Welcome to Slavakis Lab
We are a research team (est. April 2021) at the Institute of Science Tokyo (formerly known as Tokyo Institute of Technology), Department of Information and Communications Engineering, that designs computational methods for signal processing and machine learning.
Our designs aim at a wide variety of themes and applications. Examples are:
- reinforcement learning;
- nonparametric regression;
- manifold learning;
- biomedical imaging;
- network time-series analysis;
- online learning and adaptive filtering;
- optimization with a focus on nonexpansive and monotone mappings; and
- quantum signal processing and machine learning.
Software developed by our team can be found here.
信号処理と機械学習の研究を行なっている。信号やデータの背景に潜む幾何学的構造を抽出し、これを最大限活 用することによって、これまでの信号処理と機械学習を凌駕する柔軟なフレームワークの構築を目指している。 広い領域への応用を視野に入れているが、現在、特に強化学習、医用イメージング、脳ネットワーク、確率近似、 最適化および逆問題への応用に注力し、研究を進めている。
We are a small team, but we wish to grow! We are looking for passionate new PhD/MSc/BSc students as well as PostDocs to join the team (more info) !
Call for papers: Special issue of Signal Processing: Signal Processing and Learning with Manifolds and Lie Groups
News
January 2026:
Prof. Suzuki’s work on nonconvex regularization for sparse reinforcement learning, and Thien’s work on manifold-constrained kernel tensor decompositions have been accepted for presentation at the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Barcelona: Spain, May 4-8, 2026.
December 2025:
Shiwen and Jiayi’s work—Jiayi was a visiting student in our Lab—on a high-performance Rust platform for quantum computations appears as a preprint in arXiv.
December 2025:
Minh’s work on Gaussian mixtures for modeling Q-functions in reinforcement learning appears as a preprint in arXiv.
October 2025:
Call for papers: Special issue of Signal Processing: Signal Processing and Learning with Manifolds and Lie Groups
September 2025:
Xinru and Pan have joined the Lab as MSc students. Welcome aboard!
September 2025:
Thien’s work on tensor trains and multi-way data [TechRxiv], Kyohei’s work on sparse reinforcement learning [arXiv], and Minh’s results on online reinforcement learning via Gaussian mixture models [arXiv] have been all posted online as preprints.
August 2025:
Thien’s conference paper, entitled “Estimating dynamic graph flows with kernel models and Hadamard-structured Riemannian constraints,” will be presented at the 17th Asia Pacific Signal and Information Processing Association (APSIPA) Annual Summit and Conference, Shangri-la, Singapore, 22-24 October, 2025.
June 2025:
Minh’s conference paper, entitled “Riemannian Q-functions for policy iteration in reinforcement learning,” has been nominated as one of the finalists for the European Signal Processing Conference (EUSIPCO), 8-12 September, 2025, best student paper award. Congrats!
May 2025:
Kotaro’s work on robust invariant representation learning appears as a preprint in [arXiv].