UROP Openings

Have a UROP opening you would like to submit?

Please fill out the form.

Submit your UROP opening

Automatic Encoding Selection


Term:

Fall

Department:

6: Electrical Engineering and Computer Science

Faculty Supervisor:

Prof. Kraska

Faculty email:

kraska@mit.edu

Apply by:

9/9/80

Contact:

Andreas Kipf, kipf@mit.edu

Project Description

Selecting an appropriate encoding (e.g., run-length encoding) for a particular data column is an important problem in data warehousing. The goal of this UROP is to develop an automatic encoding selector that takes column statistics and workload requirements as input and recommends an encoding. Specifically, the task is to develop a cost model that predicts the compression ratio and the (de)compression speed based on the data representation.

Pre-requisites

Machine learning, deep learning, reinforcement learning (PyTorch). Experience with C++ is a plus.