Data Management Strategies for Space-Efficient Decoding and Planning

Abstract

Several key computer science tasks are traditionally solved via dynamic programming and need to work within the constraints of low-memory devices. This lecture presents two solutions that enhance space efficiency in such tasks. First, the speaker will show how to achieve space-efficient Viterbi decoding, used in speech recognition and probabilistic context-free grammar parsing. Then, he will outline how to make optimal planning decisions space-efficiently in a finite-horizon Markov Decision Process. Thereby, the speaker will showcase how data management expertise can deliver solutions in other domains. Lastly, he will glimpse into alternative time-efficient strategies for those problems.

About the Speaker

Prof. Panagiotis KARRAS is a Professor of Computer Science at the University of Copenhagen. His research interests include designing robust and versatile methods for data access, mining, analysis, and representation. He received the MSc in Electrical and Computer Engineering from the National Technical University of Athens and the PhD in Computer Science from The University of Hong Kong. He received the 2008 Young Scientist Award in Physical/Mathematical science by the Hong Kong Institution of Science, Lee Kuan Yew Postdoctoral Fellowship at the National University of Singapore, Teaching Excellence Fellowship at Rutgers Business School, and Best Faculty Performance Award at the Skolkovo Institute of Science and Technology. His work has been published in PVLDB, SIGMOD, ICDE, KDD, AAAI, IJCAI, NeurIPS, ICLR, USENIX Security, TheWebConf, SIGIR, ACL, and INTERSPEECH. 

For Attendees' Attention

Seating is on a first come, first served basis.