About the workshop
Relational data represents the vast majority of data present in the enterprise world. Yet none of the ML computations happens inside a relational database where data reside. Instead a lot of time is wasted in denormalizing the data and moving them outside of the databases in order to train models. Relational learning, which takes advantage of relational data structure, has been a 20 year old research area, but it hasn’t been connected with relational database systems, despite the fact that relational databases are the natural space for storing relational data. Recent advances in database research have shown that it is possible to take advantage of the relational structure in data in order to accelerate ML algorithms. Research in relational algebra originating from the database community has shown that it is possible to further accelerate linear algebra operations. Probabilistic Programming has also been proposed as a framework for AI that fits can be realized in relational databases. Data programming, a mechanism for weak/self supervision is slowly migrating to the natural space of storing data, the database. At last as models in deep learning grow several systems are being developed for model management inside relational databases. This workshop aspires to start a conversation on the following topics:
- What is the impact of relations/relational structure in machine learning?
- Why has relational learning not been more successful? Why we don’t have yet the equivalent of tensorflow/pytorch in relational learning?
- Why is there no deep network structure for structured relational data? Are we just not there yet, or is there something intrinsic in random forest/boosted trees that work better for relational data?
- Can relational databases take advantage of the relational nature of graph neural network
- The algorithms and db communities have completely different approaches to relational learning, what is the connection?
- How does data programming connect to relational learning and can it be accelerated with the algorithmic primitives of relational databases?
- The attention network has been interpreted and used as a mechanism for discovering and expressing relations. It has also been considered as a storage mechanism of knowledge in Large Language Models (Transformers). Are transformers equivalent to databases?
Call for papers
Areas of particular interest for the workshop include (but are not limited to):
- Data Management in Machine Learning Applications
- Definition, Execution and Optimization of Complex Machine Learning Pipelines
- Systems for Managing the Lifecycle of Machine Learning Models
- Systems for Efficient Hyperparameter Search and Feature Selection
- Machine Learning Services in the Cloud
- Modeling, Storage and Provenance of Machine Learning Artifacts
- Integration of Machine Learning and Dataflow Systems
- Integration of Machine Learning and ETL Processing
- Definition and Execution of Complex Ensemble Predictors
- Sourcing, Labeling, Integrating, and Cleaning Data for Machine Learning
- Data Validation and Model Debugging Techniques
- Privacy-preserving Machine Learning
- Benchmarking of Machine Learning Applications
- Responsible Data Management
- Transparency and Accountability of Machine-Assisted Decision Making
- Impact of Data Quality and Data Preprocessing on the Fairness of ML Predictions
Submission: Submissions can be short papers (4 pages) or long papers (up to 8 pages, plus unlimited references). Authors are requested to prepare submissions following the NeurIPS proceedings format. DBAI is a single-blind workshop, authors must include their names and affiliations on the manuscript cover page. Submission Website: To be announce upon acceptance of the workshop Inclusion and Diversity in Writing: http://2021.sigmod.org/calls_papers_inclusion_and_diversity.shtml
Conflicts: Workshops are not a venue for work that has been previously published in other conferences on machine learning or related fields. Work that is presented at the main NeurIPS conference will not be accepted in the workshop, including as part of an invited talk..
Paper submission deadline:
Sep 17, 2021, 11:59 PM (AoE, UTC-12) Extended Deadline: Sep 24, 2021, 11:59 PM (AoE, UTC-12)
Acceptance notification: Oct 22, 2021 EOD
Mandatory SlidesLive upload for speaker videos: Nov 08, 2021
Workshop day: Dec 13, 2021
Submission link: https://openreview.net/group?id=NeurIPS.cc/2021/Workshop/DBAI
All questions about submissions should be emailed to this address.
- Nikolaos Vasiloglou (RelationalAI)
- Maximilian Schleich (University of Washington)
- Nantia Makrynioti (Centrum Wiskunde & Informatica)
- Parisa Kordjamshidi (Michigan State University)
- Kirk Pruhs (University of Pitsburg)
- Zenna Tavares (MIT)