Neural Collaborative Filtering
∗
Xiangnan He
National University of
Singapore, Singapore
xiangnanhe@gmail.com
Lizi Liao
National University of
Singapore, Singapore
liaolizi.llz@gmail.com
Hanwang Zhang
Columbia University
USA
hanwangzhang@gmail.com
Liqiang Nie
Shandong University
China
nieliqiang@gmail.com
Xia Hu
Texas A&M University
USA
hu@cse.tamu.edu
Tat-Seng Chua
National University of
Singapore, Singapore
dcscts@nus.edu.sg
ABSTRACT
In recent years, deep neural networks have yielded immense
success on speech recognition, computer vision and natural
language processing. However, the exploration of deep neu-
ral networks on recommender systems has received relatively
less scrutiny. In this work, we strive to develop techniques
based on neural networks to tackle the key problem in rec-
ommendation — collaborative filtering — on the basis of
implicit feedback.
Although some recent work has employed deep learning
for recommendation, they primarily used it to model auxil-
iary information, such as textual descriptions of items and
acoustic features of musics. When it comes to model the key
factor in collaborative filtering — the interaction between
user and item features, they still resorted to matrix factor-
ization and applied an inner product on the latent features
of users and items.
By replacing the inner product with a neural architecture
that can learn an arbitrary function from data, we present
a general framework named NCF, short for Neural network-
based Collaborative Filtering. NCF is generic and can ex-
press and generalize matrix factorization under its frame-
work. To supercharge NCF modelling with non-linearities,
we propose to leverage a multi-layer perceptron to learn the
user–item interaction function. Extensive experiments on
two real-world datasets show significant improvements of our
proposed NCF framework over the state-of-the-art methods.
Empirical evidence shows that using deeper layers of neural
networks offers better recommendation performance.
Keywords
Collaborative Filtering, Neural Networks, Deep Learning,
Matrix Factorization, Implicit Feedback
∗
NExT research is supported by the National Research
Foundation, Prime Minister’s Office, Singapore under its
IRC@SG Funding Initiative.
c
2017 International World Wide Web Conference Committee
(IW3C2), published under Creative Commons CC BY 4.0 License.
WWW 2017, April 3–7, 2017, Perth, Australia.
ACM 978-1-4503-4913-0/17/04.
http://dx.doi.org/10.1145/3038912.3052569
.
1. INTRODUCTION
In the era of information explosion, recommender systems
play a pivotal role in alleviating information overload, hav-
ing been widely adopted by many online services, including
E-commerce, online news and social media sites. The key to
a personalized recommender system is in modelling users’
preference on items based on their past interactions (e.g.,
ratings and clicks), known as collaborative filtering [31, 46].
Among the various collaborative filtering techniques, matrix
factorization (MF) [14, 21] is the most popular one, which
projects users and items into a shared latent space, using
a vector of latent features to represent a user or an item.
Thereafter a user’s interaction on an item is modelled as the
inner product of their latent vectors.
Popularized by the Netflix Prize, MF has become the de
facto approach to latent factor model-based recommenda-
tion. Much research effort has been devoted to enhancing
MF, such as integrating it with neighbor-based models [21],
combining it with topic models of item content [38], and ex-
tending it to factorization machines [26] for a generic mod-
elling of features. Despite the effectiveness of MF for collab-
orative filtering, it is well-known that its performance can be
hindered by the simple choice of the interaction function —
inner product. For example, for the task of rating prediction
on explicit feedback, it is well known that the performance
of the MF model can be improved by incorporating user
and item bias terms into the interaction function
1
. While
it seems to be just a trivial tweak for the inner product
operator [14], it points to the positive effect of designing a
better, dedicated interaction function for modelling the la-
tent feature interactions between users and items. The inner
product, which simply combines the multiplication of latent
features linearly, may not be sufficient to capture the com-
plex structure of user interaction data.
This paper explores the use of deep neural networks for
learning the interaction function from data, rather than a
handcraft that has been done by many previous work [18,
21]. The neural network has been proven to be capable of
approximating any continuous function [17], and more re-
cently deep neural networks (DNNs) have been found to be
effective in several domains, ranging from computer vision,
speech recognition, to text processing [5, 10, 15, 47]. How-
ever, there is relatively little work on employing DNNs for
recommendation in contrast to the vast amount of literature
1
http://alex.smola.org/teaching/berkeley2012/slides/8_
Recommender.pdf
arXiv:1708.05031v2 [cs.IR] 26 Aug 2017