Lots of learning tasks require dealing with graph data which contains rich relation information among elements. Modeling physics systems, learning molecular fingerprints, predicting protein interface, and classifying diseases demand a model to learn from graph inputs. In other domains such as learning from non-structural data like texts and images, reasoning on extracted structures (like the dependency trees of sentences and the scene graphs of images) is an important research topic which also needs graph reasoning models. Graph neural networks (GNNs) are neural models that capture the dependence of graphs via message passing between the nodes of graphs. In recent years, variants of GNNs such as graph convolutional network (GCN), graph attention network (GAT), graph recurrent network (GRN) have demonstrated ground-breaking performances on many deep learning tasks. In this survey, we propose a general design pipeline for GNN models and discuss the variants of each component, systematically categorize the applications, and propose four open problems for future research.

Graph Neural Networks: A Review of Methods and Applications

What is Twitter

Sequential location recommendation plays an important role in many applications such as mobility prediction, route planning and location-based advertisements. In spite of evolving from tensor factorization to RNN-based neural networks, existing methods did not make effective use of geographical information and suffered from the sparsity issue. To this end, we propose a Geography-aware sequential recommender based on the Self-Attention Network (GeoSAN for short) for location recommendation. On the one hand, we propose a new loss function based on importance sampling for optimization, to address the sparsity issue by emphasizing the use of informative negative samples. On the other hand, to make better use of geographical information, GeoSAN represents the hierarchical gridding of each GPS point with a self-attention based geography encoder. Moreover, we put forward geography-aware negative samplers to promote the informativeness of negative samples. We evaluate the proposed algorithm with three real-world LBSN datasets, and show that GeoSAN outperforms the state-of-the-art sequential location recommenders by 34.9%. The experimental results further verify significant effectiveness of the new loss function, geography encoder, and geography-aware negative samplers.

Geography-Aware Sequential Location Recommendation

A range of applications, from predicting the spread of human and electronic viruses to city planning and resource management in mobile communications, depend on our ability to foresee the whereabouts and mobility of individuals, raising a fundamental question: To what degree is human behavior predictable? Here we explore the limits of predictability in human dynamics by studying the mobility patterns of anonymized mobile phone users. By measuring the entropy of each individual's trajectory, we find a 93% potential predictability in user mobility across the whole user base. Despite the significant differences in the travel patterns, we find a remarkable lack of variability in predictability, which is largely independent of the distance users cover on a regular basis.

http://absimage.aps.org/image/MAR10/MWS_MAR10-2009-007155.pdf

Limits of predictability in human mobility

The recent growth of social networking platforms also led to the emergence of social spammers, who overwhelm legitimate users with unwanted content. The existing social spammer detection methods can be characterized into two categories: features based ones and propagation-based ones. Features based methods mainly rely on matrix factorization using tweet text features, and regularization using social graphs is incorporated. However, these methods are fully supervised and can only utilize labeled part of social graphs, which fail to work in a real-world semi-supervised setting. The propagation-based methods primarily employ Markov Random Fields (MRFs) to capture human intuitions in user following relations, which cannot take advantages of rich text features. In this paper, we propose a novel social spammer detection model based on Graph Convolutional Networks (GCNs) that operate on directed social graphs by explicitly considering three types of neighbors. Furthermore, inspired by the propagation-based methods, we propose a MRF layer with refining effects to encapsulate these human insights in social relations, which can be formulated as a RNN through mean-field approximate inference, and stack on top of GCN layers to enable end-to-end training. We evaluate our proposed method on two real-world social network datasets, and the results demonstrate that our method outperforms the state-of-the-art approaches.

/pdf/graph-convolutional-networks-with-markov-random-field-2nwko8y9o3.pdf

Graph Convolutional Networks with Markov Random Field Reasoning for Social Spammer Detection

Inferring social relations from user trajectory data is of great value in real-world applications such as friend recommendation and ride-sharing. Most existing methods predict relationship based on a pairwise approach using some hand-crafted features or rely on a simple skip-gram based model to learn embeddings on graphs. Using hand-crafted features often fails to capture the complex dynamics in human social relations, while the graph embedding based methods only use random walks to propagate information and cannot incorporate external semantic data provided. We propose a novel model that utilizes Graph Convolutional Networks (GCNs) to learn user embeddings on the User Mobility Heterogeneous Graph in an unsupervised manner. This model is capable of propagating relation layer-wisely as well as combining both the rich structural information in the heterogeneous graph and predictive node features provided. Our method can also be extended to a semi-supervised setting if a part of the social network is available. The evaluation on three real-world datasets demonstrates that our method outperforms the state-of-the-art approaches.

/pdf/graph-convolutional-networks-on-user-mobility-heterogeneous-4o147kpmim.pdf

Graph Convolutional Networks on User Mobility Heterogeneous Graphs for Social Relationship Inference

Graph convolutional networks (GCNs) have recently enabled a popular class of algorithms for collaborative filtering (CF). Nevertheless, the theoretical underpinnings of their empirical successes remain elusive. In this paper, we endeavor to obtain a better understanding of GCN-based CF methods via the lens of graph signal processing. By identifying the critical role of smoothness, a key concept in graph signal processing, we develop a unified graph convolution-based framework for CF. We prove that many existing CF methods are special cases of this framework, including the neighborhood-based methods, low-rank matrix factorization, linear auto-encoders, and LightGCN, corresponding to different low-pass filters. Based on our framework, we then present a simple and computationally efficient CF baseline, which we shall refer to as Graph Filter based Collaborative Filtering (GF-CF). Given an implicit feedback matrix, GF-CF can be obtained in a closed form instead of expensive training with back-propagation. Experiments will show that GF-CF achieves competitive or better performance against deep learning-based methods on three well-known datasets, notably with a 70% performance gain over LightGCN on the Amazon-book dataset.

How Powerful is Graph Convolution for Recommendation

Self-attention has become increasingly popular in a variety of sequence modeling tasks from natural language processing to recommendation, due to its effectiveness. However, self-attention suffers from quadratic computational and memory complexities, prohibiting its applications on long sequences. Existing approaches that address this issue mainly rely on a sparse attention context, either using a local window, or a permuted bucket obtained by locality-sensitive hashing (LSH) or sorting, while crucial information may be lost. Inspired by the idea of vector quantization that uses cluster centroids to approximate items, we propose LISA (LInear-time Self Attention), which enjoys both the effectiveness of vanilla self-attention and the efficiency of sparse attention. LISA scales linearly with the sequence length, while enabling full contextual attention via computing differentiable histograms of codeword distributions. Meanwhile, unlike some efficient attention methods, our method poses no restriction on casual masking or sequence length. We evaluate our method on four real-world datasets for sequential recommendation. The results show that LISA outperforms the state-of-the-art efficient attention methods in both performance and speed; and it is up to 57x faster and 78x more memory efficient than vanilla self-attention.

Yongji Wu

Papers

Geography-Aware Sequential Location Recommendation

Graph Convolutional Networks with Markov Random Field Reasoning for Social Spammer Detection

Graph Convolutional Networks on User Mobility Heterogeneous Graphs for Social Relationship Inference

How Powerful is Graph Convolution for Recommendation

Linear-Time Self Attention with Codeword Histogram for Efficient Recommendation