Daily Arxiv

전 세계에서 발간되는 인공지능 관련 논문을 정리하는 페이지 입니다.
본 페이지는 Google Gemini를 활용해 요약 정리하며, 비영리로 운영 됩니다.
논문에 대한 저작권은 저자 및 해당 기관에 있으며, 요약본 공유 시 출처만 명기하면 됩니다.
This service is supported by Google Gemini.

PromptRad: Knowledge-Enhanced Multi-Label Prompt-Tuning for Low-Resource Radiology Report Labeling

FineBench: Benchmarking and Enhancing Vision-Language Models for Fine-grained Human Activity Understanding

Component-Aware Structure-Preserving Style Transfer for Satellite Visual Sim2Real Data Construction

ARC-RL: A Reinforcement Learning Playground Inspired by ARC Raiders

Unlocking the Potential of Continual Model Merging: An ODE Perspective

Rebalancing Reference Frame Dominance to Improve Motion in Image-to-Video Models

Toward User Comprehension Supports for LLM Agent Skill Specifications

Agent Security is a Systems Problem

ZeroUnlearn: Few-Shot Knowledge Unlearning in Large Language Models

Exact Linear Attention

Lance: Unified Multimodal Modeling by Multi-Task Synergy

Charon: A Unified and Fine-Grained Simulator for Large-Scale LLM Training and Inference

UCSF-PDGM-VQA: Visual Question Answering dataset for brain tumor MRI interpretation

Voice ''Cloning'' is Style Transfer

Toward Template-Free Explainability for Monte Carlo Tree Search

The Impact of AI Search on the Online Content Ecosystem: Evidence from Google and Reddit

Bridging Silicon and the Hippocampus: Algebro-Deterministic Memory "VaCoAl" as a Substrate for Vector-HaSH and TEM

PBT-Bench: Benchmarking AI Agents on Property-Based Testing

MeMo: Memory as a Model

WriteSAE: Sparse Autoencoders for Recurrent State

The critical slowing down in diffusion models

Beyond GRPO and On-Policy Distillation: An Empirical Sparse-to-Dense Reward Principle for Language-Model Post-Training

Fill the GAP: A Granular Alignment Paradigm for Visual Reasoning in Multimodal Large Language Models

A Theory of Time-Sensitive Language Generation: Sparse Hallucination Beats Mode Collapse

A Comparative Study of Machine Learning and Deep Learning for Out-of-Distribution Detection

Task-Agnostic Noisy Label Detection via Standardized Loss Aggregation

APCD: Adaptive Path-Contrastive Decoding for Reliable Large Language Model Generation

Benchmarking EngGPT2-16B-A3B against Comparable Italian and International Open-source LLMs

Consistent Geometric Deep Learning via Hilbert Bundles and Cellular Sheaves

Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex

SOPE: Stabilizing Off-Policy Evaluation for Online RL with Prior Data

JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation

Graph Neural Network based Hierarchy-Aware Embeddings of Knowledge Graphs: Applications to Yeast Phenotype Prediction

HeadQ: Model-Visible Distortion and Score-Space Correction for KV-Cache Quantization

How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance

Compute Aligned Training: Optimizing for Test Time Inference

Speech Enhancement Based on Drifting Models

Differentially Private Model Merging

When Fairness Metrics Disagree: Evaluating the Reliability of Demographic Fairness Assessment in Machine Learning

TIP: Token Importance in On-Policy Distillation

Towards Autonomous Mechanistic Reasoning in Virtual Cells

Beyond Attention Scores: SVD-Based Vision Token Pruning for Efficient Vision-Language Models

Lightweight Low-Light Image Enhancement via Distribution-Normalizing Preprocessing and Depthwise U-Net

IMPACT: Influence Modeling for Open-Set Time Series Anomaly Detection

Why Aggregate Accuracy is Inadequate for Evaluating Fairness in Law Enforcement Facial Recognition Systems

Transcription and Recognition of Italian Parliamentary Speeches Using Vision-Language Models

AI-Powered Facial Mask Removal Is Not Suitable For Identification

Sustainability Is Not Linear: Quantifying Performance, Energy, and Privacy Trade-offs in On-Device Intelligence

How Open Must Language Models be to Enable Reliable Scientific Inference?

