RecoXplainer: An Extensible Toolkit for Explainable Recommender Systems

A Quarter-day Tutorial at AAAI 2021

Overview

As recommender systems today play an important role in users' online experience and are involved in a wide range of decisions, multiple stakeholders demand explanations for the corresponding algorithmic predictions.

These demands, together with the benefits of explanations (e.g., trust, efficiency, and sometimes even persuasion), have triggered significant interest from researchers in academia and industry. Nonetheless, to the best of our knowledge, no comprehensive toolkit for explainable recommender systems is available to the community yet.

Researchers are frequently faced with the challenge of re-implementing prior algorithms when creating and evaluating new approaches. Aiming to address the resulting need, we present RecoXplainer, a software toolkit that includes several state-of-the-art explainability methods, and evaluation metrics.

Goal

The goal of the tutorial is to provide a review of eXplainable Recommender Systems (XRSs), and a practical hands-on session using a recently developed software toolkit called RecoXplainer.

RecoXplainer is a unified, extendable and easy to use Python toolkit that includes several explainability techniques that are useful for various groups of stakeholders. The goal of RecoXplainer is to equally help applied data scientists and researchers who focus their efforts on improving the field of XRSs.

The tutorial's audience will learn about model-based and post-hoc explanations of recommendations.

Syllabus

The tutorial will be organised in two parts:

  • Part I: An Introduction to Explainable Recommender Systems: The first part will provide the necessary background knowledge about Explainable AI and Explainable Recommender Systems, overviewing different categories and formats of explanations.
  • Part II: RecoXplainer Overview and Hands-on Session: The second part will consist of an overview of the RecoXplainer toolkit, and the implemented techniques such as model-based, post-hoc explanations, offline and online evaluation of explanations. After this introductory part, tutorial's attendees will be guided to the use of RecoXplainer, and to generate several examples of explainable recommendations.

Schedule

The tutorial is scheduled on Wednesday, February 3, 2021:

  • [8:30am-9:10am] Part I: An Introduction to Explainable Recommender Systems
  • [9:15am-10:00am] Part II: RecoXplainer Overview and Hands-on Session

Resources

Slides:

Code:

Presenters

Ludovik Coba is a post-doctoral researcher at the Free University of Bozen-Bolzano. He currently performs research in explainability and innovative algorithms for recommender systems. He publishes his results in venues like RecSys or IUI and journals like IT and Tourism, UMUAI or DKE.

Roberto Confalonieri is an Assistant Professor at the Free University of Bozen-Bolzano. Previously he was Senior Research Scientist in AI and XAI team lead at Alpha, the first European Moonshot projects company funded by Telefonica Research, working on explainable models of Artificial Intelligence in the health domain. A major focus in his current research is on Explainable AI, in particular, on how the integration of symbolic and non-symbolic reasoning approaches can convey human-understandable explanations of black-box models. He is an associate editor of Cognitive System Research published by Elsevier. He co-edited the book 'Concept Invention: Foundations, Implementation, Social Aspects and Applications' published by Springer in 2018. He organised scientifi events: he was the chair of an invited symposium at CogSci 2019, and co-chair of the series of International workshops of Methods for Interpretation of Industrial Event Logs (MIEL @IDEAL 2018, MIEL @BPM 2019), and Data meets Applied Ontologies (DAO @JOWO 2017, DAO-SI @JOWO 2019), and of C3GI 2018. He regularly serves in a number of programme committees as Senior PC and PC Member (IJCAI, AAAI, ECAI).

Markus Zanker is a professor at the Faculty of Computer Science of the Free University of Bozen-Bolzano, where he also served as vice dean for studies and director for study programmes. Before moving to Bolzano he was an associate professor at the Alpen-Adria-Universitaet Klagenfurt, Austria. His research focuses on knowledge-based information systems supporting decision making processes such as personalized information filtering and retrieval and product recommendation. Until 2013 Markus Zanker was also a co-founder and director of a recommendation service company for more than 10 years. He is an associate editor of Information Technology & Tourism (Springer). Besides organizing numerous workshops in the field, he was a program chair of the 4th ACM Conference on Recommender Systems in 2010 and he co-chaired the 9th ACM Conference on Recommender Systems in 2015.

Contact us

Please contact us for more information.