RIKEN AIP and NCU Joint Workshop on Mathematical Foundations of Machine Learning

Date and Time: September 26th 2:50pm - 9:00 pm (JST)
Venue: Online
Language: English

Scope:
The astounding performance and applicability of machine learning in an ever-increasing number of applications have not been paired with a sufficient understanding of its mathematical foundations, preventing it from application to critical fields where human safety and health are of interest. This workshop aims to present and discuss results laying firm mathematical foundations of design, implementation, and evaluation of machine learning algorithms using tools of linear algebra, functional analysis, probability theory, statistics, optimization, and information theory.

Agenda:

14:50~15:00 JST 7:50~8:00 CEST
Welcome Remarks
Wojciech WYSOTA Vice-Rector for Research, Nicolaus Copernicus University, Toruń, Poland

Session 1

15:00~15:20 JST 8:00~8:20 CEST
Classifier Learning Using Multi-objective Optimization - Threats to Validity
Michał WOŹNIAK Wrocław University of Science and Technology, Wrocław, Poland
15:20~15:40 JST 8:20~8:40 CEST Convergence of Fixed Point Iterations of Positive Concave Mappings with Applications to Design of Machine Learning Algorithms
Renato L. G. CAVALCANTE HHI Fraunhofer Institute, Berlin, Germany
15:40~16:00 JST 8:40~9:00 CEST
Fixed Points of Nonnegative Neural Networks
omasz PIOTROWSKI Nicolaus Copernicus University, Toruń, Poland
16:00~16:20 JST 9:00~9:20 CEST Break Time

Session 2

16:20~16:40 JST 9:20~9:40 CEST
Fisher-Rao Metric and Infinite-dimensional Divergences for Gaussian Processes
Minh HA QUANG RIKEN AIP, Tokyo, Japan
16:40~17:00 JST 9:40~10:00 CEST
Efficient Machine Learning with Tensor Networks
Qibin ZHAO RIKEN AIP, Tokyo, Japan
17:00~17:20 JST 10:00~10:20 CEST
Tensor Networks for CNN Compression
Rafał ZDUNEK Wrocław University of Science and Technology, Wrocław, Poland
17:20~17:40 JST 10:20~10:40 CEST
Discovering Optimal Tensor Network Architectures: An Exploration of Tensor Network Structure Search (TN-SS)
Chao LI RIKEN AIP, Tokyo, Japan
17:40~18:00 JST 10:40~11:00 CEST Break Time

Session 3

18:00~18:20 JST 11:00~11:20 CEST
Explainable AI: Opportunities and Challenges
Przemysław BIECEK Warsaw University of Technology, Warsaw, Poland
18:20~18:40 JST 11:20~11:40 CEST
Zero-waste Machine Learning
Tomasz TRZCIŃSKI IDEAS NCBR, Warsaw, Poland and Warsaw University of Technology, Warsaw, Poland
18:40~19:00 JST 11:40~12:00 CEST
Towards Continuous Adaptation in Non-stationary Environments
Zhen-Yu ZHANG RIKEN AIP, Tokyo, Japan
19:00~19:20 JST 12:00~12:20 CEST
Exploring the Fixed Points in Cone Mapping: Enhancements to Neural Network Applications
Krzysztof RYKACZEWSKI Nicolaus Copernicus University, Toruń, Poland
19:20~20:20 JST 12:20~13:20 CEST Break Time

Session 4

20:20~20:40 JST 13:20~13:40 CEST
Recent Advances in Reliable Machine Learning
Masashi SUGIYAMA RIKEN AIP and The University of Tokyo, Tokyo, Japan
20:40~21:00 JST 13:40~14:00 CEST
Stable Gradient-based Hyperparameter Optimization
Ryuichiro HATAYA RIKEN AIP, Tokyo, Japan

For more information, please see the following site.
Workshop on Mathematical Foundations of Machine Learning

All participants are required to agree with the AIP Seminar Series Code of Conduct.
Please see the URL below.
https://aip.riken.jp/event-list/termsofparticipation/?lang=en

RIKEN AIP will expect adherence to this code throughout the event. We expect cooperation from all participants to help ensure a safe environment for everybody.