On Integrating Resilience and Human Oversight into LLM-Assisted Modeling Workflows for Digital Twins

FEAT: A Linear-Complexity Foundation Model for Extremely Large Structured Data

Informationally Compressive Anonymization: Non-Degrading Sensitive Input Protection for Privacy-Preserving Supervised Machine Learning

Deeper Thought, Weaker Aim: Understanding and Mitigating Perceptual Impairment during Reasoning in Multimodal Large Language Models

The Generation-Recognition Asymmetry: Six Dimensions of a Fundamental Divide in Formal Language Theory

Exploring Deep Learning and Ultra-Widefield Imaging for Diabetic Retinopathy and Macular Edema

MASFactory: A Graph-centric Framework for Orchestrating LLM-Based Multi-Agent Systems with Vibe Graphing

SPARC: Spatial-Aware Path Planning via Attentive Robot Communication

One Operator to Rule Them All? On Boundary-Indexed Operator Families in Neural PDE Solvers

Iterative LLM-based improvement for French Clinical Interview Transcription and Speaker Diarization

Multimodal Optimal Transport for Training-free Temporal Segmentation in Surgical Robotics

Anatomy of Agentic Memory: Taxonomy and Empirical Analysis of Evaluation and System Limitations

VLANeXt: Recipes for Building Strong VLA Models

Flow Map Language Models: One-step Language Modeling via Continuous Denoising

Explainable AI: Context-Aware Layer-Wise Integrated Gradients for Explaining Transformer Models

CompilerKV: Risk-Adaptive KV Compression via Offline Experience Compilation

CTFExplorer: Evaluating LLM Offensive Agents Through Multi-Target Web CTF Benchmarking

rePIRL: Learn PRM with Inverse RL for LLM Reasoning

SHINE: A Scalable In-Context Hypernetwork for Mapping Context to LoRA in a Single Pass

Learning Query-Aware Budget-Tier Routing for Runtime Agent Memory

Causal Discovery from Heteroscedastic Stochastic Dynamical Systems under Imperfect Physical Models

Graph Autoencoder for Process Monitoring

MCP-Atlas: A Large-Scale Benchmark for Tool-Use Competency with Real MCP Servers

When Does Adaptation Win? Scaling Laws for Meta-Learning in Quantum Control

Mind the Generative Details: Direct Localized Detail Preference Optimization for Video Diffusion Models

Deep Neural Networks as Discrete Dynamical Systems: Implications for Physics-Informed Learning

DrugRAG: Enhancing Pharmacy LLM Performance Through A Novel Retrieval-Augmented Generation Pipeline

End2Reg: Learning Task-Specific Segmentation for Markerless Registration in Spine Surgery

CHEM: Estimating and Understanding Hallucinations in Deep Learning for Image Processing

Principled RL for Flow Matching Emerges from the Chunk-level Policy Optimization

Can VLMs Unlock Semantic Anomaly Detection? A Framework for Structured Reasoning

FineVision: Open Data Is All You Need

ATLAS: Adaptive Trading with LLM AgentS Through Dynamic Prompt Optimization and Multi-Agent Coordination

A Free Lunch in LLM Compression: Revisiting Retraining after Pruning

InteractScience: Programmatic and Visually-Grounded Evaluation of Interactive Scientific Demonstration Code Generation

EvalMORAAL: Interpretable Chain-of-Thought and LLM-as-Judge Evaluation for Moral Alignment in Large Language Models

Q-Net: Queue Length Estimation via Kalman-based Neural Networks

Measuring and mitigating overreliance to build human-compatible AI

Access Paths for Efficient Ordering with Large Language Models

STM3: Mixture of Multiscale Mamba for Long-Term Spatio-Temporal Time-Series Prediction

Retrospective Sparse Attention for Efficient Long-Context Generation

FAIR-Pruner: A Flexible Framework for Automatic Layer-Wise Pruning via Tolerance of Difference

M3: Conversational LLMs Simplify Secure Clinical Data Access, Understanding, and Analysis

Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning

Time-Prompt: Integrated Heterogeneous Prompts for Unlocking LLMs in Time Series Forecasting

Code Researcher: Deep Research Agent for Large Systems Code and Commit History

Task-conditioned probing of instruction-tuned multimodal LLMs: Region-specific brain alignment patterns under naturalistic stimuli

Beyond Words: Multimodal LLM Knows When to Speak

Bridging Language Models and Financial Analysis

Do LLMs Triage Like Clinicians? A Dynamic Study of Outpatient Referral

END: Early Noise Dropping for Efficient and Effective Context Denoising

Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models

Created by

Haebom

저자

Qiguang Chen, Libo Qin, Jinhao Liu, Dengyun Peng, Jiannan Guan, Peng Wang, Mengkang Hu, Yuhang Zhou, Te Gao, Wanxiang Che

개요

본 논문은 최근 대규모 언어 모델(LLM)의 추론 능력 향상에 중요한 역할을 하는 장쇄 사고 과정(Long CoT)에 대한 종합적인 조사를 제공합니다. 기존의 단쇄 사고 과정(Short CoT)과의 차이점을 명확히 하고, 장쇄 사고 과정의 핵심 특징(심층 추론, 광범위한 탐색, 실행 가능한 반성)을 분석합니다. 또한, 과도한 사고(overthinking) 및 추론 시간 확장과 같은 현상을 조사하고, 다모달 추론 통합, 효율성 개선, 향상된 지식 프레임워크 등 미래 연구 방향을 제시합니다.