Nuisance-Label Supervision: Robustness Improvement by Free Labels
arXiv:2110.07118
Abstract
In this paper, we present a Nuisance-label Supervision (NLS) module, which can make models more robust to nuisance factor variations. Nuisance factors are those irrelevant to a task, and an ideal model should be invariant to them. For example, an activity recognition model should perform consistently regardless of the change of clothes and background. But our experiments show existing models are far from this capability. So we explicitly supervise a model with nuisance labels to make extracted features less dependent on nuisance factors. Although the values of nuisance factors are rarely annotated, we demonstrate that besides existing annotations, nuisance labels can be acquired freely from data augmentation and synthetic data. Experiments show consistent improvement in robustness towards image corruption and appearance change in action recognition.
ICCV 2021 Workshop
References in corpus (6)
- UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
- AugMix: A Simple Data Processing Method to Improve Robustness and Uncertainty
- Temporal Segment Networks: Towards Good Practices for Deep Action Recognition
- R-CNNs for Pose Estimation and Action Detection
- Why Can't I Dance in the Mall? Learning to Mitigate Scene Bias in Action Recognition
- RSA: Randomized Simulation as Augmentation for Robust Human Action Recognition