In a Massive Open Online Course (MOOC), predictive models of student behavior can support multiple aspects of learning, including instructor feedback and timely intervention. Ongoing courses, when the student outcomes...
详细信息
ISBN:
(纸本)9781450362566
In a Massive Open Online Course (MOOC), predictive models of student behavior can support multiple aspects of learning, including instructor feedback and timely intervention. Ongoing courses, when the student outcomes are yet unknown, must rely on models trained from the historical data of previously offered courses. It is possible to transfer models, but they often have poor prediction performance. One reason is features that inadequately represent predictive attributes common to both courses. We present an automated transductive transfer learning approach that addresses this issue. It relies on problem-agnostic, temporal organization of the MOOC clickstream data, where, for each student, for multiple courses, a set of specific MOOC event types is expressed for each time unit. It consists of two alternative transfer methods based on representation learning with auto-encoders: a passive approach using transductive principal component analysis and an active approach that uses a correlation alignment loss term. With these methods, we investigate the transferability of dropout prediction across similar and dissimilar MOOCs and compare with known methods. Results show improved model transferability and suggest that the methods are capable of automatically learning a feature representation that expresses common predictive characteristics of MOOCs.
Detecting topics from textual data streams is an interesting task in social networks studies. Traditional techniques have certain limitations when processing social network data such as tweets and online conversations...
详细信息
ISBN:
(纸本)9781728100036
Detecting topics from textual data streams is an interesting task in social networks studies. Traditional techniques have certain limitations when processing social network data such as tweets and online conversations, because of the large amount of data and noises. Deep learning appears to be a viable approach for harvesting and extracting valuable knowledge from complex systems. Therefore, we suggest using deep autoencoder model with Koreans-H- algorithm and work with the reconstructed data that contains less noise to detect the eventual topics within it. We evaluate the proposed model on two public datasets of annotated topics. Then, we compare our results to three well known methods. According to the results, our deep learning based method for detecting topics from social networks data outperforms all the three methods, and was able to detect perfectly the right topics in unsupervised way.
Early diagnosis of pulmonary nodules is critical for lung cancer clinical management. In this paper, a novel framework for pulmonary nodule diagnosis, using descriptors extracted from single computed tomography (CT) s...
详细信息
ISBN:
(纸本)9781538662496
Early diagnosis of pulmonary nodules is critical for lung cancer clinical management. In this paper, a novel framework for pulmonary nodule diagnosis, using descriptors extracted from single computed tomography (CT) scan, is introduced. This framework combines appearance and shape descriptors to give an indication of the nodule prior growth rate, which is the key point for diagnosis of lung nodules. Resolved Ambiguity Local Binary Pattern and 7th Order Markov Gibbs Random Field are developed to describe the nodule appearance without neglecting spatial information. Spherical harmonics expansion and some primitive geometric features are utilized to describe how the nodule shape is complicated. Ultimately, all descriptors are combined using denoising autoencoder to classify the nodule, whether malignant or benign. Training, testing, and parameter tuning of all framework modules are done using a set of 727 nodules extracted from the Lung Image Database Consortium (LIDC) dataset. The proposed system diagnosis accuracy, sensitivity, and specificity were 94:95%, 94:62%, 95:20% respectively, all of which show that our system has promise to reach the accepted clinical accuracy threshold.
Breakdown of equipment causes very large damage to the factory. Research is continuously being conducted to prevent break down of equipment by detecting abnormal signs before equipment failure. This paper proposes an ...
详细信息
ISBN:
(纸本)9783030243111
Breakdown of equipment causes very large damage to the factory. Research is continuously being conducted to prevent break down of equipment by detecting abnormal signs before equipment failure. This paper proposes an anomaly detection for system architecture based on a docker container. A docker is a virtualized container with many performance and scalability advantages. We have used the deep learning model of autoencoder to effectively anomaly detection and its performance has been proven through experiments.
Recent research has demonstrated how deep neural networks are able to learn representations to improve data clustering. By considering representation learning and clustering as a joint task, models learn clustering-fr...
详细信息
ISBN:
(纸本)9783030261429;9783030261412
Recent research has demonstrated how deep neural networks are able to learn representations to improve data clustering. By considering representation learning and clustering as a joint task, models learn clustering-friendly spaces and achieve superior performance, compared with standard two-stage approaches where dimensionality reduction and clustering are performed separately. We extend this idea to topology-preserving clustering models, known as self-organizing maps (SOM). First, we present the Deep Embedded Self-Organizing Map (DESOM), a model composed of a fully-connected autoencoder and a custom SOM layer, where the SOM code vectors are learnt jointly with the autoencoder weights. Then, we show that this generic architecture can be extended to image and sequence data by using convolutional and recurrent architectures, and present variants of these models. First results demonstrate advantages of the DESOM architecture in terms of clustering performance, visualization and training time.
Background The main goal of this study is to explore the use of features representing patient-level electronic health record (EHR) data, generated by the unsupervised deep learning algorithm autoencoder, in predictive...
详细信息
Background The main goal of this study is to explore the use of features representing patient-level electronic health record (EHR) data, generated by the unsupervised deep learning algorithm autoencoder, in predictive modeling. Since autoencoder features are unsupervised, this paper focuses on their general lower-dimensional representation of EHR information in a wide variety of predictive tasks. Methods We compare the model with autoencoder features to traditional models: logistic model with least absolute shrinkage and selection operator (LASSO) and Random Forest algorithm. In addition, we include a predictive model using a small subset of response-specific variables (Simple Reg) and a model combining these variables with features from autoencoder (Enhanced Reg). We performed the study first on simulated data that mimics real world EHR data and then on actual EHR data from eight Advocate hospitals. Results On simulated data with incorrect categories and missing data, the precision for autoencoder is 24.16% when fixing recall at 0.7, which is higher than Random Forest (23.61%) and lower than LASSO (25.32%). The precision is 20.92% in Simple Reg and improves to 24.89% in Enhanced Reg. When using real EHR data to predict the 30-day readmission rate, the precision of autoencoder is 19.04%, which again is higher than Random Forest (18.48%) and lower than LASSO (19.70%). The precisions for Simple Reg and Enhanced Reg are 18.70 and 19.69% respectively. That is, Enhanced Reg can have competitive prediction performance compared to LASSO. In addition, results show that Enhanced Reg usually relies on fewer features under the setting of simulations of this paper. Conclusions We conclude that autoencoder can create useful features representing the entire space of EHR data and which are applicable to a wide array of predictive tasks. Together with important response-specific predictors, we can derive efficient and robust predictive models with less labor in data extraction and
To establish advanced analytical methods for preserving cultural heritage, this research proposes a method to generate a time-lapse image with a super-long temporal interval. The key issue is to realize an image colle...
详细信息
ISBN:
(纸本)9781728108582
To establish advanced analytical methods for preserving cultural heritage, this research proposes a method to generate a time-lapse image with a super-long temporal interval. The key issue is to realize an image collection method using crowdsourcing and a method to improve the matching accuracy between images of cultural heritage buildings captured 50 to 100 years ago and current images. As degradation and damage to the appearance of cultural heritage buildings occurs due to ageing, rebuilding, and renovation, image features of the timed images are changed. This decreases the accuracy of the matching process that uses the appearance of patch-region. In addition, we need to give more consideration to incorrect feature correspondence that is prominent in buildings with considerable symmetry. We aim to solve these difficulties by applying an autoencoder and a guided matching method. Our method involves utilizing the function of crowdsourcing, which can easily obtain the current image captured at the same position and orientation as the past image. We propose this method to address the inability to obtain the correspondence points between two images when observation times are significantly different.
We introduce a periodic loss function and corresponding activation function, to be used for neural network regression and autoencoding task involving periodic targets. Such target features, typically represented in no...
详细信息
ISBN:
(纸本)9789811519604;9789811519598
We introduce a periodic loss function and corresponding activation function, to be used for neural network regression and autoencoding task involving periodic targets. Such target features, typically represented in non-Cartesian coordinates, arise mainly from angular distributions, but also include repeating time series, e.g. 24-h cycles or seasonal intervals. To demonstrate the use of this loss function, two different use-cases within the context of high-energy physics are presented. The first is a simple regression network, trained to predict the angle between particles emerging from the decay of a heavier, unstable particle. Next, we look at the same particle decay, but train an autoencoder to reproduce all inputs, which include both cyclic and noncyclic features. All examples show that failing to incorporate the cyclic property of the targets into the loss and activation function significantly degrades the performance of the model predictions.
In the signal processing field, there is a growing interest in speech enhancement. Recently, a lot of speech enhancement methods based on the deep neural network have been proposed. Mostly, these networks, such as SEG...
详细信息
ISBN:
(纸本)9781728130385
In the signal processing field, there is a growing interest in speech enhancement. Recently, a lot of speