Yongkun Wang

A Generalized Recurrent Neural Architecture for Text Classification with Multi-Task Learning

Multi-task learning leverages potential correlations among related tasks to extract common features and yield performance gains. However, most previous works only consider simple or weak interactions, thereby failing to model complex correlations among three or more tasks. In this paper, we propose a multi-task learning architecture with four types of recurrent neural layers to fuse information across multiple related tasks. The architecture is structurally flexible and considers various interactions among tasks, which can be regarded as a generalized case of many previous works. Extensive experiments on five benchmark datasets for text classification show that our model can significantly improve performances of related tasks with additional information from others.

Generative Warfare Nets: Ensemble via Adversaries and Collaborators

Generative Adversarial Nets are a powerful method for training generative models of complex data, where a Generator and a Discriminator confront with each other and get optimized in a two-player minmax manner. In this paper, we propose the Generative Warfare Nets (GWN) that involve multiple generators and multiple discriminators from two sides to exploit the advantages of Ensemble Learning. We maintain the authorities for the generators and the discriminators to enhance inter-side interactions, and utilize the mechanisms of imitation and innovation to model intra-side interactions among the generators, where they can not only learn from but also compete with each other. Extensive experiments on three natural image datasets show that GWN can achieve state-of-the-art Inception scores and produce diverse high-quality synthetic results.

Transformable Convolutional Neural Network for Text Classification

Convolutional neural networks (CNNs) have shown their promising performance for natural language processing tasks, which extract n-grams as features to represent the input. However, n-gram based CNNs are inherently limited to fixed geometric structure and cannot proactively adapt to the transformations of features. In this paper, we propose two modules to provide CNNs with the flexibility for complex features and the adaptability for transformation, namely, transformable convolution and transformable pooling. Our method fuses dynamic and static deviations to redistribute the sampling locations, which can capture both current and global transformations. Our modules can be easily integrated by other models to generate new transformable networks. We test proposed modules on two state-of-the-art models, and the results demonstrate that our modules can effectively adapt to the feature transformation in text classification.

Discovering and modeling meta-structures in human behavior from city-scale cellular data

Abstract For a long time, researchers explore spatio-temporal properties in mobility to understand human behavior. They have discovered many statistical laws about human dynamics. Unfortunately, we still have limited knowledge about the spatio-temporal structure of individuals’ movement at a large scale. In this paper, we studied the unified spatio-temporal structures (i.e., meta-structures ) in human mobility. We hereby propose a meta-structure discovery algorithm by coupling both topology and spatio-temporal attributes of mobility graphs. With the construction of individual profiles from meta-structure analyses, we provided a novel mobility model from a process-driven perspective, which reduced the dependence of many existing models on the consistency between local and global mobility statistics. We gained some insights on the dominating meta-structures in human mobility by leveraging mobile data in a large city. The statistical distribution of meta-structures is found to be determined by the intrinsic heterogeneity of spatio-temporal properties in human behavior. Our model evaluation showed that a process with basic rules could demonstrate the key statistical properties in mobility meta-structures. We believe that these approaches and observations would be a good reference for management of human mobility in mobile networks and transportation systems.