About
Our mission is to explore and establish a methodology for evolutionary intelligence that fosters highly adaptive, autonomous, and theoretically reliable learning and optimization abilities. Our current research topics include the followings:
-
Expensive optimizationMany real-world applications require optimizing expensive-to-evaluate objectives, as objective values are evaluated often with computationally expensive simulations or costly experiments. An example is the optimization of aircraft wing using CFD simulation, which is computationally expensive. To solve such expensive optimization problems, we explore sample-efficient optimization techniques. Especially, high-dimensional multi-objective problems are our main focus to be studied.
-
Evolutionary machine learningIntegrating meta-heuristics into machine learning approaches bring highly adaptive and autonomous ability, such as auto-design of learning models, auto-tuning of hyper-parameters, and extraction of explainable models. We explore evolutionary machine leaning techniques; evolutionary rule-based learning, evolutionary symbolic regression, evolutionary neural architecture search, and etc. We are interested in realizing the interplay between evolution and learning on a computer intelligence scheme.
Highlights
15 Apr. 2025
A/Prof. Nakata has received the Young Scientists' Award, the 2025 Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology.
14 Mar. 2025
We have proposed an evolutionary rule-based learning algorithm that automatically adapts appropriate rule representations to each subspace of the input space. This novel approach enables flexible modeling of both crisp and fuzzy decision boundaries across different regions of data, effectively addressing the challenge of representation selection in complex problems. This contribution was published on IEEE Transactions on Evolutionary Computation.
Paper
Code
16 Feb. 2025
In evolutionary fuzzy rule-based classification systems, we have proposed a novel class inference scheme to theoretically quantify model uncertainty (termed the "I don’t know" state) using the Dempster-Shafer Theory of Evidence. Specifically, for binary classification tasks, this method enables models to assign confidence scores not only to "Class 1" and "Class 2" but also to a dedicated "I don’t know" category, providing a mathematically grounded measure of indecision. This contribution has been published in ACM Transactions on Evolutionary Learning and Optimization.
Paper
Code
Contact
Our laboratory is on the eighth floor of N6-2, Hodogaya-campus, Yokohama National University. See here for more detail.
Affiliation
Faculty of Engineering, Yokohama National University
Address
Room 801/812, N6-2, Tokiwadai 79-5, Yokohama, Japan, 240-8501.
E-mail
nakata-masaya-tb at ynu.ac.jp（to Masaya Nakata)