The 2019 SIGMOD conference is coming up in a week, and the WuLab is both organizing events at the conference and presenting some of our current work. This post highlights what we’ll be up to. Come see our work and find us in the hallways!
All of this work is thanks to the excellent students in the WuLab and our collaborators around the world.
Tutorials and Events
Sunday June 30 Eugene Wu is presenting the tutorial Towards Democratizing Relational Data Visualization. Visualization is one of the most powerful ways that humans can interact with, and make use of, data. The tutorial will survey the exciting connections between relational data visualization and data management, and the research frontiers that lie ahead. I will share my perspective on performance and efficiency in this area. Presented with Nan Tang and Guoliang Li.
Monday July 1–2 Eugene will co-chair the SIGMOD Student Research Competition (SRC). This competition offers undergraduate and graduate students a unique forum to experience the research world, present their research results to conference attendees, and compete for prizes. We are excited to promote research from junior researchers around the world — come see the future researchers in our community! Organized with Jana Giceva.
Tuesday 2:30–4:30 in SIGMOD Research 3 (room: Administratiezaal) Pei Wang will present Progressive Deep Web Crawling Through Keyword Queries For Data Enrichment. Data scientists and data journalists spend a lot of time scraping the web (say, by querying a search-based API such as Yelp) to enrich their local database with additional information. It’s too painful to fully scrape all of Yelp, and the search APIs often have rate limits. Thus, DEEPER figures out how to issue these search queries to specifically enrich the data that you already have, rather than waste resources scraping data that you don’t need. Coauthors: Jiannan Wang and Ryan Shea at SFU. Check out our online demo as well!
Wednesday 2:30–4:30 in SIGMOD Research 9 (room: Administratiezaal) Eugene will present Mining Precision Interfaces From Query Logs. Custom-built interactive visualization speed up data analysis and are easier to use than writing SQL or using generic exploration tools like Tableau. They are also more accessible to non-technical end-users. However, creating custom tailored (i.e., precise) interfaces is too expensive for the vast majority of people and use cases. Precision Interfaces is a first step towards automating the interface creation process by leveraging analysis queries. Coauthors: Owen Zhang, Haoci Zhang, Thibault Sellam.
Thursday 11:30–1 in SIGMOD Research 11 (room: Effectenbeurszaal) Thibault Sellam will present DeepBase: Deep Inspection of Neural Networks. It is our system for generalizing, scaling up, and making it easy to find neurons in a trained NN model that have learned higher level information (e.g., face detectors, language patterns). We believe it is a building block towards better understanding and testing neural network models, and our earlier blog post introduces this work in more detail. Coauthors: Kevin Lin, Ian Huang, Michelle Yang, and Carl Vondrick.