Goal-oriented conversational agents – a proposed approach for practical domains

Journal Title: Revista Romana de Interactiune Om-Calculator - Year 2018, Vol 11, Issue 4

Abstract

Conversational agents are nowadays desired in many fields, mainly for task automation, but also for entertaining, therapy or other purposes. This report will introduce the main Dialogue Systems categories, along with some of the most successful frameworks and algorithms, reported as state-of-the-art over the last few years, with a focus on Goal Oriented (GO) chatbots. Due to the domain specificity, large accomplishments are the most difficult to attain, conversational agents being able, in general, to encompass some of the most problematic natural language processing tasks such as intent classification, information extraction, text generation. Similar to other subfields of artificial intelligence and machine learning, in this field we find the same vital role of large quantities of annotated data. Some of the research discussed here with significant input towards the industry tries to maximize knowledge gain over decreasing the number of annotated instances to be fed to the system as training data. I further carry forth our experiments on a telecom dataset, along with results, conclusions and future work.

Authors and Affiliations

Stefania Budulan

Keywords

Related Articles

Recommendation Technique for the “Cold-Start” Problem

Usually, users that benefit from a Recommender System outputs only get a list of items that the system assumes the best match their needs, without having any clue regarding how the system managed to figure out what they...

Methodology for Identification and Evaluation of Web Application Performance Oriented Usability Issues

This paper aims to illustrate a methodology for identifying and assessing a set of performance issues encountered in a particular web application, with impact on the usability level. Throughout this methodology, several...

Interactions in Smart Environments and the Importance of Modelling

One challenge in software engineering is the development of smart environments that help users to intuitively accomplish their tasks. The ideal smart environment dynamically manages a diverse collection of devices, is ac...

ReadME – A System for Automated Writing Evaluation and Learning Management

Writing is an essential skill in the digital era, whether we are referring to essays, books, discourses, emails, or newsletters. Writing transcends individual words or sentences, and it considers the structuring of ideas...

Tehnici de vizualizare a imaginilor satelitare prin servicii geospaţiale

Analiza unui volum mare de date, extragerea informaţiilor de interes, expunerea şi vizualizarea interactivă a rezultatelor reprezintă procesele de bază în modelarea fenomenelor naturale, bazate pe prelucrarea imaginilor...

Download PDF file
  • EP ID EP672049
  • DOI -
  • Views 114
  • Downloads 0

How To Cite

Stefania Budulan (2018). Goal-oriented conversational agents – a proposed approach for practical domains. Revista Romana de Interactiune Om-Calculator, 11(4), 266-280. https://europub.co.uk/articles/-A-672049