Pyo Min Hong

Halyn Yu

Beom Soo Lee

Junseok Jang

JiHyun Jung

Ki-wook Hong

Dayeon Kim

Beomseok Seo

Yebin Wang

Sanghoon Kang

Yunfei Gao

Minhyeok Cho

Jinhyuk Park

Haonan Yang

The Artificial Intelligence and Information Systems Laboratory (AI-ISL) focuses on various aspects of Artificial Intelligence and Information Systems, including Machine Learning, Data Mining, Natural Language Processing, and Human-Computer Interaction.

인공지능 및 정보 시스템 연구실(AI-ISL)은 머신러닝, 데이터 마이닝, 자연어 처리, 인간-컴퓨터 상호작용을 포함한 인공지능 및 정보 시스템의 다양한 측면에 중점을 둡니다.

Deterministic matrices matching the compressed sensing phase transitions of Gaussian random matrices

Markov Decision Policies for Dynamic Video Delivery in Wireless Caching Networks

Cooperative sequence clustering and decoding for DNA storage system with fountain codes

Prune Your Model Before Distill It

Information Measures: The Curious Case of the Binary Alphabet

Strong Successive Refinability and Rate-Distortion-Complexity Tradeoff

Reference based genome compression

Learned Smartphone ISP on Mobile GPUs with Deep Learning, Mobile AI &amp; AIM 2022 Challenge: Report

Rateless Lossy Compression via the Extremes

Error Rate-Based Log-Likelihood Ratio Processing for Low-Density Parity-Check Codes in DNA Storage

Universality of logarithmic loss in lossy compression

SPIHT Algorithm With Adaptive Selection of Compression Ratio Depending on DWT Coefficients

Reducing cost in DNA-based data storage by sequence analysis-aided soft information decoding of variable-length reads

An Information-Theoretic Justification for Model Pruning

Generative Adversarial Networks for DNA Storage Channel Simulator

PyNET-Q×Q: An Efficient PyNET Variant for Q×Q Bayer Pattern Demosaicing in CMOS Image Sensors

Rate-Distortion Theoretic Model Compression: Successive Refinement for Pruning

Age Estimation from fMRI Data Using Recurrent Neural Network

FCLQC: fast and concurrent lossless quality scores compressor

Universality of Logarithmic Loss in Successive Refinement

Title of Paper 1

Title of Paper 2

Title of Paper 3: A Deep Dive

The Confidence Shortcut: A Reasoning Failure Mode of Masked Diffusion Models

Where Rollouts Begin: Low-Load, High-Leverage First-Token Diversification for RLVR

VLMs Trace Without Tracking: Diagnosing Failures in Visual Path Following

Slower Generalization, Faster Memorization: A Sweet Spot in Algorithmic Learning

Few-Shot Truly Benign DPO Attack for Jailbreaking LLMs

BenchPreS: A Benchmark for Context-Aware Personalized Preference Selectivity of Persistent-Memory LLMs

A Theoretical Analysis of Why Masked Diffusion Models Mitigate the Reversal Curse

Position: The Term “Machine Unlearning” Is Overused in LLMs

dgMARK: Decoding-Guided Watermarking for Diffusion Language Models

Preserve-Then-Quantize: Balancing Rank Budgets for Quantization Error Reconstruction in LLMs

Dependency-Aware Parallel Decoding via Attention for Diffusion LLMs

DUSK: Do Not Unlearn Shared Knowledge

Memorization In Stable Diffusion Is Unexpectedly Driven by CLIP Embeddings

Rethinking Benign Relearning: Syntax as the Hidden Driver of Unlearning Failures

A2D: Any-Order, Any-Step Safety Alignment for Diffusion Language Models

Rainbow Padding: Mitigating Early Termination in Instruction-Tuned Diffusion LLMs

Differentially Private Adaptation of Diffusion Models via Noisy Aggregated Embeddings

An Information Theoretic Evaluation Metric For Strong Unlearning

Information-Theoretic Discrete Diffusion

SAFEPATH: Preventing Harmful Reasoning in Chain-of-Thought via Early Alignment

SEPS: A Separability Measure for Robust Unlearning in LLMs

R-TOFU: Unlearning in Large Reasoning Models

Assigning Distinct Roles to Quantized and Low-Rank Matrices Toward Optimal Weight Decomposition

