Older News

• Jul-2023: Delighted to recieve a grant from Columbia’s Center of AI Technology to study differentially private dataset search with Rachel Cummings.
• May-2023: SIGMOD 2023 is next month! We have exciting work on a Data Interface Grammar, how to detect and avoid logical errors in semantic layers, CPU-GPU acceleration for In-DBMS machine learning, a suite of XTutor visualization libraries for data education, and the OM3 progressively encoded data structure for fast and massive-scale time series visualizations!
• Dec-2022: Zachary recieved the Columbia Data Science Institute’s Avanessian PhD fellowship!
• Sep-2022: IEEE VIS 2022 is next month! Excited to present our work on generating fully interactive interface from natural language questions at the nlviz workshop, our recommendations for how to write captions for data visualizations at the viscomm workshop, and our new formalism and interaction design for ad hoc comparisons at the main conference!
• Aug-2022: A nice Columbia Engineering article about our research to help insure African farmers from droughts and floods.
• Jun-2022: SIGMOD 2022 is this week! We are presenting 3 full papers, 1 DEEM workshop paper, 2 demos, and student research competition posters! We summarize this work in our SIGMOD 22 Blog post!
• Mar-2022: PI2 is the first system that generates fully interactive visualization analysis interfaces from queries. We are excited to present it at SIGMOD 2022!
• Feb-2022: Excited to have been recommended for tenure at Columbia University!
• Feb-2022: our paper View Composition Algebra for Ad hoc Comparisons was accepted to TVCG 2022 and will be presented at VIS later this year.
• Jan-2022: 2 more paper acceptances to SIGMOD 2022: our demo paper Interactive Interface Generation in Notebooks extends jupyter to automatically generate interactive analysis interfaces during SQL analysis, and our full paper Complaint-Driven Training Data Debugging at Interactive Speeds enables training data debugging by simply identifying errors in analytic results in milliseconds!
• Jan-2022: FLAWLESS VICTORY! Our mortal kombat-themed system, Reptile reformulates data cleaning as a top-down iterative process. It has been used in Zambia and Ethiopia to clean nation-wide farmer survey data in order to design their national drought insurance policies. Zachary will present it at SIGMOD 2022.
• Older News
• Jun-2021: Looking forward to giving a keynote at the SEA DATA workshop at VLDB 2021 this summer!
• Aug-2020: For Highly Interactive Apps, Prediction Is Not Enough! is a blog post to introduce our Khameleon paper. Haneen also recorded a short YouTube video summarizing our work.
• Jul-2020: FLAWLESS VICTORY! Khameleon, our rethink of client-server communication for interactive applications will be presented at VLDB 2020! With Haneen Mohammed, Tracy Wei, and Ravi Netravali. This makes the sixth mortal kombat-based system that has fallen to science.
• Jun-2020: Haneen participated in, and won, first place at the 2020 SIGMOD student research competition for her work on Khameleon!!
• Mar-2020: FATALITY! A new mortal kombat-themed system has been beaten into submission. Our full paper about Rain: Complaint-driven Training Data Debugging for Query 2.0 will be presented at SIGMOD 2020! With Young Wu, Lampros Flokas, and Jiannan Wang.
• Feb-2020: Excited to share our demo for Physical Visualization Design at SIGMOD 2020! With Lana Ramjit, Zhaoning Kong, and Ravi Netravali.
• Feb-2020: Lampros will present our work Towards Complaint-driven ML Workflow Debugging at the MLOps workshop, part of MLSys! With Young Wu and Jiannan Wang.
• Feb-2020: Yiru presented our work on layout-sensitive interface generation at the Intelligent Process Automation workshop at AAAI!
• Dec-2019: Delighted to recieve the NSF CAREER for automatically generating interactive data analysis interfaces!
• Aug-2019: Eugene gives a tutorial at the VLDB Summer School in Beijing, talk at 8th Workshop for Big Data Analysis and Systems in China, and at Tsinghua University.
• Jul-2019: Our undergrads Lauren Arnett and Robert Netzorg will be presenting their late breaking work on studying popularity on live streaming platforms at CHI next month!
