Interpretable Triplet Importance for Personalized Ranking

TL;DR

This paper introduces a novel interpretable importance measure for triplets in personalized ranking, using Shapley values and Monte Carlo approximation to improve model performance and transparency.

Contribution

It proposes the Triplet Shapley method, providing a trustworthy, interpretable importance score for triplets in ranking models, and demonstrates its effectiveness across multiple datasets.

Findings

Outperforms state-of-the-art methods in experiments

Provides stable and unbiased importance scores

Enhances model learning through importance-guided resampling

Abstract

Personalized item ranking has been a crucial component contributing to the performance of recommender systems. As a representative approach, pairwise ranking directly optimizes the ranking with user implicit feedback by constructing (\textit{user}, \textit{positive item}, \textit{negative item}) triplets. Several recent works have noticed that treating all triplets equally may hardly achieve the best effects. They assign different importance scores to negative items, user-item pairs, or triplets, respectively. However, almost all the generated importance scores are groundless and hard to interpret, thus far from trustworthy and transparent. To tackle these, we propose the \textit{Triplet Shapley} -- a Shapely value-based method to measure the triplet importance in an interpretable manner. Due to the huge number of triplets, we transform the original Shapley value calculation to the Monte Carlo (MC) approximation, where the guarantee for the approximation unbiasedness is also provided. To stabilize the MC approximation, we adopt a control covariates-based method. Finally, we utilize the triplet Shapley value to guide the resampling of important triplets for benefiting the model learning. Extensive experiments are conducted on six public datasets involving classical matrix factorization- and graph neural network-based recommendation models. Empirical results and subsequent analysis show that our model consistently outperforms the state-of-the-art methods.