Conversational Agent for Mental Health

Oct 21, 2020 ml

This project was mentored by Prof. Ashutosh Modi where we attempted to improve a chat-bot that is in development by IIT Kanpur. We did the following -

  • Created a semi-automated self-help conversational agent that serves as a cost-effective mental health regulator.
  • Designed a retrieval based system where the input is compared to existing conversations in the database using deep learning.
  • Used neural network based architectures to find out the similarity between two pieces of texts to generate a response.
  • Explored Natural Language Generation (NLG) techniques based on RNN and transformer based architectures.
  • Pre-processed to convert the textual data into word embeddings to facilitate comparison among different texts.
ml
Data Scientist & Coordinator at Stamatics

I’m a senior undergraduate in Mathematics and Scientific Computing who works in Statistical Inference.