Date of Submission

Spring 2021

Academic Program

Computer Science

Project Advisor 1

Sven Anderson

Abstract/Artist's Statement

Neural Networks, a form of machine learning, are used in increasingly important roles in the modern world. They are being used in self-driving cars and medical diagnoses. However, they are “Black Boxes”: they cannot be easily interpreted by humans. This project combines two methods of explaining a neural network’s decisions in an attempt to improve their accuracy. This new method, relevance-based testing with concept activation vectors (R-TCAV), yields promising results on two small experiments but is less precise than the previous TCAV method.

Open Access Agreement

Open Access

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

This work is protected by a Creative Commons license. Any use not permitted under that license is prohibited.