Multivariate Time Series Classification (MTSC) is a prevalent but challenging problem in data mining. With the development of Deep Neural Networks (DNN), hundreds of deep models for MTSC have been proposed. However, most prior works only explicitly model the relationship between time series and classes and ignore the diversity of the relationship, suffering from insufficient information exploitation. In this paper, we propose a novel module named Multi-Relationship Modeling Module(MRM2) for more effective MTSC. MRM2 uses the classified labels to explicitly model not only the relationship between time series and classes, but also the relationship among time series, enabling the backbone to generate distinguishable embeddings. In addition, MRM2 is versatile because it can be combined with the existing backbones of DNN for end-to-end training. Finally, we conduct a series of ablation studies and comparative experiments on the real multivariate time series archive UEA. Experimental results indicate that MRM2 can significantly improve classification performance in most cases. Codes are available on GitHub1.

In some computer vision tasks, in addition to images, we also have a graph indicating the relationship of the images. In this case, we can use Graph Convolutional Network (GCN) to get extra information from the graph. GCN is a widely-used deep neural network with graph convolutional layers extracting high-level feature maps from the graph. In this paper, we design a Dual Convolution Neural Network (Dual-CNN) with a CNN solving the images and a GCN solving the graph. We take the lung nodule classification task as an example to introduce our model. In this task, we have features including a CT image and several categorical attributes for each nodule, and the purpose is to classify if the nodule is malignant. Conventional classifiers take these features as the input and then give the prediction, yet we additionally construct a graph based on attributes and perform graph convolution on it. We also design experiments to demonstrate the effectiveness of our Dual-CNN models. Codes are available on https://github.com/temp2244/lung_nodules_classification.

Recently, significant progress has been made in time series classification with deep learning. However, using deep learning models to solve time series classification generally suffers from expensive calculations and difficulty of data labeling. In this work, we study self-supervised time series pre-training to overcome these challenges. Compared with the existing works, we focus on the universal and unlabeled time series pretraining. To this end, we propose a novel end-to-end neural network architecture based on self-attention, which is suitable for capturing long-term dependencies and extracting features from different time series. Then, we propose two different self-supervised pretext tasks for time series data type: Denoising and Similarity Discrimination based on DTW (Dynamic Time Warping). Finally, we carry out extensive experiments on 85 time series datasets (also known as UCR2015 [2]). Empirical results show that the time series model augmented with our proposed self-supervised pretext tasks achieves state-of-the-art / highly competitive results.