Sebastian Hanus / Deborah Hanus

Deborah Hanus

Deborah has done machine learning research at MIT, Harvard, and Google Brain. Her work in machine learning has spanned developing models of human perception to exploring medical data. She has been a teaching assistant for undergraduate classes at MIT, graduate classes at Harvard, and the Boston Python Workshop. Before working in machine learning, she did education research and taught in Cambodia as a Fulbright Scholar. She has spoken at PyTennessee, SciPy Conference, AI With the Best, QConNY, and PyCon US.

Sebastian Hanus

Sebastian loves data analysis, programming, and teaching. As a student research assistant at MIT, he used Python, NumPy, Pandas, and Keras to wrangle gigabytes of voice data (stored as text) into a neural network to detect vocal trauma. As a research assistant at the University of Nebraska, he used Python, NumPy, and sklearn on text data for computer security. In his spare time, he collects and analyzes data to improve his minecraft civilization. He regularly teaches programming and design to the robotics team that he founded and basic computer skills to senior citizens.

Speaker home page

Using Keras & Numpy to Detect Voice Disorders

Python & Libraries, AI & Data, Novice
8/18/2018 | 11:00 AM-11:30 AM | Robertson

Description

This talk is for Python programmers who want to learn how to use Keras to get started with deep learning. The audience should expect to learn what deep learning is, develop an intuitive understanding of how it works, and learn how to avoid some common mistakes. All of this is done via a recurring example, using utterance data to determine whether a medical patient might have a voice disorder.

Abstract

Deep learning is a useful tool for problems in computer vision, natural language processing, and medicine. While it might seem difficult to get started in deep learning, Python libraries, such as Keras make deep learning quite accessible. In this talk, we will discuss what deep learning is, introduce NumPy and Keras, and discuss common mistakes and debugging strategies. Throughout the talk, we will return to an example project in the medical domain, which used deep learning on vocal data to determine whether a patient has a voice disorder called vocal hyperfunction.