Byoungjip Kim (김병집)

"Try not to become a man of success, but rather try to become a man of value."   -- Albert Einstein

Recent Research Topics

• Large Language Models (LLMs), Code LLMs, LLM Agents, LLM Alignment

• Reinforcement Learning (RL)

• Self-Supervised Learning (SSL)

• Generative Models (GANs, VAEs, Diffusion Models)

Education

• 2004.03 ~ 2013.02: Ph.D. in Computer Science, KAIST (Advisor: Prof. Junehwa Song)

• 2002.03 ~ 2004.02: M.S. in Electrical and Electronic Engineering, KAIST (Advisor: Prof. Daeyeon Park)

• 1998.03 ~ 2002.02: B.S. in Electrical and Electronic Engineering, KAIST (Korea Advanced Institute of Science and Technology)

• 1996.03 ~ 1998.02: Hansung Science High School

Professional Experiences

• 2021.08 ~ present: Research Scientist (& Code LLM Squad Leader) at LG AI Research, Advanced ML Lab.

• 2018.12 ~ 2021.07: Principal Engineer at Samsung SDS, AI Research Center, AI Core Lab.

• 2018.05 ~ 2018.11: Chief AI Officer at Speclipse Inc.

• 2017.03 ~ 2018.05: Research Scientist at SK Telecom, T-Brain

• 2015.09 ~ 2017.02: Assistant Professor at Sangmyung University, Department of Information and Communication Engineering

• 2013.05 ~ 2015.08: Senior Software Engineer at Samsung Electronics, Mobile Communication Division, Health Service Lab.

• 2005.06 ~ 2006.06: Research Intern at IBM T.J. Watson Research Center, Enterprise Linux Group

Recent Publications

• [Multimodal] Show, Think, and Tell: Thought-Augmented Fine-Tuning of Large Language Models for Video Captioning

  • Byoungjip Kim, Dasol Hwang, Sungjun Cho, Youngsoo Jang, Honglak Lee, Moontae Lee

  • CVPR 2024 Workshop Multi-Modal Foundation Models (MMFM) (accepted)

• [LLM] Degeneration-free Policy Optimization: RL Fine-Tuning of Language Models without Degeneration 

  • Youngsoo Jang, Geon-Hyeong Kim, Byoungjip Kim, Yu Jin Kim, Honglak Lee, Moontae Lee

  • ICML 2024 (accepted)

  • ICLR 2024 Workshop GenAI4DM (accepted)

• [LLM] On Sample-Efficient Code Generation [pdf] 

  • Hojae Han, Yu Jin Kim, Byoungjip Kim, Youngwon Lee, Kyungjae Lee, Kyungmin Lee, Moontae Lee, Kyunghoon Bae, Seung-won Hwang

  • EMNLP 2023 (Industry Track)

• [LLM] Prospector: Improving LLM Agents with Self-Asking and Trajectory Ranking [pdf] [blog (en)] [blog (kr)]  

  • Byoungjip Kim, Youngsoo Jang, Lajanugen Logeswaran, Geon-Hyeong Kim, Yu Jin Kim, Honglak Lee, Moontae Lee

  • NeurIPS 2023 Workshop Foundation Models for Decision Making (FMDM)

• [RL] SafeDICE: Offline Safe Imitation Learning with Non-Preferred Demonstrations [pdf] [video]   

  • Youngsoo Jang, Geon-Hyeong Kim, Jongmin Lee, Sungryull Sohn, Byoungjip Kim, Honglak Lee, Moontae Lee

  • NeurIPS 2023

• [Multimodal] Transferring Pre-trained Multimodal Representations with Cross-modal Similarity Matching [pdf] [video] [blog (en)] [blog (kr)]  

  • Byoungjip Kim, Sungik Choi, Dasol Hwang, Moontae Lee, Honglak Lee

  • NeurIPS 2022

• [CV] ContraCluster: Learning to Classify without Labels by Contrastive Self-Supervision and Prototype-Based Semi-Supervision [pdf] 

  • Seongho Joe, Byoungjip Kim, Hoyoung Kang, Kyoungwon Park, Bogun Kim, Jaeseon Park, Joonseok Lee, Youngjune Gwon

  • IEEE ICPR 2022

• [CV] VaB-AL: Incorporating Class Imbalance and Difficulty with Variational Bayes for Active Learning [pdf] 

  • Jongwon Choi, Kwang Moo Yi, Jihoon Kim, Jincho Choo, Byoungjip Kim, Jin-Yeop Chang, Youngjune Gwon, Hyung Jin Chang

  • CVPR 2021

• [CV] SelfMatch: Combining Contrastive Self-Supervision and Consistency for Semi-Supervised Learning [pdf] 

  • Byoungjip Kim, Jinho Choo, Yeong-Dae Kwon, Seongho Joe, Seungjai Min, Youngjune Gwon

  • NeurIPS 2020 Workshop Self-Supervised Learning

• [RL] POMO: Policy Optimization with Multiple Optima for Reinforcement Learning [pdf] [code] 

  • Yeong-Dae Kwon, Jinho Choo, Byoungjip Kim, Iljoo Yoon, Seungjai Min, Youngjune Gwon

  • NeurIPS 2020

• [GAN] DefogGAN: Predicting Hidden Information in the StarCraft Fog of War with Generative Adversarial Nets [pdf] [code] 

  • Yonghyun Jeong, Hyunjin Choi, Byoungjip Kim, Youngjune Gwon

  • AAAI 2020 (Oral)

• [GAN] Unsupervised Visual Attribute Transfer with Reconfigurable Generative Adversarial Networks [pdf] 

  • Taeksoo Kim, Byoungjip Kim, Moonsu Cha, Jiwon Kim

  • arXiv 1707.09798

Other Publications

Google Scholar 

Acedemic Services

• NeurIPS Reviewer

• ICLR Reviewer

• ICML Reviewer

• CVPR Reviewer

• ECCV Reviewer

• ACL Reviewer

• ACM IMWUT Associated Editor

• IEEE MDM 2017 Finance Chair

Awards

• 2023 LG AI Research Best Challenge Award

• 2020 Samsung Best Paper Award Bronze

• IEEE MDM 2006 Best Paper Award