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.

Evaluating how LLM annotations represent diverse views on contentious topics

QualiSpeech: A Speech Quality Assessment Dataset with Natural Language Reasoning and Descriptions

ShED-HD: A Shannon Entropy Distribution Framework for Lightweight Hallucination Detection on Edge Devices

OS-Kairos: Adaptive Interaction for MLLM-Powered GUI Agents

MathFusion: Enhancing Mathematical Problem-solving of LLM through Instruction Fusion

Transformers without Normalization

Compute Optimal Scaling of Skills: Knowledge vs Reasoning

Identifying Trustworthiness Challenges in Deep Learning Models for Continental-Scale Water Quality Prediction

InfiniSST: Simultaneous Translation of Unbounded Speech with Large Language Model

From Euler to AI: Unifying Formulas for Mathematical Constants

C2-DPO: Constrained Controlled Direct Preference Optimization

Less is More: Improving LLM Alignment via Preference Data Selection

Quantifying Memorization and Parametric Response Rates in Retrieval-Augmented Vision-Language Models

ProMedTS: A Self-Supervised, Prompt-Guided Multimodal Approach for Integrating Medical Text and Time Series

Boosting Generalization in Diffusion-Based Neural Combinatorial Solver via Inference Time Adaptation

HARBOR: Exploring Persona Dynamics in Multi-Agent Competition

SelfCite: Self-Supervised Alignment for Context Attribution in Large Language Models

Truth Knows No Language: Evaluating Truthfulness Beyond English

PoGDiff: Product-of-Gaussians Diffusion Models for Imbalanced Text-to-Image Generation

Multi-Knowledge-oriented Nighttime Haze Imaging Enhancer for Vision-driven Intelligent Systems

Life-Code: Central Dogma Modeling with Multi-Omics Sequence Unification

Is attention all you need to solve the correlated electron problem?

Optimizing Temperature for Language Models with Multi-Sample Inference

Speak Easy: Eliciting Harmful Jailbreaks from LLMs with Simple Interactions

Adapt-Pruner: Adaptive Structural Pruning for Efficient Small Language Model Training

The Other Side of the Coin: Unveiling the Downsides of Model Aggregation in Federated Learning from a Layer-peeled Perspective

Satori: Reinforcement Learning with Chain-of-Action-Thought Enhances LLM Reasoning via Autoregressive Search

Activation-Informed Merging of Large Language Models

Layer by Layer: Uncovering Hidden Representations in Language Models

Search-Based Adversarial Estimates for Improving Sample Efficiency in Off-Policy Reinforcement Learning

Activation by Interval-wise Dropout: A Simple Way to Prevent Neural Networks from Plasticity Loss

ASAP: Learning Generalizable Online Bin Packing via Adaptive Selection After Proposal

Which Optimizer Works Best for Physics-Informed Neural Networks and Kolmogorov-Arnold Networks?

Rethinking Table Instruction Tuning

Concurrent Learning with Aggregated States via Randomized Least Squares Value Iteration

Video Depth Anything: Consistent Depth Estimation for Super-Long Videos

Foundations of Large Language Models

Benchmarking Rotary Position Embeddings for Automatic Speech Recognition

An Investigation into Value Misalignment in LLM-Generated Texts for Cultural Heritage

Retrieval-Augmented Dynamic Prompt Tuning for Incomplete Multimodal Learning

Towards End-to-End Neuromorphic Voxel-based 3D Object Reconstruction Without Physical Priors

AnalogXpert: Automating Analog Topology Synthesis by Incorporating Circuit Design Expertise into Large Language Models

Achieving Collective Welfare in Multi-Agent Reinforcement Learning via Suggestion Sharing

ClusterKV: Manipulating LLM KV Cache in Semantic Space for Recallable Compression

AccDiffusion v2: Towards More Accurate Higher-Resolution Diffusion Extrapolation

Immune: Improving Safety Against Jailbreaks in Multi-modal LLMs via Inference-Time Alignment

FastRAG: Retrieval Augmented Generation for Semi-structured Data

A dataset of questions on decision-theoretic reasoning in Newcomb-like problems

Knowledge-Augmented Multimodal Clinical Rationale Generation for Disease Diagnosis with Small Language Models

Mind Your Step (by Step): Chain-of-Thought can Reduce Performance on Tasks where Thinking Makes Humans Worse

Building, Reusing, and Generalizing Abstract Representations from Concrete Sequences

Chemical Language Model Linker: blending text and molecules with modular adapters

Context-Augmented Code Generation Using Programming Knowledge Graphs

Convex Markov Games: A New Frontier for Multi-Agent Reinforcement Learning

FrameBridge: Improving Image-to-Video Generation with Bridge Models

Accurate and Regret-aware Numerical Problem Solver for Tabular Question Answering

MoH: Multi-Head Attention as Mixture-of-Head Attention

On Information-Theoretic Measures of Predictive Uncertainty

Compositional Shielding and Reinforcement Learning for Multi-Agent Systems

Advanced Gesture Recognition in Autism: Integrating YOLOv7, Video Augmentation and VideoMAE for Video Analysis

Upcycling Large Language Models into Mixture of Experts

Geometric Representation Condition Improves Equivariant Molecule Generation

How Much Can We Forget about Data Contamination?

On the Feasibility of Fully AI-automated Vishing Attacks

Improving Clinical Note Generation from Complex Doctor-Patient Conversation

Toward End-to-End Bearing Fault Diagnosis for Industrial Scenarios with Spiking Neural Networks

AgentCourt: Simulating Court with Adversarial Evolvable Lawyer Agents

A Survey of Text-to-SQL in the Era of LLMs: Where are we, and where are we going?

VideoQA in the Era of LLMs: An Empirical Study

TrialBench: Multi-Modal Artificial Intelligence-Ready Clinical Trial Datasets

The Remarkable Robustness of LLMs: Stages of Inference?

Navigating LLM Ethics: Advancements, Challenges, and Future Directions

OR-Bench: An Over-Refusal Benchmark for Large Language Models

DP-IQA: Utilizing Diffusion Prior for Blind Image Quality Assessment in the Wild

Does learning the right latent variables necessarily improve in-context learning?

Consistency of Neural Causal Partial Identification

Generalized Multi-Objective Reinforcement Learning with Envelope Updates in URLLC-enabled Vehicular Networks

Aptly: Making Mobile Apps from Natural Language

Personalized Wireless Federated Learning for Large Language Models

Unifying Lane-Level Traffic Prediction from a Graph Structural Perspective: Benchmark and Baseline

Counterfactual contrastive learning: robust representations via causal image synthesis

On the Completeness of Invariant Geometric Deep Learning Models

A Comprehensive Survey on Vector Database: Storage and Retrieval Technique, Challenge

Causal Deep Learning

MogaNet: Multi-order Gated Aggregation Network

Conformal Risk Control

Regret Minimization and Convergence to Equilibria in General-sum Markov Games

Towards a Cascaded LLM Framework for Cost-effective Human-AI Decision-Making

A Benchmark for Generalizing Across Diverse Team Strategies in Competitive Pok\'emon

Unreal Patterns

Control Tax: The Price of Keeping AI in Check

General agents need world models

Distinguishing Autonomous AI Agents from Collaborative Agentic Systems: A Comprehensive Framework for Understanding Modern Intelligent Architectures

Multi-RAG: A Multimodal Retrieval-Augmented Generation System for Adaptive Video Understanding

Beyond Chemical QA: Evaluating LLM's Chemical Reasoning with Modular Chemical Operations

Applying Cognitive Design Patterns to General LLM Agents

Requirements for Recognition and Rapid Response to Unfamiliar Events Outside of Agent Design Scope

Affordable AI Assistants with Knowledge Graph of Thoughts

Engineering Scientific Assistants using Interactive Structured Induction of Programs

Reference-Aligned Retrieval-Augmented Question Answering over Heterogeneous Proprietary Documents

Collaborative Memory: Multi-User Memory Sharing in LLM Agents with Dynamic Access Control

Created by

Haebom

저자

Alireza Rezazadeh, Zichao Li, Ange Lou, Yuying Zhao, Wei Wei, Yujia Bao

개요

본 논문은 여러 사용자와 여러 에이전트 환경에서 비대칭적이고 시간에 따라 변화하는 접근 제어를 갖는 협업 메모리 프레임워크를 제시합니다. 이 프레임워크는 개별 사용자에게만 보이는 개인 메모리와 선택적으로 공유되는 공유 메모리의 두 가지 메모리 계층을 유지합니다. 각 메모리 단편은 변경 불가능한 출처 속성(기여 에이전트, 접근한 리소스, 타임스탬프)을 가지며, 이를 통해 소급적인 권한 확인을 지원합니다. 세분화된 읽기 정책은 현재 사용자-에이전트-리소스 제약 조건을 적용하고 기존 메모리 단편을 필터링된 변환된 보기로 투영합니다. 쓰기 정책은 메모리 업데이트에 컨텍스트 인식 변환을 적용하여 단편 유지 및 공유를 결정합니다. 읽기 및 쓰기 정책 모두 시스템, 에이전트 및 사용자 수준 정보를 조건으로 설계될 수 있습니다. 이 프레임워크는 비대칭적이고 시간에 따라 변화하는 정책을 준수하고 메모리 작업을 완벽하게 감사할 수 있도록 하여 안전하고 효율적이며 해석 가능한 사용자 간 지식 공유를 가능하게 합니다.

시사점, 한계점

•

시사점:

◦

비대칭적이고 시간에 따라 변화하는 접근 제어를 갖는 다중 사용자, 다중 에이전트 환경에서 안전하고 효율적인 지식 공유 가능성 제시

◦

메모리 작업의 완벽한 감사 기능을 통한 투명성 및 신뢰성 향상

◦

개인 메모리와 공유 메모리의 이중 계층 구조를 통한 유연한 접근 제어

◦

컨텍스트 인식 변환을 통한 지능적인 메모리 관리

•

한계점:

◦

제안된 프레임워크의 실제 구현 및 성능 평가에 대한 정보 부족

◦

복잡한 접근 제어 정책 설계 및 관리의 어려움

◦

대규모 환경에서의 확장성 문제

◦

다양한 유형의 데이터와 에이전트에 대한 적용성 검증 부족

Made with Slashpage