Counterfactual Explainable Recommendation

Tan, Juntao; Xu, Shuyuan; Ge, Yingqiang; Li, Yunqi; Xu, Chen; Zhang, Yongfeng

doi:10.1145/3459637.3482420

Cited by 92 publications

(63 citation statements)

References 51 publications

(59 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The former investigates how people perceive different styles of explanation [8,9,19], while the latter provides explanations by designing new explainable recommendation algorithms, to which our work is more related. There exist various types of explanation style, such as pre-defined templates [24,49,60], item features [20,54], ranked text [5,12,27], image visualizations [10], knowledge graph paths [1,18,55,56], and reasoning rules [7,46,63], but in this work we focus on generating natural language explanations. Previous works [6,13,26,57] mostly rely on RNN, e.g., LSTM [22] and GRU [14], or unpretrained Transformer [29], leaving the potentially more effective pre-trained models under-explored, which motivates this work.…”

Section: Explainable Recommendationmentioning

confidence: 99%

“…Meanwhile, explanations that justify why these recommendations are made are becoming more and more important, as they can help users make better and faster decisions, increase the system's ease of use, and gain their trust in the system [50,58]. There is a variety of explanation style, such as pre-defined templates [24,49,60], highlighted image regions [10] and automatically generated sentences [6,26,29]. The last type has gained increasing attention recently, mainly due to the availability of textual data on online commercial platforms, such as Amazon and Yelp, which encourage users to express their opinions by writing reviews (see Fig.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Personalized Prompt Learning for Explainable Recommendation

Li¹,

Zhang²,

Chen³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Providing user-understandable explanations to justify recommendations could help users better understand the recommended items, increase the system's ease of use, and gain users' trust. A typical approach to realize it is natural language generation. However, previous works mostly adopt recurrent neural networks to meet the ends, leaving the potentially more effective pre-trained Transformer models under-explored. In fact, user and item IDs, as important identifiers in recommender systems, are inherently in different semantic space as words that pre-trained models were already trained on. Thus, how to effectively fuse IDs into such models becomes a critical issue. Inspired by recent advancement in prompt learning, we come up with two solutions: find alternative words to represent IDs (called discrete prompt learning), and directly input ID vectors to a pre-trained model (termed continuous prompt learning). In the latter case, ID vectors are randomly initialized but the model is trained in advance on large corpora, so they are actually in different learning stages. To bridge the gap, we further propose two training strategies: sequential tuning and recommendation as regularization. Extensive experiments show that our continuous prompt learning approach equipped with the training strategies consistently outperforms strong baselines on three datasets of explainable recommendation.

show abstract

Section: Explainable Recommendationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Personalized Prompt Learning for Explainable Recommendation

Li¹,

Zhang²,

Chen³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Explainable AI has been an important topic in recommender systems [5,6,13,36,41,46,47], natural language processing [8,16,20] and computer vision [7,10,15,25,38]. To improve the transparency of deep neural networks, many explanation techniques have been proposed in recent years.…”

Section: Related Work 21 Explainability In Deep Learning and Aimentioning

confidence: 99%

“…They usually map the final prediction to the input space by gradient back-propagation or by linking hidden features to inputs via interpolation. Perturbation/casual-based methods [11,12,15,27,36,37,39] learn the feature importance through observing the change of predictions with respect to the input perturbation. The idea behind these methods are intuitive: determining which part of the inputs are important by either removing the least important information (i.e., pixels in image, words in text, nodes in graph) to keep the model prediction the same (factual reasoning) or removing the most important information to change the model prediction (counterfactual reasoning).…”

Section: Related Work 21 Explainability In Deep Learning and Aimentioning

confidence: 99%

Learning and Evaluating Graph Neural Network Explanations based on Counterfactual and Factual Reasoning

