1. Research Area
In modern society, our lives are closely influenced by software, and its influence is increasing. As a result, the quality of software also has a significant influence on our lives, and there are also a number of cases where software defects result in trillions of won in losses.
In order to reduce the probability of malfunction doe to software defects and the damage caused by malfunctions, safety standards are established that reflect the characteristics of various industry domain. However, despite these efforts, improving software quality is a difficult task that requires a lot of cost. Our lab is conducting research to improve software quality al low cost through software test automation research and software quality verification research.
2. Research Overview
- Deep Neural Network Evaluation based on Data and Structure
This research develops an integrated framework for evaluating DNN based on data and structure. In DNN evaluation based on data evaluates DNN performance by generating test data expressing various contexts using metamorphic relation based on learning data. In DNN evaluation based on structure studies the technique of evaluating DNN based on the operation information of the components(neurons, weights, etc.).
- Software Safety Assurance
This research studies and develops testing/verification technique to ensure the safety of software. We define formal modeling techniques that define safety elements and study techniques for automatically generating safety test cases using safety specifications. In addition, we research and develop technology to test unexpected safety problem that may arise from the interaction and operation of software functions.
3. Research Achievements
- J. H. Bae and H. S. Chae, "Systematic Approach for Constructing an Understandable State Machine from a Contract-based Specification - Technology-Oriented and Human-Oriented Controlled Experiments," Software and Systems Modeling, vol. 15, no. 3, pp. 847-879, July, 2016.
- Y. W. Choi and Y. W. Lee, H. S. Chae, "Development of Image Data Augmentation Tool for Evaluating CNN Model," Journal of the Software Engineering Society, vol. 29, no. 1, pp. 13-21, 2020.
- Y. W. Lee and H. S. Chae, "Patterns of Detecting Feature Interaction in Autonomous Car," Journal of KIISE, vol. 46, no. 10, pp. 1001-1011, 2019.
- J. H. Park and H. J. Choi, H. S. Chae, "Model Evaluation of the White Box Method by Correlation Analysis between the neuron active values and the Epoch of ResNet," Proceedings of the 22nd Korea Software Engineering Conference, vol. 22, no. 1, pp. 58-65, 2020.
- G. C, Lee and H. J. Choi, H. S. Chae, "Method for Generating Safety Test Cases using Event Sequence-based Risk Factors," Korea Computer Science Society Conference, pp. 690-692, 2016.
- H. S. Chae and Y. W. Lee, "Methods and Devices for Identifying Feature Interaction of Autonomous Vehicles using Patterns," Application Number: 10-2018-0167806, Application Date: 2018.12.21, Registration Number: 10-2092482, Registration Date: 2020.03.17.