Generating preferred images using generative adversarial networks (GANs) is challenging owing to the high-dimensional nature of latent space. In this study, we propose a novel approach that uses simple user-swipe interactions to generate preferred images for users. To effectively explore the latent space with only swipe interactions, we apply principal component analysis to the latent space of the StyleGAN, creating subspaces composed of principal components that significantly change the appearance of the image. We use a multi-armed bandit algorithm to decide the dimensions to explore, focusing on the preferences of the user. Experiments show that our method is more efficient in generating preferred images than the baseline methods. Furthermore, changes in preferred images during image generation or the display of entirely different image styles were observed to provide new inspirations, subsequently altering user preferences.
When adding another service to an existing service, focusing on the stakeholders who have a need for the additional service may omit consideration of the impact on other stakeholders when eliciting requirements. On the other hand, stakeholders who do not have a direct need for additional services may be less interested in adding services, making it difficult to elicit their needs through interviews. In addition, service additions may cause problems due to the interaction between services, but it is generally difficult to anticipate such problems, and it is difficult to elicit problems and countermeasures through interviews. Therefore, we devised a method to identify the needs of stakeholders using the CATWOE analysis, a requirements analysis method, and to elicit requirements using the STAMP/STPA safety analysis method, which focuses on interactions by considering failure to satisfy needs as a loss. A case study is presented in which the devised method is applied to a past accident case, and the results are shown to confirm its effectiveness.
Fuzzing is a technique for detecting vulnerabilities through rapid test case generation and execution. AFL, a well-known fuzzing tool, efficiently explores test cases that pass through previously undiscovered paths by observing execution paths at the basic block level. In this study, we investigate the change in AFL path search efficiency by aggregating code clones of the target source code. We hypothesize that aggregating code clones that contain basic blocks would aggregate paths that AFL can detect, reducing the number of paths observed in AFL and making it easier to reach undiscovered paths. The experimental results showed no statistically significant difference in the number of paths discovered by AFL, but code clone aggregation did change the test cases generated by AFL, detecting one undiscovered crash.