UROP Openings

Have a UROP opening you would like to submit?

Please fill out the form.

Submit your UROP opening

Emergent deep neural network architectures for biomedical and clinical datasets for improving human health (Remote UROP)


Term:

Fall

Department:

MAS: Media Arts and Sciences

Faculty Supervisor:

Pratik Shah

Faculty email:

pratiks@mit.edu

Apply by:

September 10

Contact:

Sam Ghosal: sghosal@media.mit.edu

Project Description

Co-authorship on publications and opportunity to apply deep learning to develop deployable solutions . Deidentified numerical measurements, images, and videos from clinical datasets from human subjects are available to develop novel deep learning tools. Your will be trained to create visualizations that allow us to understand the data along its multiple dimensions and to identify areas for deep learning analyses. This includes creating a landing page for these visualizations and this project in general. You will also implement novel deep learning algorithms to classify this and other publicly available datasets for real-world use. Preferred technical skills: Loading and processing large data from files in text format (e.g., csv, json, xml, etc.) using programming tools (e.g., Python Pandas, R, etc.). Computational graph and auto-differentiation tools (e.g., Pytorch, Tensorflow, Theano, etc.) for deep learning models. Visualization tools (e.g. Matplotlib, Tableau, Jupyter Notebook, etc.). Parallel processing tools (e.g. Python Multiprocessing, MPI, etc.). Basic statistical tools (e.g. linear regression, null hypothesis tastings, model comparison, etc.).

Pre-requisites

Prerequisites: Familiarity and interest in working with data visualization, web programming for clinical deep learning methods. Knowledge/coursework in statistics and EECS is preferred. Include a brief cover letter, resume, a list of related coursework, and other relevant material (projects, portfolio, etc.).