Abhijit Nargund


2022

pdf
PAR: Persona Aware Response in Conversational Systems
Abhijit Nargund | Sandeep Pandey | Jina Ham
Proceedings of the 19th International Conference on Natural Language Processing (ICON)

To make the Human Computer Interaction more user friendly and persona aligned, detection of user persona is of utmost significance. Towards achieving this objective, we describe a novel approach to select the persona of a user from pre-determine list of personas and utilize it to generate personalized responses. This is achieved in two steps. Firstly, closest matching persona is detected from a set of pre-determined persona for the user. The second step involves the use of a fine-tuned natural language generation (NLG) model to generate persona compliant responses. Through experiments, we demonstrate that the proposed architecture generates better responses than current approaches by using a detected persona. Experimental evaluation on the PersonaChat dataset has demonstrated notable performance in terms of perplexity and F1-score.

2021

pdf
MRE : Multi Relationship Extractor for Persona based Empathetic Conversational Model
Bharatram Natarajan | Abhijit Nargund
Proceedings of the 18th International Conference on Natural Language Processing (ICON)

Artificial intelligence(AI) has come a long way in aiding the user requirements in many fields and domains. However, the current AI systems do not generate human- like response for user query. Research in these areas have started gaining traction recently with explorations on persona or empathy based response selection. But the combination of both the parameters in an open domain haven’t been explored in detail by the research community. The current work highlights the effect of persona on empathetic response. This research paper concentrates on improving the response selection model for PEC dataset, containing both persona information and empathetic response. This is achieved using an enhanced multi relationship extractor and phrase based information for the selection of response.