Keynote Speaker
Prof. Masaaki Fukasawa
The University of Osaka, Japan
Speech Title: Martingale expansion for stochastic volatility
Abstract: The martingale
expansion provides a refined approximation to the
marginal distributions of martingales beyond the normal
approximation implied by the martingale central limit
theorem. We develop a martingale expansion framework
specifically suited to continuous stochastic volatility
models. Our approach accommodates both small
volatility-of-volatility and fast mean-reversion models,
yielding first-order perturbation expansions under
essentially minimal conditions.
Biodata: Masaaki Fukasawa is Professor at The University of Osaka, charing Stochastic Analysis Group at Graduate School of Engineering Science, since 2016. He has been Co-Editor of Finance and Stochastics since 2018. He has been working on Mathematical Finance and related issues, in particular, Asymptotic Analysis of financial stochastic models.
Invited Speaker
Prof. Hsi-Sheng Goan
National Taiwan University, Taiwan
Speech Title: Quantum Machine Learning
Abstract: Quantum computing and machine learning (the core of contemporary artificial intelligence) are emerging and promising technologies that would greatly impact human life and society in the future. It is interesting to explore the interaction between quantum computing and machine learning, and study how to use the results and technologies of one field to solve problems in another field. In this talk, I will first give a brief introduction to machine learning and quantum computing.
We are currently in the so-called noisy intermediate scale quantum (NISQ) era. Due to the lack of quantum error correction, NISQ machines are unsuitable for deep quantum circuit architectures. The variational quantum algorithm with tunable parameters in quantum circuits, which are optimized iteratively by a classical computer, is a hybrid quantum-classical approach suitable for NISQ machines. It leverages the strengths of quantum and classical computation. I will present the results obtained for some machine learning tasks using variational quantum algorithms.
After that, I will introduce the quantum-train (QT) framework, a novel approach that integrates quantum computing with classical machine learning algorithms to tackle significant challenges in data encoding, model compression, and inference hardware requirements. Despite a slight drop in accuracy, QT achieves impressive results by utilizing a quantum neural network alongside a classical mapping model, significantly reducing the parameter count from M to O(polylog(M)) during training. Our experiments demonstrate QT’s effectiveness in classification tasks, providing insights into its potential to revolutionize machine learning by exploiting quantum computational advantages. This approach not only enhances model efficiency but also reduces generalization errors, highlighting QT’s promise across various machine learning applications.
Biodata: Hsi-Sheng Goan received the Ph.D. degree in physics from the University of Maryland, College Park, USA, in 1999. He then worked as a Postdoctoral Research Fellow at the University of Queensland, Brisbane, Australia, from 1999 to 2001. From 2002 to 2004, he was a Senior Research Fellow awarded the Hewlett-Packard Fellowship with the Center for Quantum Computer Technology, University of New South Wales, Sydney, Australia, before he took up a faculty position at the Department of Physics, National Taiwan University (NTU), in 2005. He is currently a Professor of physics with NTU working in the fields of Quantum Computing and Quantum Information, Quantum Control, Mesoscopic (Nano) Physics, Quantum Optics, and Quantum Optomechanical and Electromechanical Systems. He has served as a member of the Editorial Boards of several international scientific journals, such as the International Journal of Quantum Information, European Physical Journal: Quantum Technology, the Chinese Journal of Physics, and Frontiers in ICT: Quantum Computing.(Based on document published on 20 July 2020).