• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.3.2023.tde-05022024-111423
Document
Author
Full name
Douglas Luan de Souza
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Paulo, 2023
Supervisor
Committee
Cozman, Fabio Gagliardi (President)
Veiga, Allan Koch
Peres, Sarajane Marques
Title in Portuguese
Algoritmos para explicação responsável de recomendações.
Keywords in Portuguese
Aprendizado computacional
Interpretabilidade e explicabilidade
Sistemas de recomendação
Abstract in Portuguese
Sistemas de Recomendação tem sido cada vez mais presentes em produtos digitais como e-commerce e redes sociais. Um das razões para o seu sucesso é a sua crescente capacidade de predizer quais itens irão agradar seus usuários. No entanto, há cenários onde as recomendações podem não estar sendo geradas segundo os interesses dos usuários. Isto pode acontecer porque o sistema incentiva que o usuário passe horas assistindo vídeos curtos, ou porque ele recomenda algum produto que não tem a relação custo-benefício que seria melhor para o cliente. Nestes casos, seria desejável que os usuários pudessem entender as razões que levaram à recomendação. Além disso, usuários se beneficiariam ao compreender quais são as desvantagens de seguir a recomendação dada. Neste trabalho, propomos métodos para geração de razões a favor e razões contra uma determinada recomendação, baseados na teoria de Snedegar para raciocínio prático. Mostramos que estes métodos são aplicáveis na prática em um contexto educacional, usando um sistema de recomendação de disciplinas.
Title in English
Algorithms for responsible explanation of recommendations.
Keywords in English
Explainable Artificial Intelligence (XAI)
Machine learning
Recommender systems
Abstract in English
Recommender Systems have been increasingly deployed as part of digital products such as e-commerce and social networks. Part of the reason for their success is their ability to predict which items the user will like. However, there are scenarios where their recommendation may not be in the best interest of the user. It could be because the recommendations reinforce a habit the users do not want, such as spending hours watching short videos, or because it suggests a product that does offer the best cost-benefit ratio. In these cases, it would be helpful for the users to understand the reasons behind the recommendation. Furthermore, it would be in the interest of the user to know what are the drawbacks of taking the recommendation. In this work, we propose methods for generating reasons for and reasons against a given recommendation, based on Snedegar theory of practical reasoning. We demonstrate that these methods are feasible in the context of education through a course recommender.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2024-02-06
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.