Associate professor in the Kim Jaechul School of Artificial Intelligence and School of Computing at KAIST. Research interest is mainly on developing novel models and algorithms for tackling new challenges in deploying artificial intelligence systems to various real-world application domains.

KAIST 김재철 인공지능대학원 및 전산학부 부교수. 인공지능 시스템을 다양한 실제 응용 분야에 배치하는 데 있어서 새로운 과제를 해결하기 위한 새로운 모델과 알고리즘 개발에 주로 관심을 가지고 연구한다.

Continual Learning: Forget-free Winning Subnetworks for Video Representations

READRetro: Natural Product Biosynthesis Predicting with Retrieval-Augmented Dual-View Retrosynthesis

Learning the Compositional Domains for Generalized Zero-shot Learning

Learning the Relative Importance of Objects from Tagged Images for Retrieval and Cross-Modal Search

Reading Between the Lines: Object Localization Using Implicit Cues from Image Tags

Distilling LLM Agent into Small Models with Retrieval and Code Tools

Delta Attention: Fast and Accurate Sparse Attention Inference by Delta Correction

Continuous Diffusion Model for Language Modeling

Mol-LLaMA: Towards General Understanding of Molecules in Large Molecular Language Model

System Prompt Optimization with Meta-Learning

FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA

Cost-Sensitive Freeze-thaw Bayesian Optimization for Efficient Hyperparameter Tuning

Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching

Efficient Real-time Refinement of Language Model Text Generation

Database-Augmented Query Representation for Information Retrieval

Efficient Long Context Language Model Retrieval with Compression

VideoRAG: Retrieval-Augmented Generation over Video Corpus

SafeRoute: Adaptive Model Selection for Efficient and Accurate Safety Guardrails in Large Language Models

AutoML-Agent: A Multi-Agent LLM Framework for Full-Pipeline AutoML

Bayesian Neural Scaling Laws Extrapolation with Prior-Fitted Networks

VideoICL: Confidence-based Iterative In-context Learning for Out-of-Distribution Video Understanding

Silent Branding Attack: Trigger-free Data Poisoning Attack on Text-to-Image Diffusion Models

ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models

A Training-Free Sub-quadratic Cost Transformer Model Serving Framework with Hierarchically Pruned Attention

HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models

Learning Diverse Attacks on Large Language Models for Robust Red-Teaming and Safety Tuning

Training Free Exponential Context Extension via Cascading KV Cache

Diffusion-based Neural Network Weights Generation

KOALA: Empirical Lessons Toward Memory-Efficient and Fast Diffusion Models for Text-to-Image Synthesis

Set-based Neural Network Encoding Without Weight Tying

Latent Paraphrasing: Perturbation on Layers Improves Knowledge Injection in Language Models

Identity Decoupling for Multi-Subject Personalization of Text-to-Image Models

Concept-skill Transferability-based Data Selection for Large Vision-Language Models

Rethinking Code Refinement: Learning to Judge Code Efficiency

STELLA: Continual Audio-Video Pre-training with SpatioTemporal Localized Alignment

Generative Modeling on Manifolds Through Mixture of Riemannian Diffusion Processes

Graph Generation with Diffusion Mixture

Drug Discovery with Dynamic Goal-aware Fragments

BECoTTA: Input-dependent Online Blending of Experts for Continual Test-time Adaptation

EVEREST: Efficient Masked Video Autoencoder by Removing Redundant Spatiotemporal Tokens

One Prompt is not Enough: Automated Construction off a Mixture-of-Experts Prompts

Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexity

Carpe diem: On the Evaluation of World Knowledge in Lifelong Language Models

ECLIPSE: Efficient Continual Learning in Panoptic Segmentation with Visual Prompt Tuning

SEA: Sparse Linear Attention with Estimated Attention Mask

DiffusionNAG: Predictor-guided Neural Architecture Generation with Diffusion Models

Self-Supervised Dataset Distillation for Transfer Learning

Progressive Fourier Neural Representation for Sequential Video Compilation

Learning to Verify Knowledge-Augmented Language Models

Co-training and Co-distillation for Quality Improvement and Compression of Language Models

Test-Time Self-Adaptive Small Language Models for Question Answering

Knowledge-Augmented Reasoning Distillation for Small Language Models in Knowledge-Intensive Tasks

Generalizable Lightweight Proxy for Robust NAS against Diverse Perturbations

Effective Targeted Attacks for Adversarial Self-Supervised Learning

STXD: Structural and Temporal Cross-Modal Distillation for Multi-View 3D Object Detection

