Preprint / Version 1

Evaluation of a Multimodal Approach to Chatbots

##article.authors##

  • Julian Weng

Keywords:

chatbots, multi-modal approach, human-based, machine-based

Abstract

Chatbots, software applications that can conduct conversations, often fall into one of two categories: “FAQ” style chatbots that can answer domain-specific questions and sometimes refer the user to a human operator, and “Conversational” chatbots that aim to be more general assistants or even just virtual companions. This paper will investigate the usage of a multi-modal approach that combines these two functions. It will evaluate two different methods, one human-based and one machine-based, for comparing the efficacy of this approach, and present final findings and their implications. This approach was found to perform better in domain-specific tasks with mixed results in conversational and general-knowledge tasks.

References or Bibliography

Britannica, T. Editors of Encyclopaedia (2020, May 7). Turing test. Encyclopedia Britannica. https://www.britannica.com/technology/Turing-test

Debecker, A. (2017, August 11). Chatbot ELIZA: Deconstructing Your Friendly Therapist. Ubisend. Retrieved September 12, 2021, from https://blog.ubisend.com/discover-chatbots/chatbot-eliza

Facebook AI. (2020, April 29). A state-of-the-art open source chatbot. Retrieved September 12, 2021, from https://ai.facebook.com/blog/state-of-the-art-open-source-chatbot

Garcia brustenga, G., Fuertes alpiste, M., & Molas castells, N. (2018, September 1). Briefing paper: Chatbots in education. Universitat Oberta de Catalunya (UOC). https://doi.org/10.7238/elc.chatbots.2018

Hart, B., & Risley, T. R. (1995). Meaningful Differences in the Everyday Experience of Young American Children (Early Education for All, Comp.). Brookes Publishing. http://www.strategiesforchildren.org/doc_research/05_MeaningfulDifferences.pdf

Rasa. (2021a). [Features of Rasa]. Retrieved September 12, 2021, from https://rasa.com/product/features/

Rasa. (2021b, July 12). Introduction to Rasa Action Servers. Retrieved September 12, 2021, from https://rasa.com/docs/action-server

Rasa. (2021c, July 12). Stories. Retrieved September 12, 2021, from https://rasa.com/docs/rasa/stories/

Sedoc, J. (2018, September 21). ChatEval (A Tool for Evaluating Chatbots). Github. Retrieved September 12, 2021, from https://github.com/chateval/archive/tree/master/eval_data

Stephanie. (2020, May 8). Chatbot FAQs – The Most Common Chatbot Questions Answered! Onlim. https://onlim.com/en/chatbot-faqs-the-most-common-chatbot-questions-answered/

Stillman, J. (2018, October 3). NEWSLETTERS SUBSCRIBE STARTUP LIFE A 70-Year Study of 70,000 Children Says This Is the Secret to Raising Successful Kids. Inc. Retrieved September 12, 2021, from https://www.inc.com/jessica-stillman/scientists-followed-thousands-of-kids-for-70-years-this-is-biggest-takeaway-for-parents.html

Streamlit. (2021). [Streamlit home page]. Retrieved September 12, 2021, from https://streamlit.io/

Zhang, Y., Sun, S., Galley, M., Chen, Y., Brockett, C., Gao, X., Gao, J., Liu, J., & Dolan, W. (2020). DIALOGPT : Large-Scale Generative Pre-training for Conversational Response Generation. ACL.

Downloads

Posted

10-26-2021