Inspiration

Problem Statement - Expert service providers/ counselors are not aware of the cultural backgrounds of their immigrant client's in order to provide them mental health problem solutions. There is a need to resolve the cultural barriers of the two entities.

Enhance Cultural Understanding Through AI Suggestions and Live Q&A Personalized Client-Provider Matching Continuous Learning Through Feedbacks

What it does

Enhance Cultural Understanding Through AI Suggestions and Live Q&A Personalized Client-Provider Matching Continuous Learning Through Feedbacks

Web interface - Registers the immigrant/ patients information and the experts information. It then shortlists top 3 suitable Experts for the immigrant's counselling. The experts is provided with AI suggestions from our model and can chat with the AI model as well.

How we built it

We created LLM model on local machine. Provided it all the personal information of immigrants and added it in the prompt to the LLM model. It also

Challenges we ran into

Understanding accurate requirements of the project to provide a suitable solution. Limited multicultural data.

Accomplishments that we're proud of

Created a working AI Assistant using LLM model that runs on local machine with no additional cost. It provides suggestions curated to the immigrant's profile. Also built a functioning Web Interface to register the Immigrant and filters top 3 suitable Experts for the immigrant's counselling.

What we learned

Learned creating the AI assistant and Live Q&A chatbot using LLMs on local machine.

What's next for AI Powered Cultural Bridge

-> There is immense possibility to enhance the model performance by fine-tuning the existing LLM model by providing the feedbacks of the immigrant (client's) and experts. Enable reinforcement learning to continuously provide new feedbacks to the models. -> A RAG based LLM model can also be developed to provide additional and relevant database of information for different cultures. -> Use multilingual NLP to enable seamless communication across languages.

Built With

Share this project:

Updates