Implementation of Artificial Intelligence Based Chatbot System with Long Term Memory

EasyChair Preprint 3202

Abstract

This paper mainly explores a specific deep learning method to build a conversational agent. Nowadays the popularity of chatbot systems is on rise as they attempt to get into daily life and achieve some commercial success. Previous approaches used simple keywords & pattern matching methodologies, answering in a static manner irrespective of previous conversions. As an improvement to this technology would be a system that will work with sequence to sequence framework. Our proposed model makes use of this framework. Given the previous sentence or sentences and the next sentence in a conversation, the model converses by predicting the next sentence. The distinctive feature of our model is that it can be trained end-to-end hence requires much fewer hand-crafted rules. This straightforward model can generate simple conversations given a large conversational training dataset.

Keyphrases: LSTM model, Recurrent Neural Network, deep learning, end to end memory, seq-to-seq model

@booklet{EasyChair:3202,
  author    = {Shruti Katkade and Urjita Kerkar and Pravin Bhilare and Manish Gupta and Payel Thakur},
  title     = {Implementation of Artificial Intelligence Based Chatbot System with Long Term Memory},
  howpublished = {EasyChair Preprint 3202},
  year      = {EasyChair, 2020}}