Tan,

Geng,

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Structural data well exists in Web applications, such as social networks in social media, citation networks in academic websites, and threads data in online forums. Due to the complex topology, it is difficult to process and make use of the rich information within such data. Graph Neural Networks (GNNs) have shown great advantages on learning representations for structural data. However, the non-transparency of the deep learning models makes it non-trivial to explain and interpret the predictions made by GNNs. Meanwhile, it is also a big challenge to evaluate the GNN explanations, since in many cases, the ground-truth explanations are unavailable.In this paper, we take insights of Counterfactual and Factual (CF 2 ) reasoning from causal inference theory, to solve both the learning and evaluation problems in explainable GNNs. For generating explanations, we propose a model-agnostic framework by formulating an optimization problem based on both of the two casual perspectives. This distinguishes CF 2 from previous explainable GNNs that only consider one of them. Another contribution of the work is the evaluation of GNN explanations. For quantitatively evaluating the generated explanations without the requirement of ground-truth, we design metrics based on Counterfactual and Factual reasoning to evaluate the necessity and sufficiency of the explanations. Experiments show that no matter ground-truth explanations are available or not, CF 2 generates better explanations than previous state-of-the-art methods on real-world datasets. Moreover, the statistic analysis justifies the correlation between the performance on ground-truth evaluation and our proposed metrics. CCS CONCEPTS• Computing methodologies → Machine learning; • Mathematics of computing → Computing most probable explanation; Graph algorithms.

show abstract

“…Personalized recommender systems (RS), which are extensively employed in e-commerce platforms, have been acknowledged for their capacity to deliver high-quality services that bridge the gap between products and customers [7,17,44,51]. Despite these huge advantages, several recent studies also raised concerns that RS may be vulnerable to algorithmic bias in several aspects, which may result in detrimental consequences for underrepresented or disadvantaged groups [19,29,43,59].…”

Section: Introductionmentioning

confidence: 99%

Toward Pareto Efficient Fairness-Utility Trade-off in Recommendation through Reinforcement Learning

Zhao

et al. 2022

Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining

Self Cite

View full text Add to dashboard Cite

The issue of fairness in recommendation is becoming increasingly essential as Recommender Systems (RS) touch and influence more and more people in their daily lives. In fairness-aware recommendation, most of the existing algorithmic approaches mainly aim at solving a constrained optimization problem by imposing a constraint on the level of fairness while optimizing the main recommendation objective, e.g., click through rate (CTR). While this alleviates the impact of unfair recommendations, the expected return of an approach may significantly compromise the recommendation accuracy due to the inherent trade-off between fairness and utility. This motivates us to deal with these conflicting objectives and explore the optimal trade-off between them in recommendation. One conspicuous approach is to seek a Pareto efficient/optimal solution to guarantee optimal compromises between utility and fairness. Moreover, considering the needs of real-world e-commerce platforms, it would be more desirable if we can generalize the whole Pareto Frontier, so that the decision-makers can specify any preference of one objective over another based on their current business needs. Therefore, in this work, we propose a fairness-aware recommendation framework using multi-objective reinforcement learning (MORL), called MoFIR (pronounced "more fair"), which is able to learn a single parametric representation for optimal recommendation policies over the space of all possible preferences. Specially, we modify traditional Deep Deterministic Policy Gradient (DDPG) by introducing conditioned network (CN) into it, which conditions the networks directly on these preferences and outputs Q-value-vectors. Experiments on several real-world recommendation datasets verify the superiority of our framework on both fairness metrics and recommendation measures when compared with all other baselines. We also extract the approximate Pareto Frontier on real-world datasets generated by MoFIR and compare to state-of-the-art fairness methods.

show abstract

Counterfactual Explainable Recommendation

Cited by 92 publications

References 51 publications

Personalized Prompt Learning for Explainable Recommendation

Personalized Prompt Learning for Explainable Recommendation

Learning and Evaluating Graph Neural Network Explanations based on Counterfactual and Factual Reasoning

Toward Pareto Efficient Fairness-Utility Trade-off in Recommendation through Reinforcement Learning

Contact Info

Product

Resources

About