Tokyo Institute of Technology School of Computing 

Dept. of Computer Science  Artificial Intelligence Course

Machine Learning and Dependable AI Lab. (tentative)

This is a new laboratory established in April 2023. We study theory and methodologies of machine learning and dependable artificial intelligence, especially AI security and privacy, explainable AI, and AI fairness. Also, we are working with external universities on AI-based pathological diagnosis of malignant lymphoma, AI-based attacks on automated driving systems and security.

The PI of this lab (Sakuma) is also the PI of the AI Security and Privacy Team, RIKEN Center for Advanced Intelligence Project (AIP).


Research

Please refer to google scholar and DBLP for publication outputs. Recent research interests include:

AI security (attacks on AI, adversarial examples, model poisoning, model inversion), AI privacy (differential privacy, multiparty computation), explainable AI, AI fairness, copyright protection of AI models and AI-generated knowledge, and language model security and privacy.

Real-world application problems working with outside research organizations includes the application of explainable AI to the pathological diagnosis of malignant lymphomas, and attacks on and defense of AI-assisted automated driving systems.

MDL slides.pdf

Major projects (as of Apr. 2023)

Understanding Attack Mechanisms against AI Considering Causal Structure in Discrimination and Building Defense Strategies (Grant-in-aid for scientific research, Fundamental Research A, PI, FY2023-2027)

 Static and dynamic reliability assurance of AI-driven hypotheses and their application in medicine (JST CREST, co-PI, FY2021-2026)

Understanding Adversarial Attack on Automated Driving Systems and Developing Countermeasure Technologies (Grant-in-aid for scientific research, Fundamental Research A, co-PI, FY2022-2024)

Differential privacy for distributed social graphs (Grant-in-aid for scientific research, Fundamental Research A, co-PI, FY2022-2026)

Members (As of Apr. 2023)

PI  Jun Sakuma (Professor, Tokyo tech and U. Tsukuba, Team leader, RIKEN AIP)

Cooperating Faculties  Yohei Akimoto (Assoc. Prof, U. Tsukuba),  Kazuto Fukuchi(Assist. Prof, U. Tsukuba)

Post Docs  YU ZHE (RIKEN AIP/Tokyo Tech.), Ryoma Senda(RIKEN AIP/Tokyo Tech.), , Joshua Butke (RIKEN AIP/Tokyo Tech.)

Doctoral Students Kazuya Kakizaki (D3, U. Tsukuba) Daiki Nishiyama (D1, Tokyo Tech.)

Master Students XU KAIWEN (M2, U. Tsukuba) Mitsuhiro Fujikawa (M2, U. Tsukuba) Rei Nagaike (M2, U. Tsukuba) Dai Shengtian (M2, U. Tsukuba) Hideyuki Oiso (M1, U. Tsukuba) Nozomu Koujiguchi (M1, U. Tsukuba) Atsuya Miyauchi (M1/, U. Tsukuba) Chikato Yamasoba (M1, U. Tsukuba) 

Undergrad Students Shojirou Yamabe (B4, Tokyo Tech.)  Yuto Yoshinari (B4, Tokyo Tech. ) Takaki Kato (B4, U. Tsukuba)

Research Students Win Lei Thwe(U. Tsukuba)


Almuni

Masahiro Fujisawa(RIKEN AIP/Tokyo Tech)  -- RIKEN

Access:

152-8550 2-12-1, Ookayama, Meguro, Tokyo, Japan

Ookayama campus, West #8 bldg.  E-905


Our lab is located on the 9th floor, and the entrance is locked.

Please call Mitsudera (三寺) san to open the gate. 

The phone number of Mitsudera san is shown at the gate.