Invited Keynotes

We are very pleased to announce the EDBT/ICDT 2020 invited keynote speakers.

Keynote 1: Facets of Probabilistic Databases
by Benny Kimelfeld

Abstract: Probabilistic databases are commonly known in the form of the tuple-independent model, where the validity of every tuple is an independent random event. Conceptually, the notion is more general, as a probabilistic database refers to any probability distribution over ordinary databases. A central computational problem is that of marginal inference for database queries: what is the probability that a given tuple is a query answer? In this talk, I will discuss recent developments in several research directions that, collectively, position probabilistic databases as the common and natural foundation of various challenges at the core of data analytics. Examples include reasoning about uncertain preferences from conventional distributions such as the Mallows model, data cleaning and repairing in probabilistic paradigms such as the HoloClean system, and the explanation of query answers through concepts from cooperative game theory such as the Shapley value and the Banzhaf Power Index. While these challenges manifest different facets of probabilistic databases, I will show how they interrelate and, moreover, how they relate to the basic theory of inference over tuple-independent databases.

Benny Kimelfeld is an Associate Professor at the Computer Science Faculty at Technion, Israel, and the head and founder of the Technion Data and Knowledge (TD&K) Laboratory. After receiving his Ph.D. from The Hebrew University of Jerusalem and before joining Technion, he has been a Research Staff Member at IBM Research Almaden and a Computer Scientist at LogicBlox. Benny's research spans a spectrum of both foundational and systems aspects of data management, such as algorithms for query evaluation, probabilistic and inconsistent databases, and infrastructure for text analytics. His research has been funded by various agencies and programs, including the German-Israeli Project Cooperation (DIP). He currently serves as an associate editor for the Journal of Computer and System Sciences (JCSS), and he has served as a co-chair for WebDB 2016 and BUDA 2014, a group leader for SIGMOD 2017, and the program committee chair for ICDT 2018.

Keynote 2: Fair Decision-Making using Privacy Protected Data
by Ashwin Machanavajjhala

Abstract: We are in an “age of surveillance”. Mass surveillance by governments and corporations has become the norm. While the goal of such data collection is often to learn aggregate trends in populations rather than properties of individuals, current techniques for data sharing and analysis do not sufficiently protect individual privacy. Differential privacy is a principled approach that, in theory, is a perfect solution for addressing these privacy problems and is starting to see adoption by some commercial (e.g., Google and Apple) and government entities (e.g., US Census) for collecting and sharing sensitive user data. However, there are still barriers to wide adoption of differential privacy notably when applied to complex relational data types. In this talk, I will highlight the opportunities and challenges raised by differentially private data analysis techniques with a specific focus on differentially private techniques for answering queries from relational databases. Specifically, I will discuss our recent work on (a) how to specify what should be private, (b) how to optimally add noise to query answers to ensure differential privacy while minimizing error, and (c) how to use noisy answers for data driven decision making.

Ashwin Machanavajjhala is an Associate Professor in the Department of Computer Science, Duke University, and co-founder of Tumult Labs. Previously, he was a Senior Research Scientist in the Knowledge Management group at Yahoo! Research. His primary research interests lie in algorithms for privacy preserving data analytics with a focus on differential privacy. He is a recipient of a 2017 IEEE Influential paper award for the invention of L-diversity in 2006, the National Science Foundation Faculty Early CAREER award in 2013, and the 2008 ACM SIGMOD Jim Gray Dissertation Award Honorable Mention. In collaboration with the US Census Bureau, he is credited with developing the first real world deployment of differential privacy. Ashwin graduated with a Ph.D. from the Department of Computer Science, Cornell University and a B.Tech in Computer Science and Engineering from the Indian Institute of Technology, Madras.

Keynote 3: Current Challenges in Graph Databases
by Juan Reutter

Abstract: As graph databases grow in popularity, decades of work in graph query languages and models are materializing in industry standards and in the construction of new graph database systems. However, this surge in graph systems has in turn opened up a series of new, interesting research problems related to graph databases. To name a few of them: graph users need support for graph schemas and analytical queries. Graph engines must be able to deal with both patterns and path queries, which also raises questions on how paths are to be represented and computed by graph engines. Moreover, graph queries often involve complex patterns that are difficult to evaluate using standard relational techniques, and graph analytics tasks put an even higher pressure on graph database engines. In this talk I will discuss current challenges in graph database research, from the design of languages to algorithmic issues raised by the development of graph engines.

Juan Reutter works as an associate professor in the department of computer science of Pontificia Universidad Católica de Chile, and as an associate researcher in the Millenium Institute for Foundational Research on Data. He received his PhD in the University of Edinburgh on 2013. His research interest has to do with mathematical foundations of data management, web data and data science. He received the Cor Baayen Young Researcher Award from the European Research Consortium for Informatics and Mathematics and a BCS distinguished dissertation award for his phD thesis. His research has also been highlighted with a Best Paper Award at PODS and two best paper award nominations at ISWC.

Keynote 4: From Data to Models and Back: Experiences from Google's Production ML Pipelines
by Neoklis (Alkis) Polyzotis

Abstract: Building a good ML model requires good input data. Conversely, debugging a model inevitably involves data debugging and understanding. In this talk, I will present our work in supporting key steps in this bidirectional journey, motivated by the needs of large production ML pipelines at Google: connecting structured data to ML, validating the data fed to training and inference, analyzing model performance through the lens of input data, and linking everything together through lineage. Our tools are deployed in ML pipelines at Google and other large organizations (through open-source software), processing several petabytes of data per day and powering models with millions of QPS.

Neoklis (Alkis) Polyzotis is a researcher at Google Research, where he currently leads the data-management projects in Google’s TensorFlow Extended (TFX) platform for production-grade machine learning. His interests include data management for machine learning, enterprise data search, and interactive data exploration. Before joining Google, he was a professor at UC Santa Cruz. He has received a PhD in Computer Sciences from the University of Wisconsin at Madison and a diploma in engineering from the National Tech. University of Athens, Greece.