Poster Abstract

Throughout the years American Heritage School has been a longstanding academic partner with HPCC Systems. Through this partnership and multiple interns from American Heritage, an autonomous security robot was created and continually improved upon. This new project was designed to investigate a new possibility on whether generative AI could be used to generate human-like responses in robots. During the internship an interface through HPCC systems was constructed to interface with GPT and ChatGPT. During this project a generalized interface was created for HPCC Systems to access GPT and ChatGPT. From there the steps were taken to use HPCC Systems to train a neural networkmodel capable of classifying faces into different emotions. These emotions would then be processed by the interface to create a call to OpenAI’s API from which an appropriate response would be generated. This response would be designed to change based on the emotions classified by the neural network and most importantly due to the use of generative AI would be different even with the same emotional classification. However due to the limited time frame of the project, which precluded the training of a neural network for detecting emotions, the project was focused on exploring the most unknown and uncertain part of the project, OpenAI’s Generative AI and their respective API. This endeavor suggests that emotion driven interactions between humans and robotics is a promising avenue of future research.

Presentation

In this Video Recording, Carlos provides a tour and explanation of his poster content.

Practical Application of Generative AI Technology 

Click on the poster for a larger image.