Understanding and Mitigating Memorization in Generative Models via Sharpness of Probability Landscapes

Improved Communication-Privacy Trade-offs in L2 Mean Estimation under Streaming Differential Privacy

Censored Sampling of Diffusion Models Using 3 Minutes of Human Feedback

Exact Optimality of Communication-Privacy-Utility Tradeoffs in Distributed Mean Estimation

Neural Tangent Kernel Analysis of Deep Narrow Neural Networks

WGAN with an Infinitely Wide Generator Has No Spurious Stationary Points

CROMqs: Rateless Lossy Compression of Quality Scores

Universality of Logarithmic Loss in Lossy Compression

Information Divergences and the Curious Case of the Binary Alphabet

Strong Successive Refinability: Sufficient Conditions

Complexity and Rate-Distortion Tradeoff via Successive Refinement

Minimax Filtering via Relations between Information and Estimation

Reference Based Genome Compression

Joint Source-Channel Coding of one Random Variable over the Poisson Channel

Preserve then Quantize: Dominant-Subspace Guided Low-Rank Reconstruction

Understanding Memorization In Generative Models Through A Geometric Framework

Adversarial Sample-Based Approach for Tighter Privacy Auditing in Final Model-Only Scenarios

LoRA can Replace Time and Class Embeddings in Diffusion Probabilistic Models

Fully Quantized Always-on Face Detector Considering Mobile Image Sensors

Exact Optimality in Communication-Privacy-Utility Tradeoffs

Learned Smartphone ISP on Mobile GPUs with Deep Learning, Mobile AI & AIM 2022 Challenge: Report

Neural Network Pruning via Rate-Distortion Theory

A Simple Extension of Bayer Demosaicing For Lightweight Non-Bayer Demosaicing

Efficient and Low-Complexity Variable-to-VariableLength Coding for DNA Storage

Simple Drop-in LoRA Conditioning on Attention Layers Will Improve Your Diffusion Model

Needleman-Wunsch Attention: A Framework for Enhancing DNA Sequence Embedding

Adaptable DNA Storage Coding: An Efficient Framework for Homopolymer Constraint Transitions

How Does a Transformer Learn Compression? An Attention Study on Huffman and LZ4

PyNET-QxQ: An Efficient PyNET Variant for QxQ Bayer Pattern Demosaicing in CMOS Image Sensors

FCLQC: Fast and Concurrent Lossless Quality Scores Compressor

Cooperative Sequence Clustering and Decoding for DNA Storage System with Fountain Codes

Nonasymptotic Upper Bounds on Binary Single Deletion Codes via Mixed Integer Linear Programming

Universality of Logarithmic Loss in Fixed-Length Lossy Compression

SPIHT Algorithm with Adaptive Selection of Compression Ratio Depending on DWT Coefficients

Information Geometric Approach on Most Informative Boolean Function Conjecture

DNA 저장 장치의 연성 정보 기반 복호화 방법, 프로그램 및 장치

데이터 압축 및 압축해제방법

Impact of Mesoscale Circulations on Near-Inertial Wave Energy Flux in the Kuroshio Extension

Mesoscale Flow Impacts on Downward Energy Flux of Near-Inertial Waves in the Kuroshio Extension

Albert No

Comparison of Upper Ocean and Lower Atmosphere Temperatures over the Western North Pacific in Summer

Artificial Intelligence and Information Systems Laboratory (AI-ISL)

교수님 소개

노알버트

연구실 소개

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

Position: The Term “Machine Unlearning” Is Overused in LLMs

dgMARK: Decoding-Guided Watermarking for Diffusion Language Models

Preserve-Then-Quantize: Balancing Rank Budgets for Quantization Error Reconstruction in LLMs

연구실 보유 특허

대학별 입시끈

우리 연구실 모집

교수님 소개

노알버트

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

Position: The Term “Machine Unlearning” Is Overused in LLMs

dgMARK: Decoding-Guided Watermarking for Diffusion Language Models

Preserve-Then-Quantize: Balancing Rank Budgets for Quantization Error Reconstruction in LLMs

연구실 보유 특허

최근 3년간 연구실적 높은 연구실

대학별 입시끈

우리 연구실 모집