Designing and Evaluating Explanations for Recommender Systems

By Nava Tintarev, Judith Masthoff. 2011

378 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yongfeng Zhang
2. Xu Chen
3. Xiangnan He
4. Tat-seng Chua
5. Xiang Wang
6. Qingyao Ai
7. Liqiang Nie
8. W. Bruce Croft

In Information Systems

1. Yongfeng Zhang
2. Xu Chen
3. Xiangnan He
4. Tat-seng Chua
5. Xiang Wang
6. Liqiang Nie
7. Qingyao Ai
8. Fuli Feng

Citations in the corpus

• Knowledge-aware Autoencoders for Explainable Recommender Systems (2018, methodology): “In order to overcome this issue, new methods have been developed in order to generate an explainable recommendation ([18] provides an overview of the most successful approaches proposed over the years) such asMoviExplain [16], which exploits movies metadata to justify its recommendation”
• How should I explain? A comparison of different explanation types for recommender systems (2014, background): “In our work, however, we aim at evaluating different explanation types in a comprehensive manner and consider the desired effects and quality dimensions efficiency, effectiveness, persuasiveness, perceived transparency, and satisfaction [Tintarev and Masthoff, 2011] in parallel.”
• Inspectability and control in social recommenders (2012, background): “Previous studies have attempted to explain recommendations by showing the link between the recommendations and the “nearest neighbors” on which the they are based [19, 33, 45].”
• Explicit factor models for explainable recommendation based on phrase-level sentiment analysis (2014, background): “Researchers have shown that providing appropriate explanations could improve user acceptance of the recommended items [10, 37], as well as benefit user experience in various other aspects, including system transparency, user trust, effectiveness, efficiency, satisfaction and scrutability [38, 3, 2].”
• Do social explanations work?: studying and modeling the effects of social explanations in recommender systems (2013, background): “part of the information presented about a recommendation, and studies show that explanations play an important role in helping a user evaluate a recommendation [23,26].”
• Learning over Knowledge-Base Embeddings for Recommendation (2018, background): “,r) and (e h,r) in the training data. If we further compute the partial derivative of ‘(e h,et,r) with respect to x = et trans(e h,r), we have ¶‘(e h,et,r) ¶x = #(e h,e t,r)s( x) k #(e h,r)P (e )s(x) (7) When the training process has converged, the partial derivative of ‘(e h,et,r) should be 0, and ”
• Evaluating the effectiveness of explanations for recommender systems (2012, background): “recommendations (Tintarev and Masthoff 2010).”
• The FacT: Taming Latent Factor Models for Explainability with Factorization Trees (2019, background): “ngs. Positive user feedback further validates the value of our proposed solution. 2 RELATED WORKS Various studies show that accurate explanation improves transparency of automated recommender systems [26, 30], helps users make more informed decisions [9, 29], and thus increases recommendation effect”
• A generalized taxonomy of explanations styles for traditional and social recommender systems (2012, methodology): “1.1 Motivation and definition of a generalized taxonomy of explanations The motivation of our survey is based on the fact that up until now, the type of an explanation style was considered in accordance to the recommender algorithm that employed it (Tintarev and Masthoff 2011, 2007).”
• Social Collaborative Viewpoint Regression with Explainable Recommendations (2017, background): “Explainable recommendations are known to improve transparency, user trust, effectiveness and scrutability [44].”
• Recommender systems and their ethical challenges (2020, background): “It may be difficult to provide good factual explanations in practice also for computational reasons (the required computation to generate a good explanation may be too complex), and because they may have distorting effects on the accuracy of the recommendations Tintarev and Masthoff (2011).”
• Recommender systems based on user reviews: the state of the art (2015, background): “User evaluation could be conducted on such a review-based explanation interface to verify its effectiveness, trustworthiness, and persuasiveness from users’ perspectives (Tintarev and Masthoff 2011).”
• A systematic review and taxonomy of explanations in decision support and recommender systems (2017, background): “by helping users make better or more informed decisions (Tintarev and Masthoff 2011).”
• Explainable Entity-based Recommendations with Knowledge Graphs (2017, background): “ Improving the accuracy of predictions in recommender systems is animportantresearchtopic.Anequallyimportanttaskisexplaining the predictions to the user. Explanations may serve many di˛erent purposes [11]. They can show how the system works (transparency) or help users make an informed choice (e˛ect”