Computer Science Special Seminar: Panagiotis Karras "Orientations in Data Management for Scalable Data Science"
Orientations in Data Management for Scalable Data Science
Maanantai, 23. elokuuta klo 10:30
Pyydä Zoom-linkki sähköpostitse: [email protected]
Huom! linkki lähetetään tietotekniikan laitoksen henkilökunnalle
Abstract: Data science promises to deliver the power of extracting value from data at an unprecedented scale. However, several fundamental data science tasks call for data management solutions that need to address problems of time efficiency, anytime operation, and space efficiency in order to fulfill this promise. In this talk, we will examine three completed or envisaged cases of data management solutions providing scalability, anytime characteristics, and space efficiency to data science tasks. We will first investigate the question of constructing graph node representations, or embeddings, with an anytime algorithm that also provides non-trivial error guarantees. We will then examine how a classical solution for Viterbi decoding by a dynamic programming algorithm can be rendered more space efficient via an application of sophisticated data management techniques. Last, we will outline how the optimal actions in a finite-horizon Markov Decision Process can be computed in a space-efficient manner as well. We will conclude with a vision on how data management expertise can facilitate and advance the frontiers of data science.
Bio: Panagiotis Karras is an Associate Professor of Computer Science at Aarhus University. In his research he designs robust and versatile methods for data access, mining, analysis, and representation. He received an MSc in Electrical and Computer Engineering from the National Technical University of Athens and a PhD in Computer Science from the University of Hong Kong. He has been awarded a Hong Kong Young Scientist Award, a Singapore Lee Kuan Yew Postdoctoral Fellowship, a Rutgers Business School Teaching Excellence Fellowship, and a Skoltech Best Faculty Performance Award. His work has been published in PVLDB, SIGMOD, ICDE, KDD, AAAI, NeurIPS, TheWebConf, and SIGIR.