FAQ System Using Generative AI and Document Analysis

Project description

The client company CY2 wishes to make use of Large Language Models (such as Open AI’s ChatGPT) to power their chat assistant. Our team is helping by investigating how good the LLMs perform in answering questions based on custom information as well as discovering how to improve the answers. The goal is to come up with a set of requirements that show how the custom data should be presented to the LLM and what configuration should be used so that it gives the best answer in terms of grammar and factual correctness.

Context

CY2 is a Dutch company founded in 2007 with headquarters in Den Bosch, The Netherlands, that offers a wide variety of campus solutions to university institutions located in North America, Europe, Africa, and Asia. Among products and services, CY2 is developing a chat assistant working with Ai that would be able to answer open questions. The model is being trained with university data such as information about enrolment, schedules, study materials, etc.

Results

Our team is performing various tests on the data that would be displayed in a Power BI dashboard. The end product of the project is an Advisory report containing recommendations for achieving optimal responses in terms of relevance.

About the project group

We are a dedicated team of DDBL minor students with a shared passion for data. Our team includes individuals with unique skills and perspectives: Alexandru and Nikita, specializing in ICT and Software; Mariyana, with a background in ICT and Business; and Sam, whose focus lies in Commercial Economics. This diversity has proven to be our strength, allowing us to approach the project from multiple angles. Our varied expertise has been instrumental in navigating the complexities of our project.