• Jun-2019: Our papers DeepBase: Deep Inspection of Neural Network, and Mining Precision Interfaces From Query Logs will be presented at SIGMOD 2019!
• Looking forward to presenting our Deep Neural Inspection work at LearningSys@Neurips!
• News article about our At-a-Glance paper at VIS 2018!
• New blog post about WuLab at VLDB 2018
• Extremely honored to win the 2018 VLDB 10-Year Test-of-Time Award for WebTables: exploring the power of tables on the web. Has it really been 10 years?
• Our At a Glance: Pixel Approximate Entropy paper that proposes a quantitative measure of perceptual chart complexity has been accepted to InfoVIS 2018! Congrats Gabriel Ryan and Abby Mosca!
• Created a Blog for the WuLab to describe the research that we do and the assistant professor experience.
• Fotis will present the deep connections between Provenance and Interactive Interfaces at HILDA 2018, and a demo of Smoke at SIGMOD 2018!
• Thibault Sellam will demo our work with HaoCi and Viraj on automatically generating multi-modal exploration interfaces from query logs at SIGMOD 2018!
• Hamed Nilforoshan will present his Writing Feedback Generation work at ICWSM 2018 in Stanford!
• Thankful to win the Google Research Faculty Fellowship!
• Excited for Thibault Sellam to present our vision of bringing software engineering concepts to understanding neural networks!
• Won the Amazon Faculty Research Award!
• Released a probabilistic model of research
• New VLDB 18 paper: Fine-grained data lineage at SMOKIN’ fast speeds!! Fast enough and low enough overhead to power real-time interactive applications!
• Our undergrad Hamed Nilforoshan won the Snap Research Scholarship!!
• Mining Precision Interfaces from Query Logs is our work on scalably generating analysis interfaces.
• BoostClean continues our work towards automated human-in-the-loop data cleaning for data science!
• Webpage to track Mortal Kombat ninjas that we have declared flawless victory over.
• ActiveClean was covered by The Morning Paper!
• Visited and spoke at DB groups at Wisconsin-Madison, University of Chicago, and UIUC. Slides: 1 per page, and 6 per page.
• Check out our DSIA paper Load-n-go, InfoVIS poster on entropy measure for visualization, and DECISIVE paper on bayesian models of cognition for interactive visualization.
• Check out our two 2017 HILDA papers! PALM: Machine Learning Explanations For Iterative Debugging and Precision Interfaces!
• A nice article about our HILDA and SIGMOD papers. See you in Chicago!
• Hamed’s paper on Segment-Predict-Explain for Automatic Writing Feedback was accepted to Collective Intelligence! Stay tuned for the full paper!
• Our CIDR paper on Combining Design and Performance in a Data Visualization Management System was accepted!
• Our QueryFix full paper was acceped to SIGMOD 2017! Come see our presentation in Raleigh, NC!
• There is a nice article about ActiveClean on the DSI website: A Data-Cleaning Tool for Building Better Prediction Models
• Our NSF proposal with jmh and jheer on bringing declarative database ideas to interactive visualizations was funded! The DVMS train continues to roll.
• Our ActiveClean submission won Best Demo at SIGMOD 2016! Many thanks to our collaborators Sanjay, Jiannan, Mike, and Ken!
• A super simple, end-to-end implementation of ActiveClean is available at this dropbox folder
• Our ActiveClean full paper was acceped to VLDB 2016! Come see our presentation in Dehli, India. Look forward to our code release later this summer!
• Our first (Data+Media):Humans in the Loop speaker series event is Sep 28, 7-9PM! Everyone welcome.
• Talk about what provenance is and how it relates to our projects. MIT BigData Workshop 2016
• Congrats to Niranjan, Arnab, Yifan, Daniel Haas, Sanjay, and Daniel Alabi for getting four papers accepted at SIGMOD’s Hilda workshop!
• QueryFix explanation demo with Xiaolan and Alexandra accepted to SIGMOD 2016!
• CLAMShell paper with Daniel Haas for drastically speeding up crowds accepted to VLDB 2016!