Evaluating the effect of data augmentation and BALD heuristics on distillation of Semantic-KITTI dataset

TL;DR

This paper evaluates Bayesian active learning and data augmentation techniques for dataset distillation in LiDAR perception, revealing that these methods may negatively impact labeling efficiency on the Semantic-KITTI dataset.

Contribution

It provides an empirical assessment of Bayesian active learning and data augmentation effects on LiDAR dataset distillation for autonomous driving.

Findings

Data augmentation and BALD heuristics can reduce labeling efficiency.

Challenges in designing effective active learning frameworks for LiDAR data.

Negative impact observed when applying these methods to large-scale datasets.

Abstract

Active Learning (AL) has remained relatively unexplored for LiDAR perception tasks in autonomous driving datasets. In this study we evaluate Bayesian active learning methods applied to the task of dataset distillation or core subset selection (subset with near equivalent performance as full dataset). We also study the effect of application of data augmentation (DA) within Bayesian AL based dataset distillation. We perform these experiments on the full Semantic-KITTI dataset. We extend our study over our existing work only on 1/4th of the same dataset. Addition of DA and BALD have a negative impact over the labeling efficiency and thus the capacity to distill datasets. We demonstrate key issues in designing a functional AL framework and finally conclude with a review of challenges in real world active learning.