Making Sense of Touch: Unsupervised Shapelet Learning in Bag-of-words Sense

TL;DR

This paper presents NN-STNE, a neural network that uses t-SNE for dimensionality reduction and shapelet learning in time-series data, improving clustering accuracy in robotics applications.

Contribution

It introduces a novel unsupervised shapelet learning method combining t-SNE with neural networks and regularization, addressing crowding issues and optimizing shapelet length.

Findings

Improved clustering accuracy on UCR dataset.

Effective shapelet learning for robotic manipulation tasks.

Outperforms state-of-the-art feature-learning methods.

Abstract

This paper introduces NN-STNE, a neural network using t-distributed stochastic neighbor embedding (t-SNE) as a hidden layer to reduce input dimensions by mapping long time-series data into shapelet membership probabilities. A Gaussian kernel-based mean square error preserves local data structure, while K-means initializes shapelet candidates due to the non-convex optimization challenge. Unlike existing methods, our approach uses t-SNE to address crowding in low-dimensional space and applies L1-norm regularization to optimize shapelet length. Evaluations on the UCR dataset and an electrical component manipulation task, like switching on, demonstrate improved clustering accuracy over state-of-the-art feature-learning methods in robotics.