PoliSe: Reinforcing Politeness Using User Sentiment for Customer Care Response Generation

4712

Authors

nan

Abstract

AbstractThe interaction between a consumer and the customer service representative greatly contributes to the overall customer experience. Therefore, to ensure customers’ comfort and retention, it is important that customer service agents and chatbots connect with users on social, cordial, and empathetic planes. In the current work, we automatically identify the sentiment of the user and transform the neutral responses into polite responses conforming to the sentiment and the conversational history. Our technique is basically a reinforced multi-task network- the primary task being ‘polite response generation’ and the secondary task being ‘sentiment analysis’- that uses a Transformer based encoder-decoder. We use sentiment annotated conversations from Twitter as the training data. The detailed evaluation shows that our proposed approach attains superior performance compared to the baseline models.

Keywords

nan

Full Content

Proceedings of the 29th International Conference on Computational Linguistics, pages 6165–6175 October 12–17, 2022. 6165 PoliSe: Reinforcing Politeness using User Sentiment for Customer Care Response Generation Mauajama Firdaus1∗†, Asif Ekbal2 and Pushpak Bhattacharyya3 1 Department of Computing Science, University of Alberta, Canada 2Department of Computer Science and Engineering, Indian Institute of Technology Patna, India 3Department of Computer Science and Engineering, Indian Institute of Technology Bombay, India 1mauzama.03@gmail.com; 2asif@iitp.ac.in; 3pushpakbh@gmail.com Abstract The interaction between a consumer and the customer service representative greatly con- tributes to the overall customer experience. Therefore, to ensure customers’ comfort and retention, it is important that customer service agents and chatbots connect with users on so- cial, cordial and empathetic planes. In the cur- rent work, we automatically identify the sen- timent of the user and transform the neutral responses into polite responses conforming to the sentiment and the conversational history. Our technique is basically a reinforced multi- task network- the primary task being ’polite response generation’ and the secondary task being ’sentiment analysis’- that uses a Trans- former based encoder-decoder. We use sen- timent annotated conversations from Twitter as the training data. The detailed evaluation shows that our proposed approach attains su- perior performance compared to the baseline models. 1 Introduction Human-machine interactions have increased rapidly assisting humans in their everyday lives. With the growth in Artificial Intelligence (AI) and Natural Language Processing (NLP), chatbots and personal assistants, such as Microsoft’s Cortana, Apple’s Siri, etc., have predominantly become a part of our daily lives. Thus, research in recent years has been on modulating biases, styles, and control in text generation to enhance these interactions. Customer care is an essential tool used by com- panies to provide guidance, and assistance and in building stable customer relations. The ease of ac- cess, ease of following up, and immediacy of social media has made it a strong platform for companies ∗The work was done while the author was in IIT Patna †This work is licensed under a Creative Commons At- tribution 4.0 International Licence. Licence details: http: //creativecommons.org/licenses/by/4.0/ and applications to interact with their customers. In this regard, conversational agents such as shopping agents, customer service agents, and personal assis- tants have become extremely famous for handling the various queries of a customer and providing suitable responses to them. Assisting the customer through social media channels is attaining high popularity. The main reason behind this is the fact that the important elements of social media are its immediacy, trans- parency, ease of following up, and providing a human-like feel to the company/brand. In this plat- form, we see the usage of polite and emphatic lan- guage, which is the center of our current study. For the growth of any company or application, customer care agents must be cordial and amicable to the customer. Thus along with handling queries, the agents need to provide customer satisfaction by greeting, empathizing, appreciating feedback, apologizing at the right time, and thus building a strong relationship with the customer. Recently, incorporating politeness in responses …

Content truncated to 500 words

Labeled Text

AbstractThe interaction between a consumer and the customer service representative greatly contributes to the overall customer experience. Therefore, to ensure customers’ comfort and retention, it is important that customer service agents and chatbots connect with users on social, cordial, and empathetic planes. In the current work, we automatically identify the sentiment of the user and transform the neutral responses into polite responses conforming to the sentiment and the conversational history. Our technique is basically a reinforced multi-task network- the primary task being ‘polite response generation’ and the secondary task being ‘sentiment analysis’- that uses a Transformer based encoder-decoder. We use sentiment annotated conversations from Twitter as the training data. The detailed evaluation shows that our proposed approach attains superior performance compared to the baseline models.

Publication Details

Paper ID:
4712
Year:
2022
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan