Recurrent Systems for EMG Based Motor Control - Part 3


Come hear the culmination of a series of talks by UMass Applied Math MS students Connor Amorin, Gabriel P. Andrade, Chris Brissette, Matthew Gagnon, Brandon Iles, Jimmy Smith, and Lance Wrobel about their research in biological signal processing and novel classification methods using machine learning, presented by Undergraduate Researchers Interested in Data at UMass Amherst (

Abstract: During two talks given over the last few months we introduced the problem of classifying EMG signals and built some intuition about how certain recurrent networks perform the tasks that they do. This time we will wrap everything up by demonstrating how these recurrent systems can be used to classify time series data (like EMG), how well they do, why the results might be what they are, and how these results can be improved. Furthermore, we will discuss considerations that should be taken into account when deciding between one model vs. another. As has been the case during the previous talks, emphasis will be put on recurrent neural networks (RNNs) and reservoir computing but no prior knowledge is assumed.

For additional information about their project, read the more high-level project overview here ( as well as the abstract from the most recent talk here (