Talking with chatbots and question answering interfaces is nowadays common to most people using some kind of digital device. Artificial intelligence-based solutions to create chatbots are available on the market and are provided by all the big players like IBM, Microsoft, Google, and Amazon. Chatbots can differ for the type of corpus they rely upon, the type of questions they are suited for, the algorithms, and the ability to learn and expand their knowledge base. Moreover they differ for the language, with English chatbots being on average those best performing. In this paper we present ConversIAmo, a chatbot prototype designed to answer questions in Italian about introductory concepts on Artificial Intelligence. The natural language processing pipeline is built upon the IBM Watson suite and implements further modules for improving the answer selection and ordering, and to expand the knowledge base. Results are encouraging, especially with regard to the ordering of the correct answers. The paper describes the question answering dataset, the first in this domain, and outlines the dialog flow and the answering process.

Talking in Italian About AI with a Chatbot: A Prototype of a Question-Answering Agent

Coccoli M.;Torre I.;Vercelli G.
2020-01-01

Abstract

Talking with chatbots and question answering interfaces is nowadays common to most people using some kind of digital device. Artificial intelligence-based solutions to create chatbots are available on the market and are provided by all the big players like IBM, Microsoft, Google, and Amazon. Chatbots can differ for the type of corpus they rely upon, the type of questions they are suited for, the algorithms, and the ability to learn and expand their knowledge base. Moreover they differ for the language, with English chatbots being on average those best performing. In this paper we present ConversIAmo, a chatbot prototype designed to answer questions in Italian about introductory concepts on Artificial Intelligence. The natural language processing pipeline is built upon the IBM Watson suite and implements further modules for improving the answer selection and ordering, and to expand the knowledge base. Results are encouraging, especially with regard to the ordering of the correct answers. The paper describes the question answering dataset, the first in this domain, and outlines the dialog flow and the answering process.
2020
978-3-030-62364-7
978-3-030-62365-4
File in questo prodotto:
File Dimensione Formato  
my-paper.pdf

accesso chiuso

Descrizione: Contributo in volume
Tipologia: Documento in versione editoriale
Dimensione 2.24 MB
Formato Adobe PDF
2.24 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/1033115
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact