Skip to content

Using LLM paired with Answer Set Programming to improve knowledge selection

License

Notifications You must be signed in to change notification settings

sajadabvi/Concierge-bot

 
 

Repository files navigation

Concierge-bot

Description

This repo is for the project Automated Interactive Domain-Specific Conversational Agents that Understand Human Dialogs.

We constructed an AutoConcierge for recommending local restaurants around UT Dallas by the preferences of the users.

It serves people by asking people questions of their preferences, and the users are also allowed to propose their own requirements. After collecting enough information, AutoConcierge will recommend the best matched restaurants for the user.

Requirement

  • s(CASP) query-driven ASP solver: available at s(CASP) GitLab page.

  • OpenAI Python package with secret key in environmental path: The Python package can be installed by pip install openai

How To Run

  1. Clone this repo.
  2. (optional) Prepare your own knowledge base follow the format of data/knowledge.pl and cover it.
  3. Run python main.py and start chat.

About

Using LLM paired with Answer Set Programming to improve knowledge selection

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Prolog 68.7%
  • Python 30.8%
  • Other 0.5%