BOD: Blindly Optimal Data Discovery
Thomas Hoang
TL;DR
This paper tackles data discovery when the user’s utility function is unknown by introducing BOD, a human-in-the-loop framework that ranks variables across augmented datasets and iteratively filters tuples to discover high-potential data without predefined preferences. The authors provide a formal performance guarantee and show, through experiments on synthetic datasets and the Boston Housing corpus, that BOD achieves higher precision and lower mean absolute error than established baselines such as Top-K, Skyline, and Pareto. The results demonstrate BOD’s promise for data-driven tasks where user preferences are hard to specify, offering a practical approach to efficiently assemble small, informative datasets for prediction tasks.
Abstract
Combining discovery and augmentation is important in the era of data usage when it comes to predicting the outcome of tasks. However, having to ask the user the utility function to discover the goal to achieve the optimal small rightful dataset is not an optimal solution. The existing solutions do not make good use of this combination, hence underutilizing the data. In this paper, we introduce a novel goal-oriented framework, called BOD: Blindly Optimal Data Discovery, that involves humans in the loop and comparing utility scores every time querying in the process without knowing the utility function. This establishes the promise of using BOD: Blindly Optimal Data Discovery for modern data science solutions.