Improving Robot Localisation by Ignoring Visual Distraction
This addresses the challenge of ignoring visual distractors for robots in dynamic environments, representing an incremental advancement in attention mechanisms.
The paper tackles the problem of visual distraction in robot localization by introducing Neural Blindness, a method that makes neural networks incapable of representing specific distracting classes in their latent space, resulting in improved localization performance.
Attention is an important component of modern deep learning. However, less emphasis has been put on its inverse: ignoring distraction. Our daily lives require us to explicitly avoid giving attention to salient visual features that confound the task we are trying to accomplish. This visual prioritisation allows us to concentrate on important tasks while ignoring visual distractors. In this work, we introduce Neural Blindness, which gives an agent the ability to completely ignore objects or classes that are deemed distractors. More explicitly, we aim to render a neural network completely incapable of representing specific chosen classes in its latent space. In a very real sense, this makes the network "blind" to certain classes, allowing and agent to focus on what is important for a given task, and demonstrates how this can be used to improve localisation.