Spatio-Temporal Graph Convolutional Networks: A Deep Learning Framework for Traffic Forecasting
- Ting Yu, Haoteng Yin, Zhanxing Zhu
- Computer ScienceInternational Joint Conference on Artificial…
- 14 September 2017
This paper proposes a novel deep learning framework, Spatio-Temporal Graph Convolutional Networks (STGCN), to tackle the time series prediction problem in traffic domain, and builds the model with complete convolutional structures, which enable much faster training speed with fewer parameters.
You Only Propagate Once: Accelerating Adversarial Training via Maximal Principle
- Dinghuai Zhang, Tianyuan Zhang, Yiping Lu, Zhanxing Zhu, Bin Dong
- Computer ScienceNeural Information Processing Systems
- 2 May 2019
It is shown that adversarial training can be cast as a discrete time differential game, and the proposed algorithm YOPO (You Only Propagate Once) can achieve comparable defense accuracy with approximately 1/5 ~ 1/4 GPU time of the projected gradient descent (PGD) algorithm.
Spatial-Temporal Fusion Graph Neural Networks for Traffic Flow Forecasting
- Mengzhang Li, Zhanxing Zhu
- Computer Science, Environmental ScienceAAAI Conference on Artificial Intelligence
- 15 December 2020
A data-driven method of generating “temporal graph” is proposed to compensate several genuine correlations that spatial graph may not reflect, and a novel Spatial-Temporal Fusion Graph Neural Networks (STFGNN) is proposed for traffic flow forecasting.
The Anisotropic Noise in Stochastic Gradient Descent: Its Behavior of Escaping from Sharp Minima and Regularization Effects
- Zhanxing Zhu, Jingfeng Wu, Bin Yu, Lei Wu, Jinwen Ma
- Computer ScienceInternational Conference on Machine Learning
- 1 March 2018
This work studies a general form of gradient based optimization dynamics with unbiased noise, which unifies SGD and standard Langevin dynamics, and shows that the anisotropic noise in SGD helps to escape from sharp and poor minima effectively, towards more stable and flat minima that typically generalize well.
Multi-Stage Self-Supervised Learning for Graph Convolutional Networks
- Ke Sun, Zhanxing Zhu, Zhouchen Lin
- Computer ScienceAAAI Conference on Artificial Intelligence
- 28 February 2019
A novel training algorithm for Graph Convolutional Network, called Multi-Stage Self-Supervised (M3S) Training Algorithm, combined with self-supervised learning approach, focusing on improving the generalization performance of GCNs on graphs with few labeled nodes.
Spatio-temporal Graph Convolutional Neural Network: A Deep Learning Framework for Traffic Forecasting
- Bin Yu, Haoteng Yin, Zhanxing Zhu
- Computer Science
- 2017
Reinforced Continual Learning
- Ju Xu, Zhanxing Zhu
- Computer ScienceNeural Information Processing Systems
- 1 May 2018
A novel approach for continual learning is proposed, which searches for the best neural architecture for each coming task via sophisticatedly designed reinforcement learning strategies, and which outperforms existing continual learning alternatives for deep networks.
Understanding and Enhancing the Transferability of Adversarial Examples
- Lei Wu, Zhanxing Zhu, Cheng Tai, E. Weinan
- Computer ScienceArXiv
- 27 February 2018
This work systematically study how two classes of factors that might influence the transferability of adversarial examples are influenced, including model-specific factors, including network architecture, model capacity and test accuracy, and the local smoothness of loss function for constructing adversarial example.
You Only Propagate Once: Painless Adversarial Training Using Maximal Principle
- Dinghuai Zhang, Tianyuan Zhang, Yiping Lu, Zhanxing Zhu, Bin Dong
- Computer Science
- 2 May 2019
This work fully exploits structure of deep neural networks and proposes a novel strategy to decouple the adversary update with the gradient back propagation, which avoids forward and backward propagating the data too many times in one iteration, and restricts core descent directions computation to the first layer of the network, thus speeding up every iteration significantly.
Interpreting Adversarially Trained Convolutional Neural Networks
- Tianyuan Zhang, Zhanxing Zhu
- Computer ScienceInternational Conference on Machine Learning
- 23 May 2019
Surprisingly, it is found that adversarial training alleviates the texture bias of standard CNNs when trained on object recognition tasks, and helps CNNs learn a more shape-biased representation.
...
...