Byeonghu Na

Postdoc@Korea Advanced Institute of Science & Technology·韩国·OpenReview

研究方向

generative models · conditional generative models · variational autoencoders · diffusion models · scene text recognition · semi-supervised learning · positive-unlabeled learning

6.0

AMiD: Knowledge Distillation for LLMs with $\alpha$-mixture Assistant Distribution

ICLR 2026Poster

5.6

Missing Pattern Recognized Diffusion Imputation Model for Missing Not at Random

ICLR 2026Rejected

5.5

Semantic-aware Wasserstein Policy Regularization for Large Language Model Alignment

ICLR 2026Poster

一作

5.5

Distillation of Large Language Models via Concrete Score Matching

ICLR 2026Poster

4.7

AC-Sampler: Accelerate and Correct Diffusion Sampling with Metropolis-Hastings Algorithm

合作者 (20)

Byeonghu Na

AMiD: Knowledge Distillation for LLMs with $\alpha$-mixture Assistant Distribution

Missing Pattern Recognized Diffusion Imputation Model for Missing Not at Random

Semantic-aware Wasserstein Policy Regularization for Large Language Model Alignment

Distillation of Large Language Models via Concrete Score Matching

AC-Sampler: Accelerate and Correct Diffusion Sampling with Metropolis-Hastings Algorithm

Preference Optimization by Estimating the Ratio of the Data Distribution

Diffusion Bridge AutoEncoders for Unsupervised Representation Learning

Diffusion Adaptive Text Embedding for Text-to-Image Diffusion Models

Training-Free Safe Text Embedding Guidance for Text-to-Image Diffusion Models