Insurance Vehicle Damage Claim Price Prediction

The main purpose of the internship is to research and create a product that can be used by insurance companies to streamline their claim processing pipeline. The task is to take images and text provided in an insurance claim and generate a prediction for the cost to repair the damage that is closer to the real cost of repair than can be achieved through human observers. The idea would be to use a sequence of models to analyze the images and text and to generate a real value prediction of the cost of repairs. The main benefits for the partner organization are a viable product for claim price prediction and the increased knowledge of the team in the theoretical and real world application of the cutting edge of research.

Faculty Supervisor:

Richard Zemel

Student:

Partner:

Nanotech Galaxy Inc

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of Toronto

Program: