Jeongju Sohn

PhD Candidate
School of Computing
Korea Advanced Institute of Science and Technology
291 Daehak Ro, Yuseong Gu
Daejeon 34141
Republic of Korea

kasio555@kaist.com

Research Interests

Software engineering, Search-based software engineering, Genetic Programming, Fault Localisation, Defect Prediction

Education

• BSc., Computer Science and Engineering, Ewha Womans University, Republic of Korea, Mar. 2011 - Feb. 2015
• MSc., Computer Science, Korean Advanced Institute of Science and Technology, Republic of Korea, Mar. 2015 - Feb. 2017

Experience

• Database and Multimedia Lab, KAIST, Mar. 2015 - Jan. 2016
  • Research on parallel processing architecture that intergrates the relational DBMS with the DFS as a single system for big data analysis
• Computational Intelligence for Software Engineering Labroatory(COINSE), KAIST, Feb. 2016 - Present
  • Research on parameter optimization on ‘OpenCV’, using amortized deep parameter optimization
  • Research on finding possible links between fault localization and defect prediction to improve the performance of fault localization

Publication

• Sohn, J., An, G., Hong, J., Hwang., D, Yoo, S. (2021). Assisting Bug Report Assignment Using Automated Fault Localisation: An Industrial Case Study, In Proceedings of the International Conference on Software Testing (ICST) (pp. to appear). Virtual

• Sohn, J., Kamei, Y., McIntosh, S., Yoo, S. (2021). Leveraging Fault Localisation to Enhance Defect Prediction, In Proceedings of the IEEE 28th International Conference on Software Analysis, Evolution, and Reengineering (SANER) (pp. to appear). Virtual % Honolulu, Hawai

• Sohn, J. (2020). Bridging Fault Localisation and Defect Prediction, In Proceedings of the ACM/IEEE 42nd International Conference on Software Engineering: Companion Proceedings (pp. 214–217). Virtual

• Sohn, J. & Yoo, S. (2019). Empirical Evaluation of Fault Localisation Using Code and Change Metrics. IEEE Transactions on Software Engineering

• Sohn, J. & Yoo, S. (2019). Why Train-and-Select When You Can Use Them All: Ensemble Model for Fault Localisation. In Proceedings of the Genetic and Evolutionary Computation Conference. Prague, Czech

• Choi, K, Sohn, J., & Yoo, S. (2018). Learning Fault Localisation for Both Humans and Machines using Multi-Objective GP. In Proceeding of the Search Based Software Engineering. Montpellier, France

• Kang, D., Sohn, J., & Yoo, S. (2017). Empirical evaluation of conditional operators in GP based fault localization. In Proceedings of the Genetic and Evolutionary Computation Conference. Berlin, Germany

• Sohn, J. & Yoo, S. (2017). Using Code and Change Metrics to Improve Fault Localization. In Proceeding of the 2017 International Symposium on Softare Testing and Analysis. Santa Barbara, CA, USA

• Sohn, J., Lee, S., & Yoo, S. (2016). Amortised Deep Parameter Optimisation of GPGPU Work Group Size for OpenCV. In Proceeding of the 2016 Search Based Software Engineering (pp. 211–217). Raleigh, NC, USA: ACM. doi:10.1007/978-3-319-47106-8_14

Projects

• ‘FLUCCS’: Using source code metrics, have been widely used in defect prediction, to improve Fault Localization; investigating possible relations between defect prediction and fault localization and applying revealed relations to fault localization for further improvement

• ‘FL2DP’: Leveraging fault localisation to improve defect prediction. FL2DP, which stands for Fault Localisation to Defect Prediction, aims to enhance Just-In-Time defect prediction by exploiting the insights about defective code from past fault localisation.

• ‘Arachne’: Search-based repair of deep neural network. Compared to previous automatic DNN repair technique, Arachne targets to repair a specific type of faults (e.g., classify input A to B) at the expense of the others. Arachne can be applied in the situations where some misbehaviour of a DNN must be fixed even at the cost of breaking other less critical behaviour.