News
Paper on repairing a query to fulfill constraints on the query’s result accepted at VLDB 2026
We got a paper accepted at PVLDB on repairing queries to fulfill constraints expressed as aggregations over the results of a query. Such constraints are powerful enough to express common fairness metrics and other types of constraints users may want to enforce for their queries.
Revived course CS581 - Database Management Systems
In this course we will look under the hood of database systems and learn about the algorithms, systems design, and implementation techniques that are used to build database systems.
Paper on stress-testing ML pipelines accepted at VLDB 2026
We got a paper accepted at PVLDB on stress-testing ML pipelines with adversarial data corruption.
Paper on computing minimal synthetic witnesses for queries accepted at PODS 2025
Stavros’s Ph.D. student Aryan got a paper accepted at PODS 2025.
Paper on incremental maintenance of provenance sketches accepted at EDBT 2026
Pengyuan got a paper accepted EDBT 2026 on incremental maintenance of provenance sketches under updates.
Paper on repairing labeling functions accepted at KDD 2025
Chenjie got a paper accepted at KDD 2025 on repairining labeling functions in programmatic weak supervision using small amounts of labeled examples.
Boris attending Dagstuhl seminar
Semirings in Databases, Automata, and Logic
Paper on efficient model search accepted at SIGMOD 2025
We got a paper accepted at SIGMOD 2025 on improving the performance of model search for transfer learning.
Open Ph.D. positions
Are you interested in data science / databases and are passionate about research? Consider joining our group.
Papers on learning on uncertain data and probabilistic query processing accepted at NeurIPS and SIGMOD
We got a paper about learning from uncertain data accepted at NeurIPS 2024 and a paper on using approximate query processing for approximating bag probabilistic queries accepted at SIGMOD 2025.
New course CS594 - Provenance & Explanations in Fall 2024
Learn how to build explainable, trust-worthy, secure, transparent, and fair systems for data analysis and machine learning.
Boris is moving to UIC in Spring 2024
Prof. Glavic will be joining University of Illinois, Chicago in January 2024.
Workshop at Simons Institute on Logic and Algebra for Query Evaluation
Boris presents our group’s work on uncertain data management at Berkeley
Su Feng Graduation
Congrats to Su for graduating! Su will continue to work with our group as a PostDoc.
SIGMOD Blog: Why Uncertainty is Unavoidable and What We Can Do About That
Boris discusses why uncertainty is unavoidable in data analysis and gives an overview of our work on uncertain data management.