Domain-specific Generative AI  (DoGAI)

Domain-specific Generative AI  (DoGAI)

The advent of Generative AI based on pre-trained large language models (LLMs) has recently elicited much interest. Popular examples of generative AI are ChatGPT 3.5, 4.0, Bing AI, Bard and many more. So far, Generative AI systems have produced impressive and mixed results, adjudged able to do well on general knowledge questions, but poorly on domain-specific questions and questions requiring higher-order reasoning. The goal of this research is to explore the augmentation and fine-tuning of LLMs with domain-specific knowledge that will lead to improved performance in specific fields of application (such as healthcare, education, commerce, and business).

Participation: 

1. Prospective Researchers: Bachelor honours, master's, doctoral students, postdoctoral fellows, research collaborators.

2. Skill Requirements: Expertise in programming (Python/machine learning/JavaScript/web development), software development, good communication and writing skills.

For more information on DoGAI research, contact me via my email: wande.daramolaj@up.ac.za