KOINEU

Emergent Bayesian Behaviour and Optimal Cue Combination in LLMs

본 연구는 최근 인공지능 분야에서 화두가 되고 있는 “대형 언어 모델(LLM)의 내재적 추론 메커니즘”을 정신물리학적 관점에서 조명한다는 점에서 학문적 의의가 크다. 먼저, 인간이 지각 과제에서 잡음이 섞인 신호를 베이지안 방식으로 통합한다는 오래된 심리학적 사실을 LLM에 적용하려는 시도는 기존의 “프롬프트 엔지니어링”이나 “파인튜닝” 중심의 평가와는 차별화된 접근이다. 연구진은 BayesBench라는 네 가지 멀티모달 크기 추정 과제를 설계했는데, 이는 텍스트와 이미지 두 가지 입력 모달을 동시에 제공함으로써 모델이 자연스럽게

Emergent Bayesian Behaviour and Optimal Cue Combination in LLMs

Theodosian: A Deep Dive into Memory-Hierarchy-Centric FHE Acceleration

Visual Categorization Across Minds and Models: Cognitive Analysis of Human Labeling and Neuro-Symbolic Integration

Well-quasi-orders on embedded planar graphs

Cyberswarm: a novel swarm intelligence algorithm inspired by cyber community dynamics

IoTEdu: Access Control, Detection, and Automatic Incident Response in Academic IoT Networks

Q2D2: A Geometry-Aware Audio Codec Leveraging Two-Dimensional Quantization

Design Space Exploration of DMA based Finer-Grain Compute Communication Overlap

Linear socio-demographic representations emerge in Large Language Models from indirect cues

Rethinking Supervised Fine-Tuning: Emphasizing Key Answer Tokens for Improved LLM Accuracy

A Knowledge-Based Language Model: Deducing Grammatical Knowledge in a Multi-Agent Language Acquisition Simulation

Benchmarking Deep Neural Networks for Modern Recommendation Systems

SynRAG: A Large Language Model Framework for Executable Query Generation in Heterogeneous SIEM System

Uncertainty-Gated Region-Level Retrieval for Robust Semantic Segmentation

Robust TTS Training via Self-Purifying Flow Matching for the WildSpoof 2026 TTS Track

What matters for Representation Alignment: Global Information or Spatial Structure?

CycliST: A Video Language Model Benchmark for Reasoning on Cyclical State Transitions

Harnessing Large Language Models for Biomedical Named Entity Recognition

Theory of Mind for Explainable Human-Robot Interaction

시각적 지식 그래프를 활용한 대형 언어 모델 환각 탐지 및 인간‑인‑루프 피드백 프레임워크

Thinking on Maps: How Foundation Model Agents Explore, Remember, and Reason Map Environments

텍스트 프롬프트 기반 의료 영상 분할 모델 MedSAM3

NeSTR: A Neuro-Symbolic Abductive Framework for Temporal Reasoning in Large Language Models

Reeb Graph of Sample Thickenings

Warp-Cortex: An Asynchronous, Memory-Efficient Architecture for Million-Agent Cognitive Scaling on Consumer Hardware

CAuSE: Decoding Multimodal Classifiers using Faithful Natural Language Explanation

UniCoMTE: A Universal Counterfactual Framework for Explaining Time-Series Classifiers on ECG Data

소셜네트워크 기반 관광객 이동 예측을 위한 문법추론 히든마코프모델

Hands-on Evaluation of Visual Transformers for Object Recognition and Detection

Oogiri-Master: Benchmarking Humor Understanding via Oogiri

State over Tokens: Characterizing the Role of Reasoning Tokens

Comparison of different segmentation algorithms on brain volume and fractal dimension in infant brain MRIs

Predicting Startup-VC Fund Matches with Structural Embeddings and Temporal Investment Data

The changing surface of the world's roads

CATCH: A Controllable Theme Detection Framework with Contextualized Clustering and Hierarchical Generation

Dual-Stream Spectral Decoupling Distillation for Remote Sensing Object Detection

Exploring the Modular Integration of 'AI + Architecture' Pedagogy in Undergraduate Design Education: A Case Study of Architectural Design III/IV Courses at Zhejiang University

PARAN: Persona-Augmented Review ANswering system on Food Delivery Review Dataset

Rethinking Prompt Design for Inference-time Scaling in Text-to-Visual Generation

MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture

The Trust in AI-Generated Health Advice (TAIGHA) Scale and Short Version (TAIGHA-S): Development and Validation Study

Adversarially Probing Cross-Family Sound Symbolism in 27 Languages

An AI-Powered Autonomous Underwater System for Sea Exploration and Scientific Research

Web World Models

Multi-Agent Coordinated Rename Refactoring

Robust Uncertainty Quantification for Factual Generation of Large Language Models

Statistically-Guided Dual-Domain Meta-Learning with Adaptive Multi-Prototype Aggregation for Distributed Fiber Optic Sensing

Benchmarking Preprocessing and Integration Methods in Single-Cell Genomics

Evaluating the Impact of Compression Techniques on the Robustness of CNNs under Natural Corruptions

Benchmark Success, Clinical Failure: When Reinforcement Learning Optimizes for Benchmarks, Not Patients

< Category Statistics (Total: 5012) >

Start searching

No results found