Probabilistic Discriminative Models Address the Tactile Perceptual Aliasing Problem
This addresses a critical issue for robotics, particularly in uncertain environments like those using deep reinforcement learning, though it is an incremental improvement over existing methods.
The paper tackles the tactile perceptual aliasing problem in deep neural networks and discriminative models, where ambiguous tactile data leads to poor predictions, and shows that a probabilistic discriminative model using a mixture density network improves prediction accuracy by identifying aliased data.
In this paper, our aim is to highlight Tactile Perceptual Aliasing as a problem when using deep neural networks and other discriminative models. Perceptual aliasing will arise wherever a physical variable extracted from tactile data is subject to ambiguity between stimuli that are physically distinct. Here we address this problem using a probabilistic discriminative model implemented as a 5-component mixture density network comprised of a deep neural network that predicts the parameters of a Gaussian mixture model. We show that discriminative regression models such as deep neural networks and Gaussian process regression perform poorly on aliased data, only making accurate predictions when the sources of aliasing are removed. In contrast, the mixture density network identifies aliased data with improved prediction accuracy. The uncertain predictions of the model form patterns that are consistent with the various sources of perceptual ambiguity. In our view, perceptual aliasing will become an unavoidable issue for robot touch as the field progresses to training robots that act in uncertain and unstructured environments, such as with deep reinforcement learning.