Biography

I am senior research scientist at Runway. I received my Ph.D. in Mathematics from Massachusetts Institute of Technology under the supervision of Prof. Michel X. Goemans. After completion of my Ph.D., I worked at Kakao Brain as a research scientist from November 2018 to February 2024. I joined Runway as a senior research scientist in March 2024.

Conference Publications

• Locality-Aware Generalizable Implicit Neural Representation
  Doyup Lee*, Chiheon Kim*, Minsu Cho, Wook-Shin Han
  Annual Conference on Neural Information Processing Systems (NeurIPS), 2023
  [paper] [code]

• Towards End-to-End Generative Modeling of Long Videos With Memory-Efficient Bidirectional Transformers
  Jaehoon Yoo, Semin Kim, Doyup Lee, Chiheon Kim, Seunghoon Hong
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023
  [paper]

• Generalizable Implicit Neural Representations via Instance Pattern Composers
  Chiheon Kim*, Doyup Lee*, Saehoon Kim, Minsu Cho, Wook-Shin Han
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023
  Spotlight (acceptance rate 2.5%)
  [paper] [code]

• Draft-and-Revise: Effective Image Generation with Contextual RQ-Transformer
  Doyup Lee*, Chiheon Kim*, Saehoon Kim, Minsu Cho, Wook-Shin Han
  Annual Conference on Neural Information Processing Systems (NeurIPS), 2022
  [paper]

• Locally Hierarchical Auto-Regressive Modeling for Image Generation
  Tackgeun You, Saehoon Kim, Chiheon Kim, Doyup Lee, Bohyung Han
  Annual Conference on Neural Information Processing Systems (NeurIPS), 2022
  [paper] [code]

• Autoregressive image generation using residual quantization
  Doyup Lee*, Chiheon Kim*, Saehoon Kim, Minsu Cho, Wook-Shin Han
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022
  [paper] [code]

• Automated learning rate scheduler for large-batch training
  Chiheon Kim, Saehoon Kim, Jongmin Kim, Donghoon Lee, Sungwoong Kim
  8th ICML Workshop on Automated Machine Learning (ICMLW-AutoML), 2021
  [paper] [code]

• Mining Gold Samples for Conditional GANs
  Sangwoo Mo, Chiheon Kim, Sungwoong Kim, Minsu Cho, Jinwoo Shin
  Annual Conference on Neural Information Processing Systems (NeurIPS), 2019
  [paper] [code]

• Fast Autoaugment
  Sungbin Lim, Ildoo Kim, Taesup Kim, Chiheon Kim, Sungwoong Kim
  Annual Conference on Neural Information Processing Systems (NeurIPS), 2019
  [paper] [code]

• Scalable Neural Architecture Search for 3d Medical Image Segmentation
  Sungwoong Kim, Ildoo Kim, Sungbin Lim, Woonhyuk Baek, Chiheon Kim, Hyungjoo Cho, Boogeon Yoon, Taesup Kim
  22nd International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 2019
  [paper] [arxiv]

• Community Detection in Hypergraphs, Spiked Tensor Models, and Sum-of-Squares
  Chiheon Kim, Afonso Bandeira, Michel X. Goemans
  International Conference on Sampling Theory and Applications (SampTA), 2017
  [paper] [arxiv]

Journal Publications

• New Classes of Set-sequential Trees
  Louis Golowich, Chiheon Kim
  Discrete Mathematics, Vol. 343, Issue 3, 2020
  [paper] [arxiv]

• Maximum Size of a Family of Pairwise Graph-Different Permutations
  Louis Golowich, Chiheon Kim, Richard Zhou
  Electornic Journal of Combinatorics, Vol. 24, Issue 4, 2017
  [paper]

Non-refereed Publications

• NVS-Adapter: Plug-and-Play Novel View Synthesis from a Single Image
  Yoonwoo Jeong, Jinwoo Lee, Chiheon Kim, Minsu Cho, Doyup Lee
  [paper]

• torchgpipe: On-the-fly Pipeline Parallelism for Training Giant Models
  Chiheon Kim*, Heungsub Lee*, Myungryong Jeong, Woonhyuk Baek, Boogeon Yoon, Ildoo Kim, Sungbin Lim, Sungwoong Kim
  Open-source, efficient PyTorch implementation of Pipeline Parallelism a.k.a. GPipe.
  Became a part of official PyTorch implementation.
  [paper] [code]

• Stochastic Block Model for Hypergraphs: Statistical Limits and a Semidefinite Programming Approach
  Chiheon Kim, Afonso Bandeira, Michel X. Goemans
  [paper]

Education

• Massachusetts Institute of Technology (MIT), Cambridge, MA, United States
  Ph.D. in Mathematics, Sep. 2012 - Sep. 2018
  Thesis: Statistical Limits of Graphical Channel Models and a Semidefinite Programming Approach
  Advisor: Michel X. Goemans

• Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea
  B.S. in Mathematical Sciences, Mar. 2006 - Feb. 2012
  Summa Cum Laude
  Thesis: On Vector Chromatic Dimension
  Advisor: Andreas Holmsen