Most previous heterogeneous graph embedding models represent elements in a heterogeneous graph as vector representations in a low-dimensional Euclidean space. However, because heterogeneous graphs inherently possess complex structures, such as hierarchical or power-law structures, distortions can occur when representing them in Euclidean space. To overcome this limitation, we propose Hyperbolic Heterogeneous Graph Attention Networks (HHGAT) that learn vector representations in hyperbolic spaces with meta-path instances. We conducted experiments on three real-world heterogeneous graph datasets, demonstrating that HHGAT outperforms state-of-the-art heterogeneous graph embedding models in node classification and clustering tasks.
Graph anomaly detection has attracted considerable attention in recent years. This paper introduces a novel approach that leverages metapath-based semi-supervised learning, addressing the limitations of previous methods. We present a new framework, Metapath-based Semi-supervised Anomaly Detection (MSAD), incorporating GCN layers in both the encoder and decoder to efficiently propagate context information between abnormal and normal nodes. The design of metapath-based context information and a specifically crafted anomaly community enhance the process of learning differences in structures and attributes, both globally and locally. Through a comprehensive set of experiments conducted on seven real-world networks, this paper demonstrates the superiority of the MSAD method compared to state-of-the-art techniques. The promising results of this study pave the way for future investigations, focusing on the optimization and analysis of metapath patterns to further enhance the effectiveness of anomaly detection on attributed networks.
Virtual humans have gained considerable attention in numerous industries, e.g., entertainment and e-commerce. As a core technology, synthesizing photorealistic face frames from target speech and facial identity has been actively studied with generative adversarial networks. Despite remarkable results of modern talking-face generation models, they often entail high computational burdens, which limit their efficient deployment. This study aims to develop a lightweight model for speech-driven talking-face synthesis. We build a compact generator by removing the residual blocks and reducing the channel width from Wav2Lip, a popular talking-face generator. We also present a knowledge distillation scheme to stably yet effectively train the small-capacity generator without adversarial learning. We reduce the number of parameters and MACs by 28$\times$ while retaining the performance of the original model. Moreover, to alleviate a severe performance drop when converting the whole generator to INT8 precision, we adopt a selective quantization method that uses FP16 for the quantization-sensitive layers and INT8 for the other layers. Using this mixed precision, we achieve up to a 19$\times$ speedup on edge GPUs without noticeably compromising the generation quality.
Graphs are used widely to model complex systems, and detecting anomalies in a graph is an important task in the analysis of complex systems. Graph anomalies are patterns in a graph that do not conform to normal patterns expected of the attributes and/or structures of the graph. In recent years, graph neural networks (GNNs) have been studied extensively and have successfully performed difficult machine learning tasks in node classification, link prediction, and graph classification thanks to the highly expressive capability via message passing in effectively learning graph representations. To solve the graph anomaly detection problem, GNN-based methods leverage information about the graph attributes (or features) and/or structures to learn to score anomalies appropriately. In this survey, we review the recent advances made in detecting graph anomalies using GNN models. Specifically, we summarize GNN-based methods according to the graph type (i.e., static and dynamic), the anomaly type (i.e., node, edge, subgraph, and whole graph), and the network architecture (e.g., graph autoencoder, graph convolutional network). To the best of our knowledge, this survey is the first comprehensive review of graph anomaly detection methods based on GNNs.
In recent years, many recommender systems using network embedding (NE) such as graph neural networks (GNNs) have been extensively studied in the sense of improving recommendation accuracy. However, such attempts have focused mostly on utilizing only the information of positive user-item interactions with high ratings. Thus, there is a challenge on how to make use of low rating scores for representing users' preferences since low ratings can be still informative in designing NE-based recommender systems. In this study, we present SiReN, a new sign-aware recommender system based on GNN models. Specifically, SiReN has three key components: 1) constructing a signed bipartite graph for more precisely representing users' preferences, which is split into two edge-disjoint graphs with positive and negative edges each, 2) generating two embeddings for the partitioned graphs with positive and negative edges via a GNN model and a multi-layer perceptron (MLP), respectively, and then using an attention model to obtain the final embeddings, and 3) establishing a sign-aware Bayesian personalized ranking (BPR) loss function in the process of optimization. Through comprehensive experiments, we empirically demonstrate that SiReN consistently outperforms state-of-the-art NE-aided recommendation methods.
