Feature Search using Automatic Machine Learning

This research project focuses on developing an automated system to search and analyze time-series tabular features in the financial institution’s machine learning pipeline. The goal is to identify relevant features and improve efficiency in the decision-making process. The project will begin by prototyping a system to support automated feature search patterns and researching feature search approaches. The system will be tested and deployed in a use case, with appropriate governance for production systems. Successful completion of the project will contribute to the feature search automation of the machine learning pipeline at the financial institution.

Faculty Supervisor:

Gennady Pekhimenko

Student:

Partner:

Layer 6 AI

Discipline:

Computer science

Sector:

Technology; Finance and Insurance; Artificial Intelligence

University:

University of Toronto

Program: