Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Relevance meets Diversity: A User-Centric Framework for Knowledge Exploration through Recommendations

Erica Coppolillo, Giuseppe Manco, Aristides Gionis

TL;DR

This work tackles the trade-off between relevance and diversity in recommender systems by reframing it as a knowledge-exploration problem guided by user behavior. It introduces a probabilistic user model that captures quitting behavior (weariness) and a diversity objective, plus a copula-based strategy that fuses item relevance with diversity gains to produce diverse, relevant recommendations without requiring training. The approach defines two diversity notions (coverage and distance) and demonstrates, across five public datasets, that the proposed explore variants achieve higher diversity while maintaining competitive relevance compared to strong baselines. The findings suggest practical benefits for knowledge acquisition and user engagement, with robust performance and scalable computation, and point to future work on richer user actions and additional objectives like serendipity and fairness.

Abstract

Providing recommendations that are both relevant and diverse is a key consideration of modern recommender systems. Optimizing both of these measures presents a fundamental trade-off, as higher diversity typically comes at the cost of relevance, resulting in lower user engagement. Existing recommendation algorithms try to resolve this trade-off by combining the two measures, relevance and diversity, into one aim and then seeking recommendations that optimize the combined objective, for a given number of items to recommend. Traditional approaches, however, do not consider the user interaction with the recommended items. In this paper, we put the user at the central stage, and build on the interplay between relevance, diversity, and user behavior. In contrast to applications where the goal is solely to maximize engagement, we focus on scenarios aiming at maximizing the total amount of knowledge encountered by the user. We use diversity as a surrogate of the amount of knowledge obtained by the user while interacting with the system, and we seek to maximize diversity. We propose a probabilistic user-behavior model in which users keep interacting with the recommender system as long as they receive relevant recommendations, but they may stop if the relevance of the recommended items drops. Thus, for a recommender system to achieve a high-diversity measure, it will need to produce recommendations that are both relevant and diverse. Finally, we propose a novel recommendation strategy that combines relevance and diversity by a copula function. We conduct an extensive evaluation of the proposed methodology over multiple datasets, and we show that our strategy outperforms several state-of-the-art competitors. Our implementation is publicly available at https://github.com/EricaCoppolillo/EXPLORE.

Relevance meets Diversity: A User-Centric Framework for Knowledge Exploration through Recommendations

TL;DR

This work tackles the trade-off between relevance and diversity in recommender systems by reframing it as a knowledge-exploration problem guided by user behavior. It introduces a probabilistic user model that captures quitting behavior (weariness) and a diversity objective, plus a copula-based strategy that fuses item relevance with diversity gains to produce diverse, relevant recommendations without requiring training. The approach defines two diversity notions (coverage and distance) and demonstrates, across five public datasets, that the proposed explore variants achieve higher diversity while maintaining competitive relevance compared to strong baselines. The findings suggest practical benefits for knowledge acquisition and user engagement, with robust performance and scalable computation, and point to future work on richer user actions and additional objectives like serendipity and fairness.

Abstract

Providing recommendations that are both relevant and diverse is a key consideration of modern recommender systems. Optimizing both of these measures presents a fundamental trade-off, as higher diversity typically comes at the cost of relevance, resulting in lower user engagement. Existing recommendation algorithms try to resolve this trade-off by combining the two measures, relevance and diversity, into one aim and then seeking recommendations that optimize the combined objective, for a given number of items to recommend. Traditional approaches, however, do not consider the user interaction with the recommended items. In this paper, we put the user at the central stage, and build on the interplay between relevance, diversity, and user behavior. In contrast to applications where the goal is solely to maximize engagement, we focus on scenarios aiming at maximizing the total amount of knowledge encountered by the user. We use diversity as a surrogate of the amount of knowledge obtained by the user while interacting with the system, and we seek to maximize diversity. We propose a probabilistic user-behavior model in which users keep interacting with the recommender system as long as they receive relevant recommendations, but they may stop if the relevance of the recommended items drops. Thus, for a recommender system to achieve a high-diversity measure, it will need to produce recommendations that are both relevant and diverse. Finally, we propose a novel recommendation strategy that combines relevance and diversity by a copula function. We conduct an extensive evaluation of the proposed methodology over multiple datasets, and we show that our strategy outperforms several state-of-the-art competitors. Our implementation is publicly available at https://github.com/EricaCoppolillo/EXPLORE.
Paper Structure (11 sections, 16 equations, 7 figures, 7 tables, 2 algorithms)

This paper contains 11 sections, 16 equations, 7 figures, 7 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related work
  3. User model and problem formulation
  4. Recommendation strategy
  5. Experiments
  6. Datasets
  7. Competing recommendation strategies
  8. Experimental setting
  9. Results
  10. Conclusion and future work
  11. Appendix

Figures (7)

  • Figure 1: The knowledge-exploration process, illustrating the interplay among relevance, diversity, and user behavior.
  • Figure 2: Illustration of the impact of different recommendation strategies. White points are recommended items, blue circles indicate information coverage. (a) High relevance, low diversity (e.g., all about 'technology'); (b) High diversity, likely non-relevant (e.g., 'technology,' 'religion,' 'lifestyle'); (c) Optimal balance: Relevant and diverse, keeping user engaged (e.g., 'technology,' 'science,' 'engineering.').
  • Figure 3: Trade-off between $\mathrm{div}\xspace_{\!_C}$ and Recall@10 across all the datasets considered. The X-axis represents recommendation quality, while the Y-axis indicates the diversity score.
  • Figure 4: Trade-off between either $\mathrm{div}\xspace_{\!_D}$ or $\mathrm{div}\xspace_{\!_C}$ and either Recall@10, HR@10, orPrecision@10 over Movielens-1M, Coat and KuaiRec-2.0. The X-axis shows the recommendation quality while the Y-axis represents the diversity score.
  • Figure 5: Trade-off between either $\mathrm{div}\xspace_{\!_D}$ or $\mathrm{div}\xspace_{\!_C}$ and either Recall@10, HR@10, orPrecision@10 over Netflix-Prize and Yahoo-R2. The X-axis shows the recommendation quality while the Y-axis represents the diversity score.
  • ...and 2 more figures