GobTech

Name: Dale B Mc Cauley
Affiliation: College of Business - Austin Entrepreneurship Program
Phone: 5417138042
E-mail: mccauled@oregonstate.edu
Website: https://gobtech.io
Knowledge Required: An understanding of neural networks, as well as genetic algorithms, is required. Knowledge of iPhone and/or Android development environments is highly valuable for this project.
Motivation: With the recent explosion of neural network innovation, many industries are being disrupted. One such industry that has yet to see much disruption is the gaming industry, which could largely be due to the fact that different AIs in different games require different network topologies. This is where the neuroevolution comes in. Game developers could save both time and money if they had easy access to this type of technology, in the form of a NEAT engine.
Description: This project is an entrepreneurial startup called GobTech, supported by OSU InnovationX, and it is working on developing AI through a combination of genetic algorithms and neural networks. It is currently focused on testing AI technology through a demonstration app called Neural Sandbox. As this year begins, the focus will shift to developing a generalized NEAT (neuroevolution of augmenting topologies) engine for use in video game AI and/or finance.
Objectives: Our objective is to fully develop a NEAT engine, SNAP-NEAT specifically, for use for game developers. This engine should be easy to use for the end user, and as computationally efficient as possible.
Deliverables: The finished AI engine will be the main deliverable. It will be able to provide a statistically significant improvement to the AI in the demonstration app while not giving up on computational resources.
Other comments: It is very often times the best practice for a startup to pivot somewhat as they develop their product. This is an entirely possible for this project and potential extensions of this system include building an AI based trading tool for use in financial markets.

   D. Kevin McGrath
   Last modified: Mon Oct 16 16:07:06 2017