研讨会, 演讲, 讲座 | 2021年12月09日
IAS Center for Quantum Technologies Seminar Series - Benchmarking Analog Quantum Simulators Based on Emergent Randomness
Abstract One of the most basic tasks to characterize analog quantum simulators is to estimate the many-body fidelity between an ideal target state and the state obtained from experiments. In this talk, the speaker will present a simple and efficient benchmarking method that requires minimal experimental control. It only requires time-evolving a quantum system under its natural ergodic dynamics, followed by measurements in a fixed local basis. At its core, the method is made possible by using a newly-discovered phenomenon that occurs in strongly interacting quantum many-body systems, namely the emergence of universal random statistics. The speaker claims the phenomenon occurs universally in a wide class of ergodic quantum systems at infinite temperature by presenting a number of evidence based on solvable models and numerical simulations. He will discuss his benchmarking protocol based on the emergent randomness, and demonstrate it both numerically for model systems and experimentally using a Rydberg quantum simulator. About the speaker Prof. Choi Soonwon obtained his BS in Physics from the California Institute of Technology in 2012 and his PhD in Physics from the Harvard University in 2018. He then worked as a Miller Postdoctoral Fellow at the University of California, Berkeley before joining MIT as an Assistant Professor in July 2021. Prof. Choi’s research focuses on exploring dynamical phenomena that occur in strongly interacting quantum many-body systems far from equilibrium and designing their novel applications for quantum information science.
研讨会, 演讲, 讲座 | 2021年12月09日
Department of Mathematics - Seminar on FinTech and Machine Learning - Problems and probable solutions of applying recent machine learning techniques to financial time series and data
Recent machine learning techniques such as deep learning and reinforcement learning were built on specific assumptions of the underlying data generation process. Financial time series frequently do not satisfy these assumptions. In this talk, we discuss the possible problems if these techniques are applied blindly. The solutions to these problems are in general problem specific. However, some of the pain can be alleviated by combining recent machine learning techniques with more classical statistical and econometrics insights. We will discuss these probable solutions with examples.
简介会, 迎新 | 2021年12月04日
HKUST MSc in Data-Driven Modeling Information Session 2021 (Wuhan)
Application for 2022/23 Fall Term admission to HKUST MSc in Data-Driven Modeling is now open. When and How to apply? Join our Information Session in Wuhan / via Zoom on December 4, 2021 (Saturday) to meet our faculty, graduates and students to learn more about: the benefits of studying Master of Science (MSc) in Data-Driven Modeling at HKUST; and why choose HKUST MSc in Data-Driven Modeling to pursue your MSc study; and information about the program and admission details. Register now: Registration Link - MSc DDM Information Session (Wuhan)