Recap: AI Deep Dive Workshop: ChatGPT [Video and Slides] 

By: David Bolton

A new collaborative workshop co-hosted by the Institute for Experiential AI and AI Skunkworks at Northeastern University kicked off with a student-led talk about ChatGPT. The first in a regular cadence of workshops on AI tools and theories, the AI Deep Dive series is designed to be an interactive and practical introduction to current data science topics. 

According to Nik Bear Brown, associate teaching professor at the College of Engineering at Northeastern University and leader of AI Skunkworks, this series is an important step towards building a community of both experienced and nascent AI researchers. Presented by students, for students, these monthly workshops will explore the theoretical foundations of cutting-edge AI models. Additionally, open-access online tutorials will both increase AI literacy and provide attendees with an introduction to AI topics. 

“The whole point of this group is to test AI and play with things,” Brown said. “With these asynchronous talks, we are going to record videos with tutorials and hands-on instructions on how to do things. The main thing about this AI Skunkworks group and all these planned talks is learning by doing. By playing with it. We have a bunch of smart students here talking about the theory, but we want to build a community that will actually do it.”

Let’s talk ChatGPT

The inaugural workshop focused on ChatGPT, featuring talks by four Northeastern students, Shesh Narayan Gupta, Arjun Bhatia, Rishabh Indoria, and Tanvi Manohar Valkunde.  

The student speakers presented findings based on their recent engagements with this specific large language model. This included the evolution of GPT, the importance of the right inputs for neural networks, the advantages and limitations of the model, and its aptitude for sentiment analysis. In addition, the talk flagged up the opportunities for future work by the “skunks” and current concerns about AI bias, copyright in generative AI art, the limitations of data, and the need for human intervention.

“EAI is an institute filled with smart people who are doing research on AI,” Brown said. “AI Skunkworks is also a bunch of smart people, with students learning AI by doing AI. Part of that is learning how to communicate. The next generation of AI researchers needs to learn how to present ideas and how to talk outside the classroom. These are critical skills in AI.”

Flip through the slides used in the presentation below and sign up to stay in the loop about AI research, workshops and career fairs.