Tensor Networks (TNs) are factorizations of high dimensional tensors into networks of many low-dimensional tensors, which have been studied in quantum physics, high-performance computing, and applied mathematics. In recent years, TNs have been increasingly investigated and applied to machine learning and signal processing, due to its significant advances in handling large-scale and high-dimensional problems, model compression in deep neural networks, and efficient computations for learning algorithms. This talk aims to present a broad overview of recent progress of TNs technology applied to machine learning from perspectives of basic principle and algorithms, novel approaches in unsupervised learning, tensor completion, multi-task, multi-model learning and various applications in DNN, CNN, RNN and etc. We also discuss the future research directions and new trend in this area.

Yang-Mills theory was studied from mathematical perspectives in the 1970s by Atiyah and his collaborators (notably Drinfeld, Hitchin, Singer). Subsequent breakthroughs were made on dimensions 3 and 4 by Floer and Donaldson (based on deep analytic results obtained by Uhlenbeck and Taubes) in the 1980s. In 1996, Donaldson and Thomas proposed to study Yang-Mills theories on dimensions bigger than 4. In higher dimensions, the analytic method is limited and algebro-geometric method is heavily used instead. This powerful tool usually enables us to compute partition functions and lead to amazing links to other invariants in enumerative geometry, e.g. Gromov-Witten and Gopakumar-Vafa invariants. In this talk, I will review some of these inspiring stories and discuss how my works on Calabi-Yau 4-folds fit into them.

The Cosmic Microwave Background (CMB) gives a photographic image of the Universe when it was still an “infant”. We have been using it to test our ideas about the origin of the Universe. The CMB research told us a remarkable story: the structure we see in our Universe such as galaxies, stars, planets, and eventually ourselves originated from tiny quantum fluctuations in the period of the early Universe called cosmic inflation. While we have accumulated strong evidence for this picture, the extraordinary claim requires extraordinary evidence. The last prediction of inflation that is yet to be confirmed is the existence of primordial gravitational waves whose wavelength can be as big as billions of light years. To this end we have proposed to JAXA a new satellite mission called LiteBIRD, whose primary scientific goal is to find signatures of gravitational waves in the polarisation of the CMB. In this presentation we describe physics of gravitational waves from inflation, and the LiteBIRD proposal.