Thinking inside The Box

Abstract

As a step beyond traditional personalized recommendation, group recommendation is the task of suggesting items that can satisfy a group of users. In group recommendation, the core is to design preference aggregation functions to obtain a quality summary of all group members' preferences. Such user and group preferences are commonly represented as points in the vector space (i.e., embeddings), where multiple user embeddings are compressed into one to facilitate ranking for group-item pairs. However, the resulted group representations, as points, lack adequate flexibility and capacity to account for the multi-faceted user preferences. Also, the point embedding-based preference aggregation is a less faithful reflection of a group's decision-making process, where all users have to agree on a certain value in each embedding dimension instead of a negotiable interval. In this paper, we propose a novel representation of groups via the notion of hypercubes, which are subspaces containing innumerable points in the vector space. Specifically, we design the hypercube recommender (CubeRec) to adaptively learn group hypercubes from user embeddings with minimal information loss during preference aggregation, and to leverage a revamped distance metric to measure the affinity between group hypercubes and item points. Moreover, to counteract the long-standing issue of data sparsity in group recommendation, we make full use of the geometric expressiveness of hypercubes and innovatively incorporate self-supervision by intersecting two groups. Experiments on four real-world datasets have validated the superiority of CubeRec over state-of-the-art baselines. CCS CONCEPTS• Information systems → Recommender systems.

GroupMO: a memory-augmented meta-optimized model for group recommendation

“…Group recommendation aims to recommend items for a group of users. Existing group recommendation methods [5,23] are usually proposed to capture implicit group preference, and a common practice is to consider its members' preference learned from richer user-item interactions. The key is the accurate aggregation of group members' preferences, which can be performed via either score-level aggregation or feature-level aggregation.…”

“…Comparison of the performance using different methods on Mafengwo and CAMRa2011. -HHGR[10]: It is a neural group recommender that employs a self-supervised hypergraph learning framework to model group member interactions.• GroupIM[8]: GroupIM is a group recommendation method that adopts mutual information to overcome the sparse group-level interactions.• CubeRec[23]: CubeRec is a neural group recommender that employs a hypercube learning framework to model group preference.…”

Disentangled Self-Attention with Auto-Regressive Contrastive Learning for Neural Group Recommendation

“…The latter category, characterized by dense group-item interaction records, readily lends itself to the application of collaborative filtering techniques directly to group-item interactions [10,24]. The focus of this paper, however, is on the more pragmatically challenging domain of ephemeral group recommendations, a subject that has received considerable attention in the recent literature [12,20,21]. Within this domain, the primary challenge lies in the paucity of group-item interaction data [12,25].…”

“…Contemporary studies have extended these methodologies by integrating ancillary information sources such as social networks [33] and knowledge graphs [34], thereby enriching the learning paradigms for group representation. Innovations such as hypercube-based representations [21] and hypergraph convolutional neural networks [20] have been explored to encapsulate complex relational dynamics within and beyond group entities. Moreover, the group identification (GI) task, aiming to recommend the potential group to users, has recently attracted more attention from researchers.…”

“…The mutual information optimization mechanism [12], focusing on the cohesive embedding of user and group representations, has been developed to enhance preference learning. Moreover, SSL objectives exploiting the geometric properties of hypercubes [21] aim to mitigate the data sparsity issue by enriching group representations with more comprehensive informational content.…”