Diverse Complexity Measures for Dataset Curation in Self-driving
This addresses the challenge of efficient dataset curation for self-driving systems, where labeling all data is infeasible, by providing a method that is not tied to specific models or tasks, though it is incremental relative to active learning approaches.
The paper tackles the problem of selecting which raw self-driving data to label by introducing a data curation method that uses diverse criteria to quantify scene interestingness, leading to improved generalization and higher performance across various tasks and models.
Modern self-driving autonomy systems heavily rely on deep learning. As a consequence, their performance is influenced significantly by the quality and richness of the training data. Data collecting platforms can generate many hours of raw data in a daily basis, however, it is not feasible to label everything. It is thus of key importance to have a mechanism to identify "what to label". Active learning approaches identify examples to label, but their interestingness is tied to a fixed model performing a particular task. These assumptions are not valid in self-driving, where we have to solve a diverse set of tasks (i.e., perception, and motion forecasting) and our models evolve over time frequently. In this paper we introduce a novel approach and propose a new data selection method that exploits a diverse set of criteria that quantize interestingness of traffic scenes. Our experiments on a wide range of tasks and models show that the proposed curation pipeline is able to select datasets that lead to better generalization and higher performance.