| 日時(Datetime) | 2026年6月19日(金)3限 (13:30--15:00), 2026/06/19, Friday, 3rd slot |
|---|---|
| 場所(Location) | パナソニック デジタル大講義室 (L1), Panasonic Digital Lecture Hall (L1) |
| 司会(Chair) | Christopher Wiraatmaja |
| 講演者(Presenter) | Ji Zhenlan, ソフトウェア設計学研究室 (Software Design and Analysis) |
|---|---|
| 題目(Title) | Causality for Software Engineering: Principled Techniques for Analyzing and Testing Modern Software Systems |
| 概要(Abstract) | Modern software systems are growing increasingly complex, with deep learning and large language models in software systems further exemplifying the challenges of opacity, non-determinism, and intricate dependencies that resist conventional correlation-based analysis. This talk advocates causality as a principled foundation for software engineering research on such systems. After introducing the basics of causality, two distinct causality application paradigms in software engineering will be illustrated through research on deep learning: (1) designing causality-inspired testing criteria, and (2) leveraging causal analysis to understand and explain system behaviors. |
| 講演者(Presenter) | Taiki Kitazawa, 情報セキュリティ工学研究室 (information security engineering) |
|---|---|
| 題目(Title) | Electromagnetic Information Leakage in Non-Cryptographic and Cryptographic Devices: Passive and Active Approaches |
| 概要(Abstract) | This talk provides an overview of electromagnetic information leakage in electronic devices, focusing on both non-cryptographic and cryptographic devices. Specifically, it discusses passive TEMPEST attacks, which observe unintentionally emitted electromagnetic signals from devices such as displays, and active EM side-channel attacks, which induce leakage by irradiating devices with external electromagnetic waves. Through a comparison of these approaches, the talk examines the physical mechanisms by which information propagates outside devices through electromagnetic waves, as well as the challenges in designing and evaluating effective countermeasures. |