2021.02 Excellent Degree Thesis Award

SNU CSE awards the prize to graduated student.

The list below is the winner of '2021.02 Excellent Degree Thesis Award'.

We have 3 Ph.D. degree, 3 M.S. degree, 1 B.S degree winner.

o Ph.D degree: Yu YoungJae (advisor professor: Gunhee Kim)
Title: Large Scale Video Understanding with Narrative Description

Youngjae Yu contributed to artificial intelligence(AI) technology to explain images/audio/videos in language or answer questions, and improved image recognition ability by applying self-supervised learning. He has published a number of papers in CVPR, ECCV, AAAI, and ICLR, which are the best academic societies on computer vision and artificial intelligence.

o Ph.D degree: Jeong Eun Ji (Academic advisor: Byung-Gon Chun)
Title: Unifying Imperative and Symbolic Deep Learning Execution

Eunji Jeong researched new deep learning software technology that can easily program and execute artificial intelligence deep learning models quickly, and presented research papers as first authors at NSDI and EuroSys, the best academic conferences in the computer system field.

o Ph.D. degree: Yang Youngseok (Academic advisor: Byung-Gon Chun)
Title: A Flexible Architecture for Optimizing Distributed Data Processing

Youngseok Yang researched a flexible architecture for optimizing distributed data processing and presented his thesis at ATC and EuroSys, the best academic conferences in the field of computer systems. The research result was selected as the Apache Foundation Incubator project, and was selected as the official runner of Apache Beam, a Google Cloud big data system.

o M.S degree: Kim Dongjoo (Academic advisor: Gunhee Kim)
Title: Natural Audio Captioning

Dongjoo Kim conducted various studies in the field of artificial intelligence during his Master's degree. For the first time, he developed a model that can explain the sounds of nature in human language and research to overcome continuous learning in an adverse environment with unbalanced data. He also published three top-tier conference papers. Two as first-author and one second-author at NAACL, ICML, ECCV.

o M.S degree: Lee Junhee (Academic advisor: Taekyung Kwon)
Title: A TLS Downgrade Attack using Spatial Differences in Security Configurations

Junhee Lee proposed a TLS downgrade attack using spatial differences in security settings that can occur when large-scale web services in a distributed environment forge security settings to third parties such as CDNs. In addition, more than 7 million domains around the world have been analyzed for real threats, and the related research results will be presented at ASIACCS 2021, an excellent academic conference in the field of security.

o M.S degree: Kim Jongseok (Academic advisor: Gunhee Kim)
Title: Dual Composition Network in Interactive Image Retrieval

Jong-Seok Kim conducted research on an interactive image search algorithm. We proposed a method to improve performance in complementary to the existing method, and based on this, we presented a paper at AAAI, an excellent academic conference in the field of artificial intelligence.

o B.S degree: Jung Jaehoon (Academic advisor: U Kang)
Title: T-GAP: Learning to Walk across Time for Temporal KnowledgeGraph Completion

Jaehoon Jung extracted structured information from natural language data in the form of a knowledge graph, and conducted research on knowledge graph exploration according to user queries. Based on the research, thesis was published in EMNLP, the best academic conference in the field of natural language processing, and a paper was submitted to KDD.

Friday, February 26th 2021