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, 공유 시 출처만 명기하면 됩니다.

OpenS2V-Nexus: A Detailed Benchmark and Million-Scale Dataset for Subject-to-Video Generation

Does quantization affect models' performance on long-context tasks?

Homophily Enhanced Graph Domain Adaptation

EmoNet-Face: An Expert-Annotated Benchmark for Synthetic Emotion Recognition

Evaluating AI cyber capabilities with crowdsourced elicitation

PCDCNet: A Surrogate Model for Air Quality Forecasting with Physical-Chemical Dynamics and Constraints

Equivariant Representation Learning for Symmetry-Aware Inference with Guarantees

STRAP: Spatio-Temporal Pattern Retrieval for Out-of-Distribution Generalization

AmpleHate: Amplifying the Attention for Versatile Implicit Hate Detection

It's Not Just Labeling -- A Research on LLM Generated Feedback Interpretability and Image Labeling Sketch Features

Prompting Decision Transformers for Zero-Shot Reach-Avoid Policies

When Two LLMs Debate, Both Think They'll Win

Reality Check: A New Evaluation Ecosystem Is Necessary to Understand AI's Real World Effects

Restoring Real-World Images with an Internal Detail Enhancement Diffusion Model

Chain-of-Zoom: Extreme Super-Resolution via Scale Autoregression and Preference Alignment

Debate-to-Detect: Reformulating Misinformation Detection as a Real-World Debate with Large Language Models

A Survey of LLM $\times$ DATA

Task-Optimized Convolutional Recurrent Networks Align with Tactile Processing in the Rodent Brain

PhySense: Sensor Placement Optimization for Accurate Physics Sensing

Improving Generative Inverse Design of Rectangular Patch Antennas with Test Time Optimization

Towards Revealing the Effectiveness of Small-Scale Fine-tuning in R1-style Reinforcement Learning

Mixture of Low Rank Adaptation with Partial Parameter Sharing for Time Series Forecasting

TransBench: Breaking Barriers for Transferable Graphical User Interface Agents in Dynamic Digital Environments

CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training

OrionBench: A Benchmark for Chart and Human-Recognizable Object Detection in Infographics

FRIREN: Beyond Trajectories -- A Spectral Lens on Time

Select2Reason: Efficient Instruction-Tuning Data Selection for Long-CoT Reasoning

From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning

Gender and Positional Biases in LLM-Based Hiring Decisions: Evidence from Comparative CV/R\'esum\'e Evaluations

BitHydra: Towards Bit-flip Inference Cost Attack against Large Language Models

Auto-nnU-Net: Towards Automated Medical Image Segmentation

Benchmarking and Pushing the Multi-Bias Elimination Boundary of LLMs via Causal Effect Estimation-guided Debiasing

Bidirectional Variational Autoencoders

DiagnosisArena: Benchmarking Diagnostic Reasoning for Large Language Models

The Multimodal Information Based Speech Processing (MISP) 2025 Challenge: Audio-Visual Diarization and Recognition

Beyond Semantics: The Unreasonable Effectiveness of Reasonless Intermediate Tokens

RL in Name Only? Analyzing the Structural Assumptions in RL post-training for LLMs

VocalAgent: Large Language Models for Vocal Health Diagnostics with Safety-Aware Evaluation

Step-wise Adaptive Integration of Supervised Fine-tuning and Reinforcement Learning for Task-Specific LLMs

Shadow-FT: Tuning Instruct via Base

Towards Visuospatial Cognition via Hierarchical Fusion of Visual Experts

Visuospatial Cognitive Assistant

Enhance Mobile Agents Thinking Process Via Iterative Preference Learning

SoftPQ: Robust Instance Segmentation Evaluation via Soft Matching and Tunable Thresholds

Retrospex: Language Agent Meets Offline Reinforcement Learning Critic

PeerGuard: Defending Multi-Agent Systems Against Backdoor Attacks Through Mutual Reasoning

From Seeing to Doing: Bridging Reasoning and Decision for Robotic Manipulation

Enhancing Time Series Forecasting via a Parallel Hybridization of ARIMA and Polynomial Classifiers

Efficient Robotic Policy Learning via Latent Space Backward Planning

GarmentDiffusion: 3D Garment Sewing Pattern Generation with Multimodal Diffusion Transformers

Democratizing Differential Privacy: A Participatory AI Framework for Public Decision-Making

Sky-Drive: A Distributed Multi-Agent Simulation Platform for Human-AI Collaborative and Socially-Aware Future Transportation

Recursive Deep Inverse Reinforcement Learning

Information Gain-Guided Causal Intervention for Autonomous Debiasing Large Language Models

S1-Bench: A Simple Benchmark for Evaluating System 1 Thinking Capability of Large Reasoning Models

Achieving binary weight and activation for LLMs using Post-Training Quantization

H2VU-Benchmark: A Comprehensive Benchmark for Hierarchical Holistic Video Understanding

Machine Learning - Driven Materials Discovery: Unlocking Next-Generation Functional Materials -- A minireview

LLM-FE: Automated Feature Engineering for Tabular Data with LLMs as Evolutionary Optimizers

Panoramic Distortion-Aware Tokenization for Person Detection and Localization Using Transformers in Overhead Fisheye Images

ClearSight: Visual Signal Enhancement for Object Hallucination Mitigation in Multimodal Large language Models

No LLM is Free From Bias: A Comprehensive Study of Bias Evaluation in Large Language Models

How to Protect Yourself from 5G Radiation? Investigating LLM Responses to Implicit Misinformation

MA-LoT: Model-Collaboration Lean-based Long Chain-of-Thought Reasoning enhances Formal Theorem Proving

Deepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of Deepfakes Circulated in 2024

Learning Policy Committees for Effective Personalization in MDPs with Diverse Tasks

Backpropagation-free Spiking Neural Networks with the Forward-Forward Algorithm

Faithful Logic Embeddings in HOL -- Deep and Shallow

Between Circuits and Chomsky: Pre-pretraining on Formal Languages Imparts Linguistic Biases

Voting or Consensus? Decision-Making in Multi-Agent Debate

Learning to Align Multi-Faceted Evaluation: A Unified and Robust Framework

Training a Generally Curious Agent

Thinking Before Running! Efficient Code Generation with Thorough Exploration and Optimal Refinement

HybridLinker: Topology-Guided Posterior Sampling for Enhanced Diversity and Validity in 3D Molecular Linker Generation

Can Community Notes Replace Professional Fact-Checkers?

GeLLMO: Generalizing Large Language Models for Multi-property Molecule Optimization

Task-Informed Anti-Curriculum by Masking Improves Downstream Performance on Text

Evaluating link prediction: New perspectives and recommendations

MMUnlearner: Reformulating Multimodal Machine Unlearning in the Era of Multimodal Large Language Models

The Hidden Dimensions of LLM Alignment: A Multi-Dimensional Analysis of Orthogonal Safety Directions

TransMLA: Migrating GQA Models to MLA with Full DeepSeek Compatibility and Speedup

Optimizing Robustness and Accuracy in Mixture of Experts: A Dual-Model Approach

ExpProof : Operationalizing Explanations for Confidential Models with ZKPs

DHP: Discrete Hierarchical Planning for Hierarchical Reinforcement Learning Agents

Linear $Q$-Learning Does Not Diverge in $L^2$: Convergence Rates to a Bounded Set

KBQA-o1: Agentic Knowledge Base Question Answering with Monte Carlo Tree Search

Tuning LLM Judge Design Decisions for 1/1000 of the Cost

Accelerating Quantum Reinforcement Learning with a Quantum Natural Policy Gradient Based Approach

It's complicated. The relationship of algorithmic fairness and non-discrimination regulations for high-risk systems in the EU AI Act

More is not always better? Enhancing Many-Shot In-Context Learning with Differentiated and Reweighting Objectives

Unraveling Indirect In-Context Learning Using Influence Functions

ProgCo: Program Helps Self-Correction of Large Language Models

Learn from Global Correlations: Enhancing Evolutionary Algorithm via Spectral GNN

GMoE: Empowering LLMs Fine-Tuning via MoE Graph Collaboration

RoboMIND: Benchmark on Multi-embodiment Intelligence Normative Data for Robot Manipulation

Interlocking-free Selective Rationalization Through Genetic-based Learning

RL-SPH: Learning to Achieve Feasible Solutions for Integer Linear Programs

RE-Bench: Evaluating frontier AI R&D capabilities of language model agents against human experts

Double Descent Meets Out-of-Distribution Detection: Theoretical Insights and Empirical Analysis on the role of model complexity

Unleashing LLM Reasoning Capability via Scalable Question Synthesis from Scratch

Multi-agent Application System in Office Collaboration Scenarios

Created by

Haebom

저자

Songtao Sun, Jingyi Li, Yuanfei Dong, Haoguang Liu, Chenxin Xu, Fuyang Li, Qiang Liu

개요

본 논문은 사무실 협업 효율 및 업무 품질 향상을 위한 다중 에이전트 애플리케이션 시스템을 제시한다. 인공지능, 머신러닝, 자연어 처리 기술을 통합하여 작업 할당, 진행 상황 모니터링, 정보 공유 등의 기능을 구현한다. 시스템 내 에이전트는 팀 구성원의 요구사항에 기반한 개인화된 협업 지원을 제공하며, 데이터 분석 도구를 통합하여 의사결정 품질을 향상시킨다. 'Plan'과 'Solver'를 분리하는 지능형 에이전트 아키텍처를 제안하며, 다회차 질의 재작성 및 비즈니스 도구 검색 기술을 통해 에이전트의 다의도 및 다회차 대화 능력을 향상시킨다. 사무실 협업 시나리오에서의 도구 설계 및 다회차 대화 설계를 자세히 설명하고, 실험 및 평가를 통해 시스템의 효과를 검증한다. 실제 비즈니스 애플리케이션에서, 특히 질의 이해, 작업 계획 및 도구 호출에서 뛰어난 성능을 입증했으며, 동적인 환경과 대규모 다중 에이전트 시스템 내의 복잡한 상호 작용 문제 해결에 중요한 역할을 할 것으로 기대된다.

시사점, 한계점

•

시사점:

◦

다중 에이전트 시스템을 활용한 사무실 협업 효율 및 품질 향상 가능성 제시

◦

개인화된 협업 지원 및 데이터 분석 기반 의사결정 지원 기능 구현

◦

'Plan'과 'Solver' 분리 아키텍처를 통한 지능형 에이전트 설계 및 다회차 대화 능력 향상

◦

실제 비즈니스 애플리케이션에서의 효과 검증 및 뛰어난 성능 확인

◦

동적인 환경과 대규모 시스템에서의 복잡한 상호 작용 문제 해결 가능성 제시

•

한계점:

◦

논문에서 구체적인 시스템 아키텍처 및 알고리즘에 대한 상세한 설명 부족

◦

실험 및 평가의 구체적인 방법론 및 결과에 대한 자세한 정보 부족

◦

실제 적용 시 발생할 수 있는 문제점 및 한계에 대한 논의 부족

◦

시스템의 확장성 및 유지보수에 대한 고려 부족

Made with Slashpage