Dr. Jason Zutty, a senior researcher at the Georgia Tech Research Institute (GTRI), presented his work on automated machine learning in a seminar with the Human Augmented Analytics Group (HAAG). His presentation was about research in automating the development of novel machine learning models using Genetic Programming and LLMs. He showed not only how the topology of machine learning models could be searched for using Genetic Algorithms, and how LLMs could combine and alter unstructured python programs in a similar fashion. These techniques can be used to improve existing models or generate novel machine learning architectures His research will be presented at GECCO (The Genetic and Evolutionary Computation Conference) in July.
LLM Guided Evolution: The Automation of Models Advancing Models
