Deep Pairwise Hashing for Cold-start Recommendation

TL;DR

This paper introduces Deep Pairwise Hashing (DPH), a method that maps users and items into binary vectors for efficient recommendation, effectively addressing data sparsity and cold-start issues by combining content and interaction data.

Contribution

The paper proposes a novel deep hashing framework that unifies content and interaction data, using a pre-trained auto-encoder and pairwise ranking loss for improved recommendation performance.

Findings

Significantly improves recommendation efficiency using binary vectors.

Effectively alleviates cold-start and data sparsity problems.

Outperforms state-of-the-art methods on multiple datasets.

Abstract

Recommendation efficiency and data sparsity problems have been regarded as two challenges of improving performance for online recommendation. Most of the previous related work focus on improving recommendation accuracy instead of efficiency. In this paper, we propose a Deep Pairwise Hashing (DPH) to map users and items to binary vectors in Hamming space, where a user's preference for an item can be efficiently calculated by Hamming distance, which significantly improves the efficiency of online recommendation. To alleviate data sparsity and cold-start problems, the user-item interactive information and item content information are unified to learn effective representations of items and users. Specifically, we first pre-train robust item representation from item content data by a Denoising Auto-encoder instead of other deterministic deep learning frameworks; then we finetune the entire framework by adding a pairwise loss objective with discrete constraints; moreover, DPH aims to minimize a pairwise ranking loss that is consistent with the ultimate goal of recommendation. Finally, we adopt the alternating optimization method to optimize the proposed model with discrete constraints. Extensive experiments on three different datasets show that DPH can significantly advance the state-of-the-art frameworks regarding data sparsity and item cold-start recommendation.