Soohan Lim

About

I am an Integrated MS/PhD student in Computer Science at the Theory of Computation (ToC) Lab, Yonsei University. My research focuses on program analysis and code intelligence, particularly time complexity prediction, test case generation, contract-satisfying code generation, and multi-agent debate frameworks. I explore how large language models (LLMs) can reason about algorithmic properties and generate specification-aware test suites, integrating chain-of-thought prompting, reinforcement learning, and self-play mechanisms to improve the reliability and coverage of generated test cases.

Recent News

  • NEW New preprint: "SLICE: Specification-Level Isolation of Contract Enforcement" is available on arXiv. (Aug 2026)
  • Our paper "K-SpecDetox" has been accepted to KCC 2026 (Poster). (Jun 2026)
  • preprint: "STAB: Specification-driven Testing for Algorithmic Bottlenecks" is available on arXiv. (May 27, 2026)
  • Our paper "ContractEval" has been accepted to Findings of ACL 2026. (Jan 8, 2026)
  • preprint: "MEC³O: Multi-Expert Consensus for Code Time Complexity Prediction" is available on arXiv. (Oct 10, 2025)
  • Our paper "TCProF" was presented at NAACL 2025 (Oral). (2025)

Research Interests

  • Program analysis and code intelligence
  • Time complexity prediction and formal analysis of programs
  • Specification-driven test case generation using large language models
  • Contract-satisfying code generation from problem specifications
  • Multi-agent debate frameworks for reasoning and validation

Contact Information

Department of Computer Science, Yonsei University
50 Yonsei-Ro, Seodaemun-Gu, Seoul 03722, Republic of Korea
Email: aness1219 [at] yonsei [dot] ac [dot] kr

Last updated 2026/8/24