Meta Research

• Location:
  University of Washington - Paul G. Allen School of Computer Science & Engineering
  Bill & Melinda Gates Center
  3800 E Stevens Way NE
  Seattle, WA 98195-2355
  
• Check in: When you arrive at the Bill & Melinda Gates Center, take the elevators to the fourth floor. The check in table and the event will be right outside the elevators.
• Registration: If you plan to attend the event, then please register here. Although there is no registration fee, you must register to attend. Breakfast, lunch and coffee breaks will be sponsored by Meta.
• Talk submission: If you would like to present at the meeting, please submit a title and short description/abstract along with your registration.
• Attending in-person? Please visit https://www.cs.washington.edu/visitors/getting_here for information about getting to UW.
• Previous Meetings: This is the fifth meeting of the series. The first meeting was at University of Washington in 2018, the second was at Microsoft Research in 2019, the third was at Amazon in 2020, and the fourth was held virtually by Google in 2021.
• Hashtag: Please use the event hashtag for social media posts! #NWDSMeeting

Agenda

Schedule:

• 9:00 AM - Welcome & Keynote 1 - Luna Dong, Meta
  Session chair: Magda Balazinska, University of Washington
  
• 10:00 AM - Break
  
• 10:30 AM - Short Talks (10 min talk + 5 min questions)
  Session co-chairs: Muthu Annamalai & Shilpa Lawande, Meta
  1. Martin Bravenboer, Relational AI, “Building Declarative Data Apps with RelationalAI”
  2. Alexey Leonov-Vendrovskiy and Eugene Koblov, Google, "A Brief Story of BigQuery BI Engine"
  3. Bailu Ding, Microsoft, “DSB: A Decision Support Benchmark for Workload-Driven and Traditional Database Systems”
  4. Xiao Hu, Google Research and Simon Fraser, “Learnability in ML-for-DB”
  5. Jordan Tigani, MotherDuck, “The Revenge of the Single-Node DBMS”
  6. Todd Porter, Meta, and David Maier, Portland State University, “Opportunities and Challenges for Internal Streaming Systems”
  7. Arash Termehchy, Oregon State University, “Collaborative Learning & Reasoning of Humans and Data Systems”
  8. Artur Borycki, Industry Partners, “Role of database systems in genomic research.”
     
     
• 12:30 PM - Lunch
  
• 1:30 PM - Keynote 2 - Leilani Battle, University of Washington
  Session chair: Dan Suciu, University of Washington
  
• 2:30 PM - Short Talks (10 min talk + 5 min questions)
  Session chair: Dan Suciu, University of Washington
  1. Sebastian Burckhardt and Badrish Chandramouli, Microsoft, “Netherite: Efficient Execution of Serverless Workflows”
  2. Tianzheng Wang, Simon Fraser, “Rethinking the Performance/Cost of Persistent Memory and SSDs”
  3. Jyoti Leeka, Microsoft, “Query Optimizer as a Service: An Idea Whose Time Has Come!”
  4. Moe Kayali, University of Washington, “Quasi-stable Colorings for Graph Compression”
     
     
• 3:30 PM - Posters and Coffee
  
• 5:00 PM - Event Ends
  

Keynote and Speakers

• Keynote:
  Xin Luna Dong (Meta)
  
  • Title:
    Zero to One Billion: The Path to a Rich Product Knowledge Graph
  • Abstract:
    Knowledge graphs have been used to support a wide range of applications and enhance search results for multiple major search engines, such as Google and Bing. To better serve customers in eCommerce, we aim to build a product knowledge graph with authoritative knowledge for products. The thousands of product verticals we need to model, the vast number of data sources we need to extract knowledge from, the huge volume of new products we need to handle every day, and the various applications in Search, Discovery, Personalization, Voice, that we wish to support, all present big challenges in constructing such a graph.
    In this talk we describe our efforts for knowledge collection for products of thousands of types. We describe how we nail down the most important first step for delivering the data business: training high-precision models that generate accurate data. We then describe how we scale up the models with learning from limited labels, and how we increase the yields with multi-modal models and web extraction. We share the many learnings and lessons in building this product graph and applying it to support customer-facing applications.
  • Bio:
    Xin Luna Dong is the Head Scientist at Facebook AR/VR Assistant. Prior to joining Facebook, she was a Senior Principal Scientist at Amazon, leading the efforts of constructing Amazon Product Knowledge Graph, and before that one of the major contributors to the Google Knowledge Vault project, and has led the Knowledge-based Trust project, which is called the “Google Truth Machine” by Washington’s Post. She has co-authored books "Machine Knowledge: Creation and Curation of Comprehensive Knowledge Bases" and “Big Data Integration”, was awarded ACM Distinguished Member, and VLDB Early Career Research Contribution Award for “Advancing the state of the art of knowledge fusion”. She serves in the VLDB endowment and PVLDB advisory committee, and is a PC co-chair for KDD'2022 ADS track, WSDM 2022, VLDB 2021, and Sigmod 2018.
    
    
• Keynote:
  Leilani Battle (University of Washington)
  
  • Title:
    Behavior-Driven Optimization for Interactive Data Exploration
  • Abstract:
    Analysts need the ability to intuitively explore their data before deciding how to clean it, model it, and present it to key decision makers. With the abundance of massive datasets in industry and science, analysts also need exploration systems that can process data quickly and efficiently, otherwise these systems will fail to keep pace with a user’s analytic flow. Addressing these challenges requires a deeper understanding of not only how system behavior influences user performance, but also how user behavior influences system performance.
    In this talk, I will first discuss how system performance impacts the way people visually explore large datasets, in particular how system latency encourages user exploration bias. Then I will discuss how we can counteract these effects using behavior-driven optimizations, such as by learning user exploration patterns automatically, and exploiting these patterns to pre-fetch data ahead of users as they explore to reduce system latency. Then I will discuss how I synthesize evaluation methodology from HCI, visualization, and data management into executable benchmarks for testing database management systems under real-time interactive analysis scenarios. Finally, I will discuss my ongoing research to further characterize, optimize, and evaluate interactive data exploration systems to promote more reliable, rigorous, and engaging analyses.
    
  • Bio:
    Leilani Battle is an Assistant Professor in the Paul G. Allen School of Computer Science and Engineering at the University of Washington and co-director of the UW Interactive Data Lab. She was previously an Assistant Professor at the University of Maryland, College Park. Her research spans the areas of data management, HCI, and data visualization. Her research interests focus on developing interactive data-intensive systems that can aid analysts in performing complex data exploration and analysis. Prof. Battle was named one of the 35 Innovators Under 35 by the MIT Technology Review in 2020. She is also an NSF Graduate Research Fellowship Recipient (2012), and her research is supported by an Adobe Research Award, a VMWare Early Career Faculty Grant, an NSF CISE CRII Award (2019-2021), an NSF CAREER Award (2022-2027), and an ORAU Ralph E. Powe Junior Faculty Enhancement Award (2019-2020). In 2017, she completed a postdoc in the UW Interactive Data Lab. She holds an MS (2013) and PhD (2017) in Computer Science from MIT, where she was a member of the MIT Database Group, and a BS in Computer Engineering from UW (2011) as part of the UW database group.
    

Organizers

• Magdalena Balazinska — magda@cs.washington.edu
• Shilpa Lawande — slawande@fb.com
• Muthu Annamalai — muthu@fb.com
• David Accame — daccame@fb.com

Contact

• David Accame — daccame@fb.com
• Magdalena Balazinska — magda@cs.washington.edu