Search: “evaluation”

On the Nonlinearity of Learning Rate Scaling for LLM Training

arXiv:2606.29158v1 Announce Type: cross Abstract: Learning-rate transfer can reduce the cost of training large language models: instead of sweeping learning rates at target scale, practitioners extrap

researcharxiv-cs-ai

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

On the Vulnerability of Parameter-Level Defenses to Model Merging

DGX agent

arXiv:2606.30360v1 Announce Type: cross Abstract: The training-free integration of expert models via model merging has exposed significant security risks, enabling free-riders to combine specialized m

model-releasesarxiv-cs-cv

30 Jun 2026

One Scene, Two Depths: Probing Geometric Ambiguity in Monocular Foundation Models

arXiv:2606.29600v1 Announce Type: cross Abstract: A faithful 3D world representation should account for layered geometry, where a single camera ray may contain multiple visible and geometrically valid

One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining

arXiv:2606.30634v1 Announce Type: new Abstract: Modern large-scale LLM pretraining benefits from utilizing Pipeline Parallelism; however, synchronous implementations leave GPUs idle during pipeline bu

researcharxiv-cs-lg

Local Ai

Online Data Selection for Instruction Tuning via Gaussian Processes

arXiv:2606.30077v1 Announce Type: cross Abstract: With Large Language Model (LLM) pre-training and fine-tuning shifting its focus from data volume to data quality, quality data selection has emerged a

local-aiarxiv-cs-ai

Online Experiential Learning for Language Models

arXiv:2603.16856v2 Announce Type: replace Abstract: The prevailing paradigm for improving large language models relies on offline training with human annotations or simulated environments, leaving the

safetyarxiv-cs-cl

OP3DSG: Open-Vocabulary Part-Aware 3D Scene Graph Generation for Real-World Environments

model-releasesarxiv-cs-cv

arXiv:2606.29786v1 Announce Type: new Abstract: 3D scene graphs (3DSGs) provide a compact and structured abstraction of 3D environments. Although advances in foundation models have enabled open-vocabu

Tutorials

Open-Vocabulary and Referring Segmentation for 3D Gaussians Using 2D Detectors

arXiv:2606.30638v1 Announce Type: new Abstract: 3D Gaussian Splatting (3DGS) has emerged at the forefront of 3D scene reconstruction. Extending 3DGS with language-driven, open-vocabulary understanding

tutorialsarxiv-cs-cv

OpenSPM: An Environment-Transferable Robotic Key Spatial Pose Memory and Closed-Loop High-Frequency Flow-Matching Action Generation Model

arXiv:2606.29936v1 Announce Type: new Abstract: Open-environment tabletop robotic manipulation requires systems to possess semantic understanding, precise geometric pose estimation, and high-frequency

researcharxiv-cs-ro

Orca: The World is in Your Mind

arXiv:2606.30534v1 Announce Type: new Abstract: We introduce Orca, an initial instantiation of a general world foundation model. Orca learns a unified world latent space from multimodal world signals

PASTA: A Paraphrasing And Self-Training Approach for Knowledge Updating in LLMs

arXiv:2606.28898v1 Announce Type: new Abstract: Knowledge updating in pre-trained Large Language Models (LLMs) remains an important challenge. While continual training provides a potential avenue for

researcharxiv-cs-cl

PCGD: Physics-Guided Conditional Graph Diffusion for TCAD Device Simulation

model-releasesarxiv-cs-lg

arXiv:2606.29272v1 Announce Type: new Abstract: Technology computer-aided design (TCAD) semiconductor device simulation is fundamentally constrained by the high computational cost of iteratively solvi

PinNet: Keypoint-Aware Learned Local Descriptors with Geometric Embedding for Loop Closure in LiDAR SLAM

arXiv:2606.28637v1 Announce Type: new Abstract: Loop closure is essential to reduce drift and build globally consistent maps in large-scale environments. However, reliable loop closure with only geome

researcharxiv-cs-ro

PLAA: Packet-level Adversarial Attacks in Network Traffic Detection

arXiv:2606.28439v1 Announce Type: cross Abstract: Deep neural networks (DNNs) are widely applied in Network-based Intrusion Detection System (NIDS) due to their high accuracy. However, DNNs are highly

researcharxiv-cs-ai

Position: RL Researchers Need to Distinguish Between Solving Simulators and Using Simulators as a Proxy

model-releasesarxiv-cs-lg

arXiv:2606.28433v1 Announce Type: new Abstract: One goal in reinforcement learning (RL) research is to understand general-purpose sequential decision-making, using benchmark simulators as a proxy for

Predictive Objectives Discard Exogenous Control-Relevant Features: A Controlled Mechanistic Study

arXiv:2606.30068v1 Announce Type: new Abstract: Joint-embedding predictive (JEPA-style) objectives learn representations by predicting future latents. In doing so they can discard features that are ex

safetyarxiv-cs-lg

Preserving Fairness and Safety in Quantized LLMs Through Critical Weight Protection

arXiv:2601.12033v2 Announce Type: replace Abstract: Quantization is widely adopted to reduce the computational cost of large language models (LLMs); however, its implications for fairness and safety,

safetyarxiv-cs-cl

Preventing Error Propagation in Multi-Agent AI through Runtime Monitoring

arXiv:2606.29026v1 Announce Type: new Abstract: Multi-agent AI systems can improve answer selection by allowing different language models to exchange reasoning traces, revise initial predictions, and

Probing and Leveraging Video Diffusion Transformer Features for Robust Point Tracking

arXiv:2512.20606v2 Announce Type: replace Abstract: Despite achieving strong results on standard benchmarks, current point tracking methods rely on feature backbones that are rarely designed with the

applicationsarxiv-cs-cv

Projection-based coupling of infrared thermography and stereocorrelation-based digital image correlation

arXiv:2606.28905v1 Announce Type: new Abstract: Full-field measurement techniques such as digital image correlation and infrared thermography are prevalent in experimental solid mechanics. Digital ima

PromptGNN-sim: Deep Fusion and Alignment of GNN and LLMs for Text-Attributed Graph Learning

arXiv:2606.30291v1 Announce Type: new Abstract: Text-Attributed Graphs (TAGs) combine textual semantics with graph structure and are central to many graph learning tasks. However, existing fusion meth

Proof-of-Guardrail in AI Agents and What (Not) to Trust from It

arXiv:2603.05786v2 Announce Type: replace-cross Abstract: As AI agents become widely deployed as online services, users often rely on an agent developer's claim about how safety is enforced, which int

ProSpec RL: Plan Ahead, then Execute

arXiv:2407.21359v2 Announce Type: replace-cross Abstract: Imagining potential outcomes of actions before execution helps agents make more informed decisions, a prospective thinking ability fundamental

Pushing Forward Pareto Frontiers of Proactive Agents with Behavioral Agentic Optimization

arXiv:2602.11351v2 Announce Type: replace Abstract: Proactive large language model (LLM) agents aim to actively plan, query, and interact over multiple turns, enabling efficient task completion beyond

Hardware

RAGA: Real Time Ray Traced Gaussian Shadow Casting for 3DGS Avatar-Scene Interaction

arXiv:2606.29329v1 Announce Type: new Abstract: We study the problem of physically plausible shadow casting when animating 3D Gaussian Splatting (3DGS) avatars, either individually or in multi-avatar

hardwarearxiv-cs-cv

Real-time Rendering-based Surgical Instrument Tracking via Evolutionary Optimization

arXiv:2603.11404v3 Announce Type: replace-cross Abstract: Accurate and efficient tracking of surgical instruments is fundamental for Robot-Assisted Minimally Invasive Surgery. Although vision-based ro

applicationsarxiv-cs-cv

Realtime Wind Estimation using Low Cost Quadrotor Uncrewed Aerial Vehicles

arXiv:2606.30581v1 Announce Type: new Abstract: In environmental monitoring as well as emergency response applications such as wildfires, wind velocity measurement is essential. Quadrotor UAVs have be

researcharxiv-cs-ro

Reconsidering Overthinking: Penalizing Internal and External Redundancy in CoT Reasoning

arXiv:2508.02178v3 Announce Type: replace Abstract: Large reasoning models (LRMs) often exhibit overthinking, producing verbose Chain-of-Thought (CoT) traces that increase inference cost and obscure t

researcharxiv-cs-ai

Recursive Self-Evolving Agents via Held-Out Selection

arXiv:2606.28374v1 Announce Type: new Abstract: LLM agents are increasingly improved without weight updates by evolving a natural-language artifact, such as reflections, workflows, playbooks, cheatshe

Reinforcement Learning for Software Vulnerability Analysis: A Systematic Review with Emphasis on C/C++ Source Code and Static Analysis

arXiv:2606.28403v1 Announce Type: cross Abstract: Vulnerability detection in C/C++ software remains a major security challenge due to code complexity, manual memory management, and the limitations of

Reliability-Prioritized Fine-Grained Generation in Multimodal Large

model-releasesarxiv-cs-cv

arXiv:2606.29573v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) are increasingly expected to generate fine-grained descriptions of visual content. However, we observe and theo