Approximate Policy Iteration (API) algorithms alternate between (approximate) policy evaluation and (approximate) greedification. Many different approaches have been explored for approximate policy evaluation, but less is understood about approximate greedification and what choices guarantee policy improvement. In this work, we investigate approximate greedification when reducing the KL divergence between the parameterized policy and the Boltzmann distribution over action values. In particular, we investigate the difference between the forward and reverse KL divergences, with varying degrees of entropy regularization. We show that the reverse KL has stronger policy improvement guarantees, but that reducing the forward KL can result in a worse policy. We also demonstrate, however, that a large enough reduction of the forward KL can induce improvement under additional assumptions. Empirically, we show on simple continuous-action environments that the forward KL can induce more exploration, but at the cost of a more suboptimal policy. No significant differences were observed in the discrete-action setting or on a suite of benchmark problems. Throughout, we highlight that many policy gradient methods can be seen as an instance of API, with either the forward or reverse KL for the policy update, and discuss next steps for understanding and improving our policy optimization algorithms.
Federated learning has emerged as a new paradigm of collaborative machine learning; however, many prior studies have used global aggregation along a star topology without much consideration of the communication scalability or the diurnal property relied on clients' local time variety. In contrast, ring architecture can resolve the scalability issue and even satisfy the diurnal property by iterating nodes without an aggregation. Nevertheless, such ring-based algorithms can inherently suffer from the high-variance problem. To this end, we propose a novel algorithm called TornadoAggregate that improves both accuracy and scalability by facilitating the ring architecture. In particular, to improve the accuracy, we reformulate the loss minimization into a variance reduction problem and establish three principles to reduce variance: Ring-Aware Grouping, Small Ring, and Ring Chaining. Experimental results show that TornadoAggregate improved the test accuracy by up to 26.7% and achieved near-linear scalability.
Information gathering in a partially observable environment can be formulated as a reinforcement learning (RL), problem where the reward depends on the agent's uncertainty. For example, the reward can be the negative entropy of the agent's belief over an unknown (or hidden) variable. Typically, the rewards of an RL agent are defined as a function of the state-action pairs and not as a function of the belief of the agent; this hinders the direct application of deep RL methods for such tasks. This paper tackles the challenge of using belief-based rewards for a deep RL agent, by offering a simple insight that maximizing any convex function of the belief of the agent can be approximated by instead maximizing a prediction reward: a reward based on prediction accuracy. In particular, we derive the exact error between negative entropy and the expected prediction reward. This insight provides theoretical motivation for several fields using prediction rewards---namely visual attention, question answering systems, and intrinsic motivation---and highlights their connection to the usually distinct fields of active perception, active sensing, and sensor placement. Based on this insight we present deep anticipatory networks (DANs), which enables an agent to take actions to reduce its uncertainty without performing explicit belief inference. We present two applications of DANs: building a sensor selection system for tracking people in a shopping mall and learning discrete models of attention on fashion MNIST and MNIST digit classification.
Value-based approaches can be difficult to use in continuous action spaces, because an optimization has to be solved to find the greedy action for the action-values. A common strategy has been to restrict the functional form of the action-values to be convex or quadratic in the actions, to simplify this optimization. Such restrictions, however, can prevent learning accurate action-values. In this work, we propose the Actor-Expert framework for value-based methods, that decouples action-selection (Actor) from the action-value representation (Expert). The Expert uses Q-learning to update the action-values towards the optimal action-values, whereas the Actor (learns to) output the greedy action for the current action-values. We develop a Conditional Cross Entropy Method for the Actor, to learn the greedy action for a generically parameterized Expert, and provide a two-timescale analysis to validate asymptotic behavior. We demonstrate in a toy domain with bimodal action-values that previous restrictive action-value methods fail whereas the decoupled Actor-Expert with a more general action-value parameterization succeeds. Finally, we demonstrate that Actor-Expert performs as well as or better than these other methods on several benchmark continuous-action domains.