Naslov
Building a conversation module for the Museum Assistant

Autori
Medved, Damir ; Perak, Benedikt

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
XXXIII. međunarodni znanstveni skup 33rd International Conference 16.–18. svibnja 2019. / 16th–18th May 2019 Rijeka, Hrvatska / Rijeka, Croatia ZNAČENJE U JEZIKU – OD INDIVIDUALNOGA DO KOLEKTIVNOGA MEANING IN LANGUAGE – FROM INDIVIDUAL TO COLLECTIVE / Matešić, Mihaela ; Nigoević, Magdalena - Rijeka : Srednja Europa ; Hrvatsko društvo za primijenjenu lingvistiku (HDPL), 2019, 65-65

ISBN
978-953-8281-00-6

Skup
XXXIII. međunarodni znanstveni skup ZNAČENJE U JEZIKU – OD INDIVIDUALNOGA DO KOLEKTIVNOGA 16. do 18. svibnja 2019. Rijeka (Hrvatska)

Mjesto i datum
Rijeka, Hrvatska, 16.05.2019. - 18.05.2019

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
conversation module, artificial inteligence, chatbot

Sažetak
The age of the digital assistants is rising, with different computer-assisted conversation modules emerging from various technology companies such as IBM, Google, Amazon, Apple, Microsoft, Facebook, etc. The conversation module sometimes also called a chatbot (Raj 2018, Goyal et al. 2018), is a machine learning system that allows human users to have conversational experience about some domain of knowledge. The process of creating a conversation module is comprised of several phases that include: defining the conversation domain, classification of intent, building the conversational database, chatbot customization and personality. The domain is defined as a set of interactional procedures and informational resources that a particular chatbot should be used for. The classification deals with the categorization and identification of user's intents to provide an appropriate response for the given domain. Intents have Training Phrases, which are examples of different syntactic-semantic constructions a user might elicit in a conversation about the given domain in order to express a particular intent for information that is retrieved from a chatbot, ie. the database of responses. In this paper, we will present a case study of a chatbot created and deployed for the purpose of eliciting a conversational experience for the Heritage Museum of Drenova (http://bezgranica.hr/heritage-museum-of- drenova/), with a goal to promote information about the history of Drenova in a museum setting. This paper presents the process of selecting the technology framework, workflow modelling, collecting information, machine learning, as well as analysing Natural Language Processing resources. Reference: Raj, S. (2018) Building Chatbots with Python. Apress, Berkeley, CA Goyal, P., Pandey, S., & Jain, K. (2018). Developing a Chatbot. In Deep Learning for Natural Language Processing (pp. 169-229). Apress, Berkeley, CA. https://www.facebook.com/muzejdrenove/

Izvorni jezik
Engleski

Znanstvena područja
Informacijske i komunikacijske znanosti, Filologija

POVEZANOST RADA