This ongoing client-project applies A.I and machine learning to entrepreneurial prototyping. The project builds on an existing alpha-stage web-based system to improve the functionality of GANspace, a powerful explorer and visualization engine for image-centric machine learning models.
Generative Adversarial Networks (GANs) are a disruptive A.I. technology with the potential to influence design, innovation, and educational practices. The project provides the opportunity to work with an outside client to improve the workflows and capabilities of the existing system. Broadly, the system allows users to create/visualize and share new product design concepts (e.g., new breakthrough car designs or consumer product designs) for different product categories.
Working from a semester project plan, the student will work on a development team and implement and test a series of functionality enhancements and workflow improvements encompassing: interpolation videos between user-created designs, adaptive discriminator augmentation and t-SNE to simplify model training, and BigBiGan to locate existing real-world designs in created models. Additionally, the overall user-friendliness of the UI/UX of the system will be improved.
Considered candidates must be 100% professional and reliable, hitting all agreed deadlines, in addition to possessing the needed technical skills.
Should have a demonstrated passion for A.I. and machine learning; direct knowledge about GANs is helpful but not required.
Will experiment with the parameters of the GANspace model, pushing its known capabilities.
Will be expected to work hand-in-hand with other coops/team members working on this project. Students must be able to work independently and collaboratively with others.
Excellent documentation will be essential to ensure others can continue the project thereafter. Client satisfaction with the end product and student professionalism over the course of the project (i.e., professional communication, hitting deadlines, demonstrated creativity) will determine final grading.