Text-Conditioned Sampling Framework for Text-to-Image Generation with Masked Generative Models

ZET-Speech: Zero-shot adaptive Emotion-controllable Text-to-Speech Synthesis with Diffusion and Style-based Models

Direct Fact Retrieval from Knowledge Graphs without Entity Linking

Language Detoxification with Attribute-Discriminative Latent Space

A Study on Knowledge Distillation from Weak Teacher for Scaling Up Pre-trained Language Model

Phrase Retrieval for Open Domain Conversational Question Answering with Conversational Dependency Modeling via Contrastive Learning

Exploring Chemical Space with Score-based Out-of-distribution Generation

Scalable Set Encoding with Universal Mini-Batch Consistency and Unbiased Full Set Gradient Approximation

Personalized Subgraph Federated Learning

Margin-based Neural Network Watermarking

Continual Learners are Incremental Model Generalizers

DAPPER: Label-Free Performance Estimation after Personalization for Heterogeneous Mobile Sensing

The Devil is in the Points: Weakly Semi-Supervised Instance Segmentation via Point-Guided Mask Representation

Any-speaker Adaptive Text-To-Speech Synthesis with Diffusion Models

Realistic Conversational Question Answering with Answer Selection based on Calibrated Confidence and Uncertainty Measurement

Meta-prediction Model for Distillation-Aware NAS on Unseen Datasets

Self-Distillation for Further Pre-training of Transformers

Sparse Token Transformers with Attention Back Tracking

Self-Supervised Set Representation Learning for Unsupervised Meta-Learning

Exploring the Role of Mean Teacher in Self-supervised Masked Auto-Encoders

On the Soft-Subnetwork for Few-Shot Class Incremental Learning

Rethinking the Entropy of Instance in Adversarial Training

Graph Self-supervised Learning with Accurate Discrepancy Learning

Set-based Meta-Interpolation for Few-Task Meta-Learning

Factorized-FL: Personalized Federated Learning with Parameter Factorization & Similarity Matching

Learning to Generate Inversion-Resistant Model Explanations

Score-based Generative Modeling of Graphs via the System of Stochastic Differential Equations

Bitwidth Heterogeneous Federated Learning with Progressive Weight Dequantization

Set Based Stochastic Subsampling

Forgetting-free Continual Learning with Winning Subnetworks

KALA: Knowledge-Augmented Language Model Adaptation

MPViT : Multi-Path Vision Transformer for Dense Prediction

Augmenting Document Representations for Dense Retrieval with Interpolation and Perturbation

Lifelong learning with dynamically expandable networks

Learning to propagate labels: Transductive propagation network for few-shot learning

Learning to quantize deep networks by optimizing quantization intervals with task loss

Mpvit: Multi-path vision transformer for dense prediction

Score-based generative modeling of graphs via the system of stochastic differential equations

Adaptive-rag: Learning to adapt retrieval-augmented large language models through question complexity

Federated continual learning with weighted inter-client transfer

Adversarial self-supervised contrastive learning

Self-supervised label augmentation via input transformations

Federated semi-supervised learning with inter-client consistency & disjoint learning

Accurate learning of graph representations with graph multiset pooling

Fedmix: Approximation of mixup under mean augmented federated learning

Distribution aligning refinery of pseudo-label for imbalanced semi-supervised learning

Adversarial purification with score-based generative models

Meta-stylespeech: Multi-speaker adaptive text-to-speech generation

Scalable and order-robust continual learning with additive parameter decomposition

Online coreset selection for rehearsal-based continual learning

Researchagent: Iterative research idea generation over scientific literature with large language models

Sung Ju Hwang

교수님 소개

황성주

연구실 소개

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

Distilling LLM Agent into Small Models with Retrieval and Code Tools

Delta Attention: Fast and Accurate Sparse Attention Inference by Delta Correction

Continuous Diffusion Model for Language Modeling

연구실 보유 특허

대학별 입시끈

우리 연구실 모집

교수님 소개

황성주

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

Distilling LLM Agent into Small Models with Retrieval and Code Tools

Delta Attention: Fast and Accurate Sparse Attention Inference by Delta Correction

Continuous Diffusion Model for Language Modeling

연구실 보유 특허

최근 3년간 연구실적 높은 연구실

대학별 입시끈

우리 연구실 모집