Jinheon Baek

CV (Curriculum Vitae)

About Me

Hi, I’m a master’s student in the Graduate school of AI at KAIST (MLAI Lab), where I am fortunate to be advised by Prof. Sung Ju Hwang. I am interested in Machine Learning for Graphs. Prior to studying at KAIST, I obtained my B.S. (Computer Science and Engineering) and B.E. (Software Technology and Enterprise Program) degrees at Korea University.

Also, I was a technical director of Jobshopper ((주)잡쇼퍼), an AI-based education startup company in Seoul, Korea. During one year of full-time work experience at Jobshopper as a lead engineer and researcher, I led the development and research groups of more than 10 people. Now, I am technically advising Jobshopper, as well as MajorMap (메이저맵(주)) spin-off from Jobshopper.

Research Interests

My research interest is mainly on developing novel machine learning models and algorithms for graphs. I focus on representing or generating graph-structured data such as a molecule, society, and neural network, modeling interactions between knowledge from relational-structured or even raw data with graphs, and tackling challenges on real-world graphs having noise or temporality.


KAIST (Korea Advanced Institute of Science and Technology)

  • M.S. student in Artificial Intelligence, Mar 2020 - Present.
  • GPA: 4.22 / 4.30 (99.1 / 100)

Korea University

  • Bachelor of Science in Computer Science and Engineering, Mar 2016 - Feb 2020.
  • Bachelor of Engineering in Software Technology and Enterprise Program (Interdisciplinary Program), Aug 2017 - Feb 2020.
  • GPA: 4.40 / 4.50 (98.9 / 100), GPA of Computer Science and Engineering: 4.48 / 4.50 (99.5 / 100)



  • Jaehyeong Jo*, Jinheon Baek*, Seul Lee*, Dongki Kim, Minki Kang, and Sung Ju Hwang. Edge Representation Learning with Hypergraphs. arXiv:2106.15845. [paper] (*: equal contribution)
  • Wonyong Jeong*, Hayeon Lee*, Gun Park*, Eunyoung Hyung, Jinheon Baek, and Sung Ju Hwang. Task-Adaptive Neural Network Retrieval with Meta-Contrastive Learning. arXiv:2103.01495. [paper] (*: equal contribution)


  • Jinheon Baek*, MinKi Kang*, and Sung Ju Hwang. Accurate Learning of Graph Representations with Graph Multiset Pooling. International Conference on Learning Representations (ICLR), 2021. [paper] [code] (*: equal contribution)
  • Jinheon Baek, Dong Bok Lee, and Sung Ju Hwang. Learning to Extrapolate Knowledge: Transductive Few-shot Out-of-Graph Link Prediction. Conference on Neural Information Processing Systems (NeurIPS), 2020. [paper] [code]


  • Hyunjae Kim*, Yookyung Koh*, Jinheon Baek, and Jaewoo Kang. Exploring The Spatial Reasoning Ability of Neural Models in Human IQ Tests. Neural Networks, 2021. [paper] (*: equal contribution)


  • Soyeong Jeong, Jinheon Baek, ChaeHun Park, and Jong C. Park. Unsupervised Document Expansion for Information Retrieval with Stochastic Text Generation. Scholarly Document Processing at Conference of the North American Chapter of the Association for Computational Linguistics (SDP@NAACL), 2021. [paper] (Oral Presentation)

Domestic (Korean) Journals & Conferences

  • Jinheon Baek, Hayeon Kim, and Kiwon Kwon. Artificial Intelligence-Based High School Course and University Major Recommendation System for Course-Related Career Exploration. KIPS Transactions on Software and Data Engineering (KTSDE), 2021.
  • Jinheon Baek, Minki Kang, and Sung Ju Hwang. Graph Representation Learning with Attention-based Set Pooling. Conference of Korean Artificial Intelligence Association (CKAIA), 2020. (Best Paper)
  • Jinheon Baek, Gwanghoon Jang, Soyeong Jeong, Donghyeon Park, Kiwon Kwon, and Jaewoo Kang. Embedding Academic Majors and Lectures for Analyzing Departments in University. Korea Computer Congress (KCC), 2019. (Oral Presentation)
  • Donghyeon Park, Yonggyu Park, Buru Chang, Jinheon Baek, and Jaewoo Kang. Embedding Food Ingredients Based on Chemical Combination in Dense Vector Space. Korea Computer Congress (KCC), 2018. (Best Paper)
  • Jinheon Baek, Dakyeong Lee, Chaeyeon Hong, and Byeongtae Ahn. Multimodal Approach for Blocking Obscene and Violent Contents. Journal of Convergence for Information Technology (JCIT), 7(6), 2017.


  • Kiwon Kwon, and Jinheon Baek. Method for Consulting on Exploration Activities based on Target Departments and Curriculum Achievement Standards. Korean patent number: 10-2021-0017764 (filed on Feb. 08, 2021).
  • Kiwon Kwon, and Jinheon Baek. Method for Building an Artificial Intelligence based Research Topics Database. Korean patent number: 10-2021-0017765 (filed on Feb. 08, 2021).

Academic Services

Conference Reviewers

  • International Conference on Learning Representations (ICLR), 2022
  • Neural Information Processing Systems (NeurIPS), 2021
  • International Conference on Machine Learning (ICML), 2021 - Selected as one of best reviewers (Top 10%)

You can see more details about me in CV (Curriculum Vitae).