Quantum machine learning: from near-term to fault-tolerance
Dr. Junyu Liu
Pritzker School of Molecular Engineering Kadanoff Center for Theoretical Physics The University of Chicago
Abstract: Quantum technologies, such as quantum computing, are poised to revolutionize next-generation digital technologies by leveraging the principles of quantum mechanics, and are widely regarded as some of the most significant technologies of our era. Quantum machine learning, which involves running machine learning algorithms on quantum devices, is seen as a flagship application in this field. In my talk, I will explore two aspects of quantum machine learning: near-term algorithms and fault-tolerant algorithms. For near-term applications, I will delve into the use of variational quantum circuits in machine learning problems and discuss the quantum neural tangent kernel theory as an analytical tool for understanding and optimizing quantum neural networks. Regarding fault-tolerant applications that incorporate quantum error correction, I will present an end-to-end application of the HHL algorithm. This algorithm offers a provable, generic, and efficient approach to a range of machine learning challenges.
Monday, Wilder Auditorium 3:50-4:30 with tea beginning at 3:25 in the physics library
ZOOM LINK FOR REMOTE VIEWING:
https://miami.zoom.us/j/91977677717?pwd=RFQzZHBIUk9HaU1YMEd0cmxQeGRDdz09&from=addon
"Enhancing Resilience and Efficiency of Ensemble Learning Systems"
Yanzhao Wu, Ph.D.
Ungar Building, Room 305 (3rd Floor), Coral Gables Campus
Wednesday 11/15/2023 at 5PM | Zoom: https://miami.zoom.us/j/97116570518
Refreshments at 4PM (Ungar 3rd Floor)
Abstract
Deep neural network ensembles hold the potential to improve generalization performance and robustness for complex learning tasks. However, it remains an open challenge to harness model learning heterogeneity among member models to improve ensemble performance. This talk will present a two-tier heterogeneity driven ensemble framework powered by focal diversity to enhance the resilience and efficiency of ensemble learning systems. First, I will show that heterogeneous DNN models trained for solving the same learning problem, e.g., image classification or object detection, with high focal diversity, can significantly improve the generalization performance and ensemble efficiency. Second, the ensemble robustness can be further strengthened by composing ensembles of heterogeneous models for solving different learning problems, e.g., object detection and semantic segmentation, through the proposed connected component labeling (CCL) based alignment. Third, extensive experiments show that this focal diversity based two-tier heterogeneity driven ensemble framework can effectively leverage model learning heterogeneity to consistently boost ensemble robustness and efficiency. I will also present an overview of our recent studies in Edge AI and Large Language Models (LLMs) and future research in machine learning systems.
Bio
Dr. Yanzhao Wu is an Assistant Professor in the Knight Foundation School of Computing and Information Sciences at Florida International University. He obtained his bachelor’s degree from University of Science and Technology of China in 2017 and then received his Ph.D. degree in Computer Science from Georgia Institute of Technology in 2022. His research interests are primarily centered on the intersection of machine learning and computing systems, including machine learning algorithm and system optimizations, deep learning, large language models, edge AI, big data analytics, and their real-world applications. His work has been published in top venues, including CVPR, ICSE, IEEE ICDCS, IEEE ICDM, IEEE TPDS, IEEE TSC, and ACM TOIS, and won the IEEE CIC 2021 Best Paper Award. He also serves as a committee member/reviewer for top conferences and journals, such as ICDE, WWW, CVPR, ECCV, AAAI, IEEE TIFS, IEEE TKDE, and ACM TOIT.
“Designing functional DNA with Large Language Models”
Avantika Lal, Ph.D.
Ungar Building, Room 305 (3rd Floor), Coral Gables Campus
Wednesday 11/01/2023 at 5PM | Zoom: https://miami.zoom.us/j/97116570518
Refreshments at 4PM (Ungar 3rd Floor)
Abstract
Designing sequences of DNA that can perform desired functions inside living cells is a challenging task with many applications, including cell therapy, gene therapy and biomanufacturing. In this talk, Dr. Lal will show how autoregressive large language models with innovative architectures can be trained to learn the regulatory code of DNA. Using DNA language models coupled with supervised sequence-to-function models, Dr. Lal will show how to design realistic and diverse DNA sequences which perform specific functions in human and yeast cells. In the future, extensions of this approach could potentially be used to design entire genetic circuits and even genomes.
“Using big data to understand how we can improve psychological well-being”
Aaron Heller, Ph.D.
Ungar Building, Room UB305 (3rd Floor), Coral Gables Campus
Wednesday 9/13/2023 at 4:30PM | Zoom: https://miami.zoom.us/j/97116570518
Refreshments at 3:30PM
Abstract
Modern technology allows researchers to follow people as they live their daily lives. Using a combination of GPS tracking, experience sampling of emotion, and brain imaging, my lab at the University of Miami has begun to test questions like, ‘what daily behaviors are most linked to increases in well-being and reductions in depression’. This work uses ‘big data’— that is, lots of data collected from lots of people over long periods of time. In this talk I will present how we collect and analyze these data, our results, and what it could mean for society. I will also talk about how previous Computer Science students have contributed to our work by applying their classroom knowledge to real-world data.
