AI4S Seminar Series "Structured representation learning: tensor network principles toward scalable and reliable AI" Talk by Qibin Zhao (RIKEN AIP)

Part of RIKEN AIP's AI for Science (AI4S) Seminar Series, this talk by Qibin Zhao (Team Director, RIKEN AIP) covers structured representation learning through tensor network principles. Tensor Networks (TNs) factorize high-dimensional tensors into networks of low-dimensional tensors, an approach with roots in quantum physics, high-performance computing, and applied mathematics. The session presents recent progress in applying TN technology to machine learning, focusing on the basic principles and algorithms. The emphasis is on efficiency, robustness, and scalability issues in deep learning models, with the goal of building toward more scalable and reliable AI. The seminar runs in a hybrid format from 4:00pm to 5:30pm JST. Note that in-person attendance is limited to AIP researchers, so external participants join online.