arXiv subjects

Zaiqiao Meng

Publications and source records attributed to Zaiqiao Meng.

57 records · Page 4Linked to original sources

arXiv2019

Variational Bayesian Context-aware Representation for Grocery Recommendation

Grocery recommendation is an important recommendation use-case, which aims to predict which items a user might choose to buy in the future, based on their shopping history. However, existing methods only represent each user and item by single deterministic points in a low-dimensional continuous space. In addition, most of these methods are trained by maximizing the co-occurrence likelihood with a simple Skip-gram-based formulation, which limits the expressive ability of their embeddings and the resulting recommendation performance. In this paper, we propose the Variational Bayesian Context-Aware Representation (VBCAR) model for grocery recommendation, which is a novel variational Bayesian model that learns the user and item latent vectors by leveraging basket context information from past user-item interactions. We train our VBCAR model based on the Bayesian Skip-gram framework coupled with the amortized variational inference so that it can learn more expressive latent representations that integrate both the non-linearity and Bayesian behaviour. Experiments conducted on a large real-world grocery recommendation dataset show that our proposed VBCAR model can significantly outperform existing state-of-the-art grocery recommendation methods.

cs.IR↗

arXiv2019

Neural Variational Hybrid Collaborative Filtering

Collaborative Filtering (CF) is one of the most used methods for Recommender System. Because of the Bayesian nature and nonlinearity, deep generative models, e.g. Variational Autoencoder (VAE), have been applied into CF task, and have achieved great performance. However, most VAE-based methods suffer from matrix sparsity and consider the prior of users' latent factors to be the same, which leads to poor latent representations of users and items. Additionally, most existing methods model latent factors of users only and but not items, which makes them not be able to recommend items to a new user. To tackle these problems, we propose a Neural Variational Hybrid Collaborative Filtering, NVHCF. Specifically, we consider both the generative processes of users and items, and the prior of latent factors of users and items to be side informationspecific, which enables our model to alleviate matrix sparsity and learn better latent representations of users and items. For inference purpose, we derived a Stochastic Gradient Variational Bayes (SGVB) algorithm to analytically approximate the intractable distributions of latent factors of users and items. Experiments conducted on two large datasets have showed our methods significantly outperform the state-of-the-art CF methods, including the VAE-based methods.

cs.IR↗

arXiv2016

Diversified Top-k Similarity Search in Large Attributed Networks

Given a large network and a query node, finding its top-k similar nodes is a primitive operation in many graph-based applications. Recently enhancing search results with diversification have received much attention. In this paper, we explore an novel problem of searching for top-k diversified similar nodes in attributed networks, with the motivation that modeling diversification in an attributed network should consider both the emergence of network links and the attribute features of nodes such as user profile information. We formulate this practical problem as two optimization problems: the Attributed Coverage Diversification (ACD) problem and the r-Dissimilar Attributed Coverage Diversification (r-DACD) problem. Based on the submodularity and the monotonicity of ACD, we propose an efficient greedy algorithm achieving a tight approximation guarantee of 1-1/e. Unlike the expension based methods only considering nodes' neighborhood, ACD generalize the definition of diversification to nodes' own features. To capture diversification in topological structure of networks, the r-DACD problem introduce a dissimilarity constraint. We refer to this problem as the Dissimilarity Constrained Non-monotone Submodular Maximization (DCNSM) problem. We prove that there is no constant-factor approximation for DCNSM, and also present an efficient greedy algorithms achieving $1/ρ$ approximation, where $ρ\leΔ$, $Δ$ is the maximum degree of its dissimilarity based graph. To the best of our knowledge, it is the first approximation algorithm for the Submodular Maximization problem with a distance constraint. The experimental results on real-world attributed network datasets demonstrate the effectiveness of our methods, and confirm that adding dissimilarity constraint can significantly enhance the performance of diversification.

cs.IR↗