Representation Learning for Equivariant Inference with Guarantees

arXiv:2505.19809v3 Announce Type: replace-cross Abstract: In many real-world applications of regression, conditional probability estimation, and uncertainty quantification, exploiting symmetries roote

applicationsarxiv-cs-ai

Research Entity Extraction and Topic Detection from UKRI Grant Proposals

arXiv:2606.30304v1 Announce Type: cross Abstract: This paper presents preliminary findings from a UKRI-funded Metascience project comparing three LLM-based approaches, GPT-4o, Mistral, and a bespoke a

RetrDex: Efficient Object Retrieval in Cluttered Scenes with a Dexterous Hand

arXiv:2502.18423v3 Announce Type: replace Abstract: Retrieving objects buried beneath clutter is both challenging and time-consuming, as complex support relationships make manipulation particularly di

safetyarxiv-cs-ro

Tutorials

RGLD: Randomized Global-Local Density Estimation for Tabular Anomaly Detection

arXiv:2606.28970v1 Announce Type: cross Abstract: Unsupervised tabular anomaly detection requires methods that are accurate, robust across heterogeneous datasets, and computationally efficient. Classi

tutorialsarxiv-cs-ai

RIPA: Sensory-Vector Prompt Injection Attacks on LLM-Controlled ROS 2 Robots

arXiv:2606.28649v1 Announce Type: cross Abstract: We present RIPA, the first systematic multi-channel empirical study of prompt injection attacks delivered through the sensory pipeline of a ROS 2-base

Robust Extended Kalman Filter for Land Navigation Using Massive Array of MEMS IMUs

arXiv:2606.29271v1 Announce Type: cross Abstract: We propose a robust Extended Kalman Filter (EKF) architecture for land navigation using an array of hundreds of low-cost micro-electromechanical syste

safetyarxiv-cs-ro

Room Scene Discovery and Grouping in Unstructured Vacation Rental Image Collections

arXiv:2507.00263v2 Announce Type: replace Abstract: The rapid growth of vacation rental (VR) platforms has led to an increasing volume of property images, often uploaded without structured categorizat

SAD-GS: Learning Reliable 3D Semantic Gaussian Fields via Dynamic Geo-Semantic Anchoring

arXiv:2606.29376v1 Announce Type: new Abstract: Open-vocabulary 3D semantic Gaussian field learning relies on multi-view 2D supervision, whose semantic targets and spatial assignments are often unreli

SAFE-DiT: Semantics-Aware Fast-path Execution for High-Resolution Diffusion Transformers

arXiv:2606.29360v1 Announce Type: new Abstract: High-resolution Diffusion Transformer (DiT) inference contains substantial spatial redundancy, but many spatially adaptive implementations encode region

safetyarxiv-cs-cv

SafeGEO: Understanding Generative Engine Optimization Risks in Recommendation Agents

arXiv:2606.28356v1 Announce Type: cross Abstract: Generative Engine Optimization (GEO) lets content owners rewrite web content to increase their visibility in generative systems. In recommendation age

SAGA: Scene-Aware, Goal-Evolving Agents for Long-Horizon CivRealm Strategy Planning

arXiv:2606.29932v1 Announce Type: new Abstract: Long-horizon strategic planning in complex strategy games demands concurrent reasoning across multiple decision domains under imperfect information and

SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm

arXiv:2606.20523v2 Announce Type: replace-cross Abstract: Multimodal foundation models have advanced rapidly thanks to large optical benchmarks, but comparable resources for synthetic aperture radar (

SATB-VR: Training Few-Step Video Restoration Diffusion Model using SNR-Aware Trajectory Blending

arXiv:2606.28677v1 Announce Type: new Abstract: While diffusion models excel in video restoration, their reliance on extensive iterative steps limits efficiency. Conversely, aggressive single-step dis

applicationsarxiv-cs-cv

SatSplat: Geometrically-Accurate Gaussian Splatting for Satellite Imagery

model-releasesarxiv-cs-cv

arXiv:2606.28581v1 Announce Type: new Abstract: High-resolution satellite imagery demands 3D reconstruction methods that deliver both speed and geometric accuracy. Recent adaptations of 3D Gaussian Sp

ScAle: Attention Head Scaling as a Minimal Adapter for Spatial Reasoning in Vision Language Models

arXiv:2606.29579v1 Announce Type: cross Abstract: Spatial reasoning remains a persistent challenge for many vision language models (VLMs), and improving it typically requires fine-tuning with substant

Selective Memory Retention for Long-Horizon LLM Agents

arXiv:2606.29178v1 Announce Type: new Abstract: When does retention matter for memory-augmented LLM agents? We study this with TraceRetain, a lightweight framework for bounded external memory in froze

Self-Evolving World Models for LLM Agent Planning

arXiv:2606.30639v1 Announce Type: new Abstract: World models offer a principled way to equip long-horizon LLM agents with foresight: predictions of action consequences before execution. However, unrel