512.11490v2 Announce Type: replace Abstract: Satellite imagery differs from natural images in viewpoint, resolution, scale variation, and the prevalence of small objects -- demanding both regio

Local Ai

VLMs Can Describe, But Not Measure: Object-Centric Scene Understanding for Robotic Manipulation

arXiv:2609.28184v1 Announce Type: cross Abstract: Robotic operation in previously unseen environments requires both semantic understanding and reliable metric information. While vision--language model

local-aiarxiv-cs-cv

VMMU: A Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark

model-releasesarxiv-cs-cl

arXiv:2508.13680v5 Announce Type: replace Abstract: We introduce VMMU, a Vietnamese Multitask Multimodal Understanding and Reasoning Benchmark designed to evaluate how vision-language models (VLMs) in

WAInjectBench: Benchmarking Prompt Injection Detections for Web Agents

arXiv:2510.01354v2 Announce Type: replace-cross Abstract: Multiple prompt injection attacks have been proposed against web agents. At the same time, various methods have been developed to detect gener

Watch, Recall, Act: Always-On Robots in Concurrent Embodied Streams

arXiv:2609.28429v1 Announce Type: new Abstract: An always-on robot faces an endless stream that never resets: instructions arrive and lapse, the scene changes, and its own past actions reshape what it

safetyarxiv-cs-ro

Water Surface Swimming in a Centipede and its Robophysical ModeL

arXiv:2609.27088v1 Announce Type: new Abstract: Elongate multi-legged robots use coordinated body waves and distributed legs to move through cluttered terrestrial environments. However, as housing act

researcharxiv-cs-ro

Wave-Robust Passive AUV Localization Using FP-MUSIC

arXiv:2609.27712v1 Announce Type: new Abstract: Localizing an autonomous underwater vehicle without pre-deployed seabed transponders, or direct access to onboard vehicle sensors remains a core challen

agentsarxiv-cs-ro

'We'll Fix It Later': Education, AI, and the Deferral of Privacy in EdTech

arXiv:2609.28137v1 Announce Type: cross Abstract: Educational technology (EdTech) platforms collect highly sensitive student data, including behavioral logs, disability records, and academic histories

safetyarxiv-cs-ai

What Changed? Drift Detection with Real, Virtual, and Incomparable Diagnosis

arXiv:2609.27865v1 Announce Type: cross Abstract: Sharing a deep encoder does not, by itself, fix the central confound of task-comparison scores. We show that cross-evaluated heads on a frozen shared

researcharxiv-cs-ai

What Changes When Fact-Verification Scores Improve? Evidence and Answer Accounting Across Trained Verifiers and LLMs

model-releasesarxiv-cs-cl

arXiv:2609.27064v1 Announce Type: new Abstract: A joint fact-verification score assesses answers and submitted evidence together. When the score improves, how much of the gain remains if the answers a

What Confidence Routing Is Actually Doing: Auditing Routing, Calibration, and Commitment in Multi-Agent Deliberation

model-releasesarxiv-cs-cl

arXiv:2609.27822v1 Announce Type: cross Abstract: A common multi-agent design asks agents to report confidence and lets the highest-scoring agent speak next, implicitly using one scalar both to route

What Converges in the Platonic Representation Hypothesis? Structure over Geometry

arXiv:2609.27252v1 Announce Type: cross Abstract: The Platonic Representation Hypothesis suggests that increasingly capable models converge toward shared representations. Recent work narrows this clai

safetyarxiv-cs-cv

What Do Tabular Foundation Models Compute In Context? In-Situ Representation Refinement through Attention-Gated Updates

arXiv:2609.27679v1 Announce Type: new Abstract: What reusable computation should a tabular foundation model learn when every table defines a new supervised task? We develop in-situ representation refi

What Looks Like a Capability Limit in Vision-Language Models Is a Readout Limit

arXiv:2609.27408v1 Announce Type: cross Abstract: Benchmarks for vision-language models offer their answer choices in some convention: a letter, a color name, a pixel coordinate. That convention is tr

What Makes a Terminal-Bench Task Hard? Separating Genuine Hardness from Fake-Hardness on an Adjudicated Agentic Corpus

arXiv:2609.26826v1 Announce Type: cross Abstract: Frontier benchmarks need tasks that current models cannot solve. But a task that no model solves is not automatically a hard task. The same zero pass

agentsarxiv-cs-ai

'What's That Sound?': A Versatile, Robust, and Lightweight Convolutional Transformer for Environment Sound Recognition

arXiv:2609.27762v1 Announce Type: cross Abstract: The conventional hearing aid is both costly and lim- ited in usage, as it is not intended to detect non-speech audio. Our objective is to develop a ma

safetyarxiv-cs-lg

WhatWorkedBench: Benchmarking Experimental Understanding in AI Agents

arXiv:2609.27490v1 Announce Type: new Abstract: AI research agents need reliable knowledge of how their experiments change outcomes. We introduce WhatWorkedBench to measure experimental understanding,

agentsarxiv-cs-ai

When Accuracy Gaps Fail to Certify: Auditing Cross-Domain Recalibration of LLM Judges

arXiv:2609.27954v1 Announce Type: new Abstract: A scalar recalibration map fitted for an LLM judge on one task can fail when the task distribution changes, but the source-target accuracy gap is often

When Adaptation Hurts: Split Sensitivity and Person-Level Negative Transfer in Federated Wearable Onboarding

arXiv:2609.27819v1 Announce Type: new Abstract: Federated wearable models eventually serve people absent from source training, but favorable average accuracy does not establish that unlabeled onboardi

When and Where to Trust the Teacher: Unifying On-Policy Distillation and GRPO through Entropy-Calibrated Credit Assignment

arXiv:2609.28385v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) supervises mathematical reasoning through final-answer correctness, but provides little guidance

safetyarxiv-cs-ai

When Clients Are Orchestrated: Strategic Gradient Manipulation to Defeat Federated Learning Servers with Efficient Defense

arXiv:2609.27124v1 Announce Type: cross Abstract: Federated Learning enables decentralized model training by exchanging model updates--rather than raw data--with a central parameter server (PS). While

When Context Misleads: In-context Learning with Jurisdiction in Large Language Models

arXiv:2609.27603v1 Announce Type: cross Abstract: In-Context Learning (ICL) has become a cornerstone of modern LLM deployment. However, existing ICL post-training methods have a critical blind spot: t

When Entanglement Lower-Bounds Disparity: Auditing and Repairing Demographic Fairness in Audio Understanding Models

arXiv:2609.27382v1 Announce Type: cross Abstract: Speech technology penalizes some voices: recognition errs nearly twice as often for Black speakers, and accuracy declines for second-language accents

safetyarxiv-cs-cl

When Labels Are Scarce: An Oscillatory State Space Model for Vibration Diagnosis

arXiv:2609.27411v1 Announce Type: new Abstract: Machine fault diagnosis from vibration requires learning from scarce labelled fault recordings while meeting the computational constraints of edge devic

When Learned Context Planning Fails to Beat Strong Retrieval: A Controlled Study of Planning, Routing, and Reranking for Long-Context QA

arXiv:2609.26976v1 Announce Type: new Abstract: Learned context planning selects evidence atoms before an answer model reasons over them. We test whether this learned selection improves long-context m

researcharxiv-cs-cl

When Parallel Drafter Meets Parallel Speculative Decoding

arXiv:2609.27396v1 Announce Type: new Abstract: DSpark-style parallel drafters have made speculative decoding highly effective, yet their draft phase remains serialized on the critical path of every r

researcharxiv-cs-cl

When Post-Processing Fairness Constraints Help and When They Harm: Evidence from Eight Cross-Domain Evaluations

arXiv:2609.26955v1 Announce Type: new Abstract: Fairness audits in production ML typically occur once, at deployment, on a single domain. Both fail in practice: fairness can shift after retraining or