Multi-Task Learning for Robot Perception with Imbalanced Data
This addresses the challenge of limited and imbalanced data in multi-task learning for mobile robots, which is an incremental improvement in a domain-specific context.
The paper tackles the problem of multi-task learning for robot perception with imbalanced data, proposing a method that learns tasks even without ground truth labels for some tasks and showing how to identify which tasks improve others, with empirical results on semantic segmentation and depth estimation using NYUDv2 and Cityscapes datasets.
Multi-task problem solving has been shown to improve the accuracy of the individual tasks, which is an important feature for robots, as they have a limited resource. However, when the number of labels for each task is not equal, namely imbalanced data exist, a problem may arise due to insufficient number of samples, and labeling is not very easy for mobile robots in every environment. We propose a method that can learn tasks even in the absence of the ground truth labels for some of the tasks. We also provide a detailed analysis of the proposed method. An interesting finding is related to the interaction of the tasks. We show a methodology to find out which tasks can improve the performance of other tasks. We investigate this by training the teacher network with the task outputs such as depth as inputs. We further provide empirical evidence when trained with a small amount of data. We use semantic segmentation and depth estimation tasks on different datasets, NYUDv2 and Cityscapes.