Tensor Learning Team Seminar (Talk by Dr. Yicong He, Department of Engineering and Computer Science, University of Central Florida).

Speaker: Dr. Yicong He, Department of Engineering and Computer Science, University of Central Florida

Title: Robust, scalable and efficient methods for data analysis and recovery

Date: August 1, 2023

Time: 2:00 pm - 3:00 pm (JST)

Venue: Online

Abstract:

The robustness to outliers, scalability to large-scale data, and online processing ability are essential when evaluating machine learning and data analysis algorithms. Traditional algorithms may not perform well when dealing with data perturbed by outliers, leading to the requirements of developing robust methods. Scalability enables processing only a portion of the data at a time, making it feasible to analyze large-scale data with limited computing resources. Online processing requires algorithms to handle dynamically generated data streams in real-time. During this talk, I will begin by introducing my work on outlier robust methods for noisy data recovery, which includes robust compressive sensing reconstruction, matrix completion, and tensor completion. Next, I will present scalable and online methods for analyzing and recovering large-scale data, such as patch-tracking-based streaming tensor completion, scalable tensor ring decomposition, and scalable subspace clustering using optimal direction search. Finally, I will discuss my research on tensor space clustering.