Quaternion Collaborative Filtering for Recommendation
This work addresses recommendation accuracy for users by introducing a novel representation learning approach, though it appears incremental as it builds on existing collaborative filtering with a new mathematical framework.
The paper tackled the problem of improving recommendation systems by proposing Quaternion Collaborative Filtering (QCF), a method using quaternion algebra to capture intricate user-item interactions, and demonstrated its effectiveness by outperforming strong neural baselines on six real-world datasets.
This paper proposes Quaternion Collaborative Filtering (QCF), a novel representation learning method for recommendation. Our proposed QCF relies on and exploits computation with Quaternion algebra, benefiting from the expressiveness and rich representation learning capability of Hamilton products. Quaternion representations, based on hypercomplex numbers, enable rich inter-latent dependencies between imaginary components. This encourages intricate relations to be captured when learning user-item interactions, serving as a strong inductive bias as compared with the real-space inner product. All in all, we conduct extensive experiments on six real-world datasets, demonstrating the effectiveness of Quaternion algebra in recommender systems. The results exhibit that QCF outperforms a wide spectrum of strong neural baselines on all datasets. Ablative experiments confirm the effectiveness of Hamilton-based composition over multi-embedding composition in real space.