Discrete Factorization Machines for Fast Feature-based Recommendation

Abstract

User and item features of side information are crucial for accurate recommendation. However, the large number of feature dimensions, e.g., usually larger than 10 7 , results in expensive storage and computational cost. This prohibits fast recommendation especially on mobile applications where the computational resource is very limited. In this paper, we develop a generic feature-based recommendation model, called Discrete Factorization Machine (DFM), for fast and accurate recommendation. DFM binarizes the real-valued model parameters (e.g., float32) of every feature embedding into binary codes (e.g., boolean), and thus supports efficient storage and fast user-item score computation. To avoid the severe quantization loss of the binarization, we propose a convergent updating rule that resolves the challenging discrete optimization of DFM. Through extensive experiments on two realworld datasets, we show that 1) DFM consistently outperforms state-of-the-art binarized recommendation models, and 2) DFM shows very competitive performance compared to its real-valued version (FM), demonstrating the minimized quantization loss.

Compositional Coding for Collaborative Filtering

“…To circumvent the above issues, direct discrete optimization method has been proposed in Discrete Collaborative Filtering (DCF) [30] and its extension [12,16,31]. More formally, it learns the binary codes by optimizing the following objective function:…”

“…To derive compact yet informative binary codes, the balanced and de-correlated constraints were further imposed [30]. In order to incorporate content information from users and items, content-aware matrix factorization and factorization machine with binary constraints was further proposed [12,16]. For dealing with social information, a discrete social recommendation model was proposed in [15].…”

“…Nevertheless, this solution suffers from a large quantization loss [30] and thus fails to preserve the original data geometry (user-item relevance and user-user relationship) in the continuous real-valued vector space. As accuracy is arguably the most important evaluation metric for recommender systems, researchers put lots of efforts to reduce the quantization loss by direct discrete optimization [12,16,30]. In spite of the advantages of this improved optimization method, compared to real-valued vectors, CF with binary codes naturally suffers from low accuracy due to limited representation capability in each bit, which impedes it from modeling complex relationship between users and items.…”

Candidate Generation with Binary Codes for Large-Scale Top-N Recommendation

“…they depend on the specific structure of an inner-product space), and thus are hard to generalize when trying to accelerate other models. Another line of work seeks to directly learn binary codes to estimate user-item interactions, and builds hash tables to accelerate retrieval time [24,27,44,[46][47][48]. While using binary codes can significantly reduce query time to constant or sublinear complexity, the accuracy of such models is still inferior to conventional (i.e., real-valued) models, as such models are highly constrained, and may lack sufficient flexibility when aiming to precisely rank the Top-N items.…”

Discrete Trust-aware Matrix Factorization for Fast Recommendation

“…Recently, the discrete matrix factorization technique captures attention due to its recommendation efficiency. It adopts hash technology [Zhang et al, 2016;Zhang et al, 2017;Liu et al, 2018] to map latent features of users and items to a joint hamming space and transforms the items recommendation task into a similarity search problem, which greatly improves the recommendation efficiency. Specifically, binary code can be executed with bit operation.…”