Trustworthy & Smart Software Engineering Lab. 日本語 >

dummy

Web Site of Fuyuki Ishikawa & Ishikawa Lab.

Professor at Information Systems Architecture Science Research Division, National Institute of Informatics, Tokyo, Japan

Professor at Sokendai

Visiting Associate Professor at The University of Electro-Communications

Trustworthy & Smart Software Engineering

We have a wide range of activities for "Smart Systems and Smart Dependability Assurance."

We envision advanced application systems and investigate techqnieus of verification, reasoning, optimization, automated test genration, and self-adaptation by making use of a variety of models for requirements, specifications, and designs.

Our present focus is dependability in Cyber-Physical Systems and Machine Learning Systems via techniques of formal methods and automated test generation.

Our group consists of members from different organizations and also promote international and industry-academia collaborations.

Key Activities

You may check DBLP or Google Scholar

Engineering methods for machine learning-based systems, specifically automated testing and repair for safety assurance of autonomous driving systems
JST MIRAI-eAI Project PI (completed)
Key papers:
  • Lyu et al., SpectAcle: Fault Localisation of AI-Enabled CPS by Exploiting Sequences of DNN Controller Inferences, ACM TOSEM
  • Schneider et al., Filter-based Repair of Semantic Segmentation in Safety-Critical Systems, SANER'25
  • Li Calsi et al., Distributed Repair of Deep Neural Networks, ICST'23
  • Tokui et al., NeuRecover: Regression-Controlled Repair of Deep Neural Networks with Training History, SANER'22
Testing, fault localization, automate repair for autonomous driving systems
JST ERATO-MMSD Project Co-PI (completed)
Key papers:
  • Zhang et al., An Incremental Approach for Understanding Collision Avoidance of an Industrial Path Planner, IEEE TDSC
  • Luo et al., Targeting Requirements Violations of Autonomous Driving Systems by Dynamic Evolutionary Search, ASE'21
  • Calò et al., Generating Avoidable Collision Scenarios for Testing Autonomous Driving Systems, ICST'20
Complexity mitigation in applying refinement-based formal methods (Event-B)
REFENG Project PI (completed)
Key papers:
  • Kobayashi et al., Formal Modelling of Safety Architecture for Responsibility-Aware Autonomous Vehicle via Event-B Refinement, FM'23
  • Kobayashi et al., Consistency-Preserving Refactoring of Refinement Structures in Event-B Models, Formal Aspects of Computing

Full lists are found in the list of publications and activities.

Contact

f-ishikawa < a > nii.ac.jp