Somshubra Majumdar

Bio: Somshubra Majumdar is an academic researcher from University of Illinois at Chicago. The author has contributed to research in topics: Convolutional neural network & Computer science. The author has an hindex of 9, co-authored 26 publications receiving 1017 citations. Previous affiliations of Somshubra Majumdar include University of California, San Francisco & University of Illinois at Urbana–Champaign.

LSTM Fully Convolutional Networks for Time Series Classification

Abstract: Fully convolutional neural networks (FCNs) have been shown to achieve the state-of-the-art performance on the task of classifying time series sequences. We propose the augmentation of fully convolutional networks with long short term memory recurrent neural network (LSTM RNN) sub-modules for time series classification. Our proposed models significantly enhance the performance of fully convolutional networks with a nominal increase in model size and require minimal preprocessing of the data set. The proposed long short term memory fully convolutional network (LSTM-FCN) achieves the state-of-the-art performance compared with others. We also explore the usage of attention mechanism to improve time series classification with the attention long short term memory fully convolutional network (ALSTM-FCN). The attention mechanism allows one to visualize the decision process of the LSTM cell. Furthermore, we propose refinement as a method to enhance the performance of trained models. An overall analysis of the performance of our model is provided and compared with other techniques.

Insights Into LSTM Fully Convolutional Networks for Time Series Classification

Abstract: Long short-term memory fully convolutional neural networks (LSTM-FCNs) and Attention LSTM-FCN (ALSTM-FCN) have shown to achieve the state-of-the-art performance on the task of classifying time series signals on the old University of California-Riverside (UCR) time series repository. However, there has been no study on why LSTM-FCN and ALSTM-FCN perform well. In this paper, we perform a series of ablation tests (3627 experiments) on the LSTM-FCN and ALSTM-FCN to provide a better understanding of the model and each of its sub-modules. The results from the ablation tests on the ALSTM-FCN and LSTM-FCN show that the LSTM and the FCN blocks perform better when applied in a conjoined manner. Two z-normalizing techniques, z-normalizing each sample independently and z-normalizing the whole dataset, are compared using a Wilcoxson signed-rank test to show a statistical difference in performance. In addition, we provide an understanding of the impact dimension shuffle that has on LSTM-FCN by comparing its performance with LSTM-FCN when no dimension shuffle is applied. Finally, we demonstrate the performance of the LSTM-FCN when the LSTM block is replaced by a gated recurrent unit (GRU), basic neural network (RNN), and dense block.

MatchboxNet: 1D Time-Channel Separable Convolutional Neural Network Architecture for Speech Commands Recognition

Abstract: We present an MatchboxNet - an end-to-end neural network for speech command recognition. MatchboxNet is a deep residual network composed from blocks of 1D time-channel separable convolution, batch-normalization, ReLU and dropout layers. MatchboxNet reaches state-of-the-art accuracy on the Google Speech Commands dataset while having significantly fewer parameters than similar models. The small footprint of MatchboxNet makes it an attractive candidate for devices with limited computational resources. The model is highly scalable, so model accuracy can be improved with modest additional memory and compute. Finally, we show how intensive data augmentation using an auxiliary noise dataset improves robustness in the presence of background noise.

Time Series Forecasting of COVID-19 transmission in Canada Using LSTM Networks.

Abstract: On March 11 th 2020, World Health Organization (WHO) declared the 2019 novel corona virus as global pandemic. Corona virus, also known as COVID-19 was first originated in Wuhan, Hubei province in China around December 2019 and spread out all over the world within few weeks. Based on the public datasets provided by John Hopkins university and Canadian health authority, we have developed a forecasting model of COVID-19 outbreak in Canada using state-of-the-art Deep Learning (DL) models. In this novel research, we evaluated the key features to predict the trends and possible stopping time of the current COVID-19 outbreak in Canada and around the world. In this paper we presented the Long short-term memory (LSTM) networks, a deep learning approach to forecast the future COVID-19 cases. Based on the results of our Long short-term memory (LSTM) network, we predicted the possible ending point of this outbreak will be around June 2020. In addition to that, we compared transmission rates of Canada with Italy and USA. Here we also presented the 2, 4, 6, 8, 10, 12 and 14 th day predictions for 2 successive days. Our forecasts in this paper is based on the available data until March 31, 2020. To the best of our knowledge, this of the few studies to use LSTM networks to forecast the infectious diseases.

Digital Twin: Values, Challenges and Enablers From a Modeling Perspective

Abstract: Digital twin can be defined as a virtual representation of a physical asset enabled through data and simulators for real-time prediction, optimization, monitoring, controlling, and improved decision making. Recent advances in computational pipelines, multiphysics solvers, artificial intelligence, big data cybernetics, data processing and management tools bring the promise of digital twins and their impact on society closer to reality. Digital twinning is now an important and emerging trend in many applications. Also referred to as a computational megamodel, device shadow, mirrored system, avatar or a synchronized virtual prototype, there can be no doubt that a digital twin plays a transformative role not only in how we design and operate cyber-physical intelligent systems, but also in how we advance the modularity of multi-disciplinary systems to tackle fundamental barriers not addressed by the current, evolutionary modeling practices. In this work, we review the recent status of methodologies and techniques related to the construction of digital twins mostly from a modeling perspective. Our aim is to provide a detailed coverage of the current challenges and enabling technologies along with recommendations and reflections for various stakeholders.

Automatic detection of microaneurysms in color fundus images for screening of diabetic retinopathy.

Abstract: This paper addresses the automatic detection of microaneurysms in color fundus images, which plays a key role in computer assisted diagnosis of diabetic retinopathy, a serious and frequent eye disease. The algorithm can be divided into four steps. The first step consists in image enhancement, shade correction and image normalization of the green channel. The second step aims at detecting candidates, i.e. all patterns possibly corresponding to MA, which is achieved by diameter closing and an automatic threshold scheme. Then, features are extracted, which are used in the last step to automatically classify candidates into real MA and other objects; the classification relies on kernel density estimation with variable bandwidth. A database of 21 annotated images has been used to train the algorithm. The algorithm was compared to manually obtained gradings of 94 images; sensitivity was 88.5% at an average number of 2.13 false positives per image.

An empirical survey of data augmentation for time series classification with neural networks.

Abstract: In recent times, deep artificial neural networks have achieved many successes in pattern recognition. Part of this success can be attributed to the reliance on big data to increase generalization. However, in the field of time series recognition, many datasets are often very small. One method of addressing this problem is through the use of data augmentation. In this paper, we survey data augmentation techniques for time series and their application to time series classification with neural networks. We propose a taxonomy and outline the four families in time series data augmentation, including transformation-based methods, pattern mixing, generative models, and decomposition methods. Furthermore, we empirically evaluate 12 time series data augmentation methods on 128 time series classification datasets with six different types of neural networks. Through the results, we are able to analyze the characteristics, advantages and disadvantages, and recommendations of each data augmentation method. This survey aims to help in the selection of time series data augmentation for neural network applications.