Yannis Ioannidis (University of Athens, Greece)

User-Defined Functions in Relational Databases: Challenges and Promising Solutions based on YeSQL

Check out the Keynote recording

Download the slides of the keynote

Monday, 3 July 2023

Abstract

The diversity and complexity of modern data management applications have led to the extension of the relational paradigm with syntactic and semantic support for User-Defined Functions (UDFs). Although well-established in traditional DBMS settings, UDFs have become central in many application contexts as well, such as data science, data analytics, and edge computing. Still, a critical limitation of UDFs is the impedance mismatch between their evaluation and relational processing.

In this talk, I will first give an overview of the area and the technical challenges that UDF processing brings. I will then present YeSQL, an SQL extension with rich UDF support along with a pluggable architecture to easily integrate it with either server-based or embedded database engines. YeSQL currently supports Python UDFs fully integrated with relational queries as scalar, aggregator, or table functions. Key novel characteristics of YeSQL include easy implementation of complex algorithms and several performance enhancements, including tracing JIT compilation of Python UDFs, parallelism and fusion of UDFs, stateful UDFs, and seamless integration with a database engine. Experimental analysis showcases the usability and expressiveness of YeSQL and demonstrates that its techniques of minimizing context switching between the relational engine and the Python VM are very effective and achieve significant speedups of up to 68x in common, practical use cases compared to earlier approaches and alternative implementation choices.

Short Bio

Yannis Ioannidis (Ph.D., Computer Sciences, UC Berkeley; MSc, Applied Mathematics, Harvard University; Diploma, Electrical Engineering, National Technical University of Athens) is the President of the Association of Computing Machinery (ACM). He is a Professor at the Department of Informatics and Telecommunications of the National and Kapodistrian University of Athens as well as an Associated Faculty at the “Athena” Research and Innovation Center, where he also served as the President and General Director for 10 years. His research interests include Database and Information Systems, Data Science, Data and Text Analytics, Data Infrastructures and Digital Repositories, Recommender Systems and Personalization, and Human-Computer Interaction, topics on which he has published over 170 articles in leading journals and conferences and also holds four patents. His work is often inspired by and applied to data management and analysis problems that arise in industrial environments or in the context of other scientific fields (Social Sciences and Humanities, Life Sciences, Physical Sciences) and the Arts. He has been the coordinator and legal entity head of OpenAIRE, which implements the European policies on open access to research publications and data. He is the software director of the European Human Brain Project flagship initiative, the coordinator of the EOSC Future strategic project, which implements the core elements of the European Open Science Cloud, and a coordinator or partner in tens of other European and national research and innovation projects. He has also led or is currently leading the creation of new international or spin-off companies. He is an ACM and IEEE Fellow, a member of Academia Europaea, and a recipient of several research and teaching awards. He is currently the Greek delegate to the European Strategy Forum on Research Infrastructures (ESFRI) and a co-chair of the Global Climate Hub of the UN Sustainable Development Solutions Network.

Julia Stoyanovich (New York University, USA)

Responsible Data Management

Check out the Keynote recording

Download the slides of the keynote

Tuesday, 4 July 2023

Abstract

Incorporating ethics and legal compliance into data-driven algorithmic systems has been attracting significant attention from the computing research community, most notably under the umbrella of fair and interpretable machine learning. Yet, much of this work has been limited to the "last mile" of data analysis, disregarding both the data lifecycle, and the lifecycle of a system's design, development, and use. In my talk, I will argue that the decisions we make during data collection and preparation profoundly impact the robustness, fairness and interpretability of the systems we build, and that our responsibility for the operation of these systems does not stop once they are deployed. I will discuss technical work, and will place this work into the broader context of policy, education, and public outreach.

Short Bio

Dr. Julia Stoyanovich is Institute Associate Professor of Computer Science and Engineering, Associate Professor of Data Science, and Director of the Center for Responsible AI at New York University. Her goal is to make “Responsible AI” synonymous with “AI”. She works towards this goal by engaging in academic research, education and technology regulation, and by speaking about the benefits and harms of AI to practitioners and members of the public. Julia’s research interests include AI ethics and legal compliance, data management and AI systems, and computational social choice. She has co-authored over 100 academic publications, and has written for the New York Times, the Wall Street Journal, and Le Monde. Julia has been teaching courses on responsible data science and AI to students, practitioners and the general public. She is a co-author of the “Data, Responsibly” and “We are AI” comic book series. She received her M.S. and Ph.D. degrees in Computer Science from Columbia University, and a B.S. in Computer Science and in Mathematics & Statistics from UMass Amherst. Julia’s work has been generously supported by the US National Science Foundation, Pivotal Ventures, JP Morgan Chase, and Meta Responsible AI, among others. She is a recipient of an NSF CAREER award and a Senior Member of the ACM.

Carlo Curino (Microsoft, USA)

Title: Tensor Query Processing: How to ride the AI investment wave for database analytics?

Check out the Keynote recording

Download the slides of the keynote

Wednesday, 5 July 2023

Abstract

The successes of modern AI have fueled huge investments in highly parallel computational devices such as GPUs/APUs/TPUs and the development of corresponding tensor-centric runtimes/compilers/optimizers. In this talk, I summarize a significant research investment within MS to embrace this tensor-based world to accelerate database analytics. We have successfully developed a framework to compiled SQL to the same tensor abstraction used by popular runtimes such as PyTorch/ONNX/TVM. This allows us to run SQL on GPUs/APUs from multiple vendors and with limited effort we can port to TPUs or other custom tensor-based devices. With careful design of the tensor-relational-operators and leveraging custom kernel-fusion compilers we outperform on perf and price/perf state-of-the-art CPU systems (SQLServer/Spark/Snowflake), and match or outperform GPU-native analytics systems. Even more encouraging is the performance gains we obtain by newer HW generations, and the multitude of obvious optimizations our SW is still lacking. Interestingly, all of this can be done with a mere 20k lines of Python code. We conclude by showcasing other interesting side advantages of this approach that blend SQL+ML in interesting new ways. While the journey towards a production system is still long and treacherous, these initial few steps are very encouraging.

Short Bio

Carlo A. Curino is the lead of Gray Systems Lab (GSL), and applied research group working at the intersection of Databases/Systems/Machine Learning. Before this Carlo was a Principal Scientist in Cloud and Information Services Lab (CISL), working on large-scale distributed systems, with a focus on scheduling for BigData clusters; this line of research was co-developed with several team members and open-sourced as part of Apache Hadoop/YARN. Intrinsically, this research work enables us to operate the largest YARN clusters in the world (deployed on 250k + servers within Microsoft). Prior to joining Microsoft was a Research Scientist at Yahoo!; primarily working entity deduplication and scale and mobile+cloud platforms. Carlo spent two years as a Post Doc Associate at CSAIL MIT working with Prof. Samuel Madden and Prof. Hari Balakrishnan, working on relational databases in the cloud. Carlo received a Bachelor in Computer Science at Politecnico di Milano. He participated in a joint project between University of Illinois at Chicago (UIC) and Politecnico di Milano, obtaining a Master Degree in Computer Science at UIC and the Laurea Specialistica (cum laude) in Politecnico di Milano. During the PhD at Politecnico di Milano, Carlo spent two years as a visiting researcher at UCLA. .