Article Source

• Multi-Objective Recommender Systems

Multi-Objective Recommender Systems

• Workshop website: https://github.com/RecList/evalRS-CIKM-2022
• Open source project: http://reclist.io/

Abstract

Optimizing recommendations for a single objective, e.g., prediction accuracy, may be too limiting in certain applications. Instead, it is often important not only to consider multiple quality factors of recommendations, e.g., diversity, but to also take the perspectives of multiple stakeholders into account. In this talk, we will review different approaches from the literature that aim to consider multiple objectives in the recommendation process. Furthermore, we will outline open challenges and future directions in this area.

Introduction

Multi-objective recommender systems (MORS) aim to provide personalized recommendations to users while considering multiple conflicting objectives such as accuracy, diversity, novelty, serendipity, and user satisfaction. The research trends in MORS are as follows:

1. Deep Learning-Based Approaches: Deep learning-based approaches have been gaining popularity in MORS research. These approaches use various deep neural networks, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), to extract features from user-item interaction data.

2. Evolutionary Algorithms: Evolutionary algorithms, such as genetic algorithms (GAs) and particle swarm optimization (PSO), have been applied to MORS to optimize multiple objectives. These algorithms can efficiently explore the large search space of candidate solutions and find the optimal solutions.

3. Meta-Learning-Based Approaches: Meta-learning-based approaches have emerged as a new research trend in MORS. These approaches learn to learn the optimal recommendation strategies by leveraging past experiences and user feedback.

4. Hybrid Approaches: Hybrid approaches combine multiple recommendation algorithms, including collaborative filtering (CF), content-based filtering (CBF), and knowledge-based recommendation (KBR), to provide more accurate and diverse recommendations.

5. Interactive Recommendation Systems: Interactive recommendation systems allow users to provide feedback on the recommended items, which can be used to refine the recommendations in real-time.

6. Fairness and Diversity: MORS research has started to focus on addressing issues related to fairness and diversity in recommendation systems. These approaches aim to provide fair and diverse recommendations to users from different demographic groups.

In summary, the research trends in MORS are focused on developing more accurate, diverse, and personalized recommendation systems using deep learning, evolutionary algorithms, meta-learning, hybrid approaches, interactive systems, and addressing fairness and diversity issues.