Deep Learning from Sequential Data

Апстракт предавања

In this talk we will discuss the state of the art approaches for descriptive and predictive analysis of sequential data, such as text and event logs. A critical challenge in the analysis of sequential data is data representation, which refers to converting the raw data into a form that is suitable for machine learning algorithms. Many machine learning algorithms, such as neural networks, require the input to be provided as a fixed-length vector and, for a long time, this has been considered a major obstacle for successful learning from sequential data. The recent progress in machine learning has resulted in several powerful ideas for better representation and learning from sequential data. Among those ideas, probably the most powerful are distributed representations and deep learning. We will describe the intuition behind these ideas and demonstrate their promise by showing our recent results on the analysis of micro-blogging data and medical records data.

О предавачу

Slobodan Vucetic is an Associate Professor and Chair of the Department for Computer and Information Sciences at Temple University. He got his Ph.D. degree in Electrical Engineering from Washington State University in 2001, and his B.S. and M.S. degrees in Electrical Engineering are from the University of Novi Sad. His research expertise and interests are in machine learning, data science, and big data. His research focuses on solving real-life knowledge discovery problems through development of novel machine learning algorithms and is driven by open problems in a wide array of disciplines such as Public Health, Medicine, Biology, Geosciences, Education, Marketing, Social Sciences, Traffic Engineering, and Industrial Engineering. Dr. Vucetic has published over 100 research papers that have been cited over 5,000 times and his current research is funded by the U.S. National Science Foundation (NSF), the National Institutes or Health, and industry. He is a recipient of the NSF CAREER award.