Number of individual students required: 6

Be part of solving the critical problem of AI explainability. Explainability of AI and Machine Learning is a significant challenge, as is humanity's capacity to keep up with & govern ever-improving algorithms.

Objective of this project is to apply our patented technology for visualizing & explaining an AI or machine learning model of your choice. For inspiration, we have to-date applied the technology to visualizing nodes of a Tensorflow model and for seeing dimensions used to train a random forest classifier in context of model-determined outcomes vs. actual outcome from the training dataset.

Immersion Analytics' patented technology and products make it possible to see up to 18 dimensions of data as a single, intuitive visualization. This is accomplished using special effects as added axes, intensity of each effect is varied at each data point to convey additional dimensions. Our Stepwise Storyboarding capability permits adding dimensions one at a time as you tell the data story.

Students will receive a personal-use license for Immersion Analytics Visualizer, which can be used in concert with Python, MATLAB, Qlik Sense, and even CSV data files. We have many samples for use with Python, several for MATLAB, and extensive documentation and install support.

Students should be prepared to:

• Identify an AI or machine learning project where up to 18 dimensions, primarily numeric, are used to train the model. This can be a project from github or elsewhere online, or one of the student's design.
• Develop immersive visualization using our Python package or MATLAB toolbox alongside our Visualizer product
• Prepare a 2-page data storytelling walkthrough of your creation, with screenshots (greenshot is fine).

While we are open to your inspirations, a few areas of specific interest for us include:

• Predicting customer lifetime value
• Recommendation engines
• Fraud detection

Company CEO and/or CTO will mentor directly and are available up to 5 hours weekly.