KOINEU

One Layer Is Enough: Adapting Pretrained Visual Encoders for Image Generation

본 논문은 최근 시각 생성 모델이 고차원 이미지 특징을 직접 활용하기보다 압축된 잠재공간에서 작동함으로써 학습 효율성을 높이고 샘플 품질을 유지하려는 흐름을 정확히 짚어낸다. 그러나 사전 학습된 비전 트랜스포머(DINO, SigLIP 등)와 같은 강력한 이해‑지향 특징은 일반적으로 높은 차원을 유지하며, 이는 마스크된 영역에 대한 다양한 가설을 동시에 표현할 수 있다는 장점이 있다. 반면, 확산 모델이나 정규화 흐름과 같은 생성 모델은 노이즈를 정확히 전달하고 점진적으로 샘플을 복원해야 하므로, 차원이 낮고 연속적인 잠재공간을 요구

One Layer Is Enough: Adapting Pretrained Visual Encoders for Image Generation

Over-the-Air Federated Learning: Rethinking Edge AI Through Signal Processing

Self-Transparency Failures in Expert-Persona LLMs: How Instruction-Following Overrides Disclosure

Simultaneous Image Quality Improvement and Artefacts Correction in Accelerated MRI

STELLA: Guiding Large Language Models for Time Series Forecasting with Semantic Abstractions

Stochasticity in Agentic Evaluations: Quantifying Inconsistency with Intraclass Correlation

SynCraft: Guiding Large Language Models to Predict Edit Sequences for Molecular Synthesizability Optimization

UnwrapDiff: A Conditional Diffusion Model for InSAR Phase Unwrapping

EfficientFlow: Efficient Equivariant Flow Policy Learning for Embodied AI

Feasibility of Radio Frequency Based Wireless Sensing of Lead Contamination in Soil

BookRAG: A Hierarchical Structure-aware Index-based Approach for Retrieval-Augmented Generation on Complex Documents

Cross-Language Bias Examination in Large Language Models

Generative Adversarial Gumbel MCTS for Abstract Visual Composition Generation

Leveraging Spreading Activation for Improved Document Retrieval in Knowledge-Graph-Based RAG Systems

OptPO: Optimal Rollout Allocation for Test-time Policy Optimization

CoPHo: Classifier-guided Conditional Topology Generation with Persistent Homology

Flux-Preserving Adaptive Finite State Projection for Multiscale Stochastic Reaction Networks

MetaHGNIE: Meta-Path Induced Hypergraph Contrastive Learning in Heterogeneous Knowledge Graphs

OPAL: Operator-Programmed Algorithms for Landscape-Aware Black-Box Optimization

STAR: Semantic-Traffic Alignment and Retrieval for Zero-Shot HTTPS Website Fingerprinting

Systematization of Knowledge: Security and Safety in the Model Context Protocol Ecosystem

Variance-Aware Prior-Based Tree Policies for Monte Carlo Tree Search

Context-Sensitive Abstractions for Reinforcement Learning with Parameterized Actions

Detecting Perspective Shifts in Multi-agent Systems

Mage: Cracking Elliptic Curve Cryptography with Cross-Axis Transformers

QGShap: Quantum Acceleration for Faithful GNN Explanations

SA-IQA: Redefining Image Quality Assessment for Spatial Aesthetics with Multi-Dimensional Rewards

Social Comparison without Explicit Inference of Others' Reward Values: A Constructive Approach Using a Probabilistic Generative Model

The Silent Scholar Problem: A Probabilistic Framework for Breaking Epistemic Asymmetry in LLM Agents

A Multi-agent Text2SQL Framework using Small Language Models and Execution Feedback

ioPUF+: A PUF Based on I/O Pull-Up/Down Resistors for Secret Key Generation in IoT Nodes

Open-Ended Goal Inference through Actions and Language for Human-Robot Collaboration

A Time-efficient Prioritised Scheduling Algorithm to Optimise Initial Flock Formation of Drones

AI/ML in 3GPP 5G Advanced -- Services and Architecture

BEACON: A Unified Behavioral-Tactical Framework for Explainable Cybercrime Analysis with Large Language Models

rSIM: Incentivizing Reasoning Capabilities of LLMs via Reinforced Strategy Injection

Schoenfeld's Anatomy of Mathematical Reasoning by Language Models

TableGPT-R1: Advancing Tabular Reasoning Through Reinforcement Learning

A Multi-objective Optimization Approach for Feature Selection in Gentelligent Systems

AraToken: Optimizing Arabic Tokenization with Normalization Pipeline and Language Extension for Qwen3

Collaborative Edge-to-Server Inference for Vision-Language Models

CoAgent: Collaborative Planning and Consistency Agent for Coherent Video Generation

World Models That Know When They Don't Know: Controllable Video Generation with Calibrated Uncertainty

Comparative Analysis of 47 Context-Based Question Answer Models Across 8 Diverse Datasets

정보 흐름 발산을 이용한 필터·레이어 통합 압축 기법

Memories Retrieved from Many Paths: A Multi-Prefix Framework for Robust Detection of Training Data Leakage in Large Language Models

RevFFN: Memory-Efficient Full-Parameter Fine-Tuning of Mixture-of-Experts LLMs with Reversible Blocks

Probing the effectiveness of World Models for Spatial Reasoning through Test-time Scaling

Intelligent Knowledge Mining Framework: Bridging AI Analysis and Trustworthy Preservation

Towards Mass Spectrum Analysis with ASP

< Category Statistics (Total: 5012) >

Start searching

No results found