TabSketchFM: Sketch-based Tabular Representation Learning for Data Discovery over Data Lakes
This addresses the need for efficient data discovery in enterprise data lakes, though it appears incremental as it builds on existing tabular neural models with novel pre-training.
The paper tackles the problem of identifying relevant tables in data lakes for tasks like unionability, joinability, and subset detection by introducing TabSketchFM, a neural tabular model with sketch-based pre-training, resulting in significant improvements in F1 scores for table search compared to state-of-the-art techniques.
Enterprises have a growing need to identify relevant tables in data lakes; e.g. tables that are unionable, joinable, or subsets of each other. Tabular neural models can be helpful for such data discovery tasks. In this paper, we present TabSketchFM, a neural tabular model for data discovery over data lakes. First, we propose novel pre-training: a sketch-based approach to enhance the effectiveness of data discovery in neural tabular models. Second, we finetune the pretrained model for identifying unionable, joinable, and subset table pairs and show significant improvement over previous tabular neural models. Third, we present a detailed ablation study to highlight which sketches are crucial for which tasks. Fourth, we use these finetuned models to perform table search; i.e., given a query table, find other tables in a corpus that are unionable, joinable, or that are subsets of the query. Our results demonstrate significant improvements in F1 scores for search compared to state-of-the-art techniques. Finally, we show significant transfer across datasets and tasks establishing that our model can generalize across different tasks and over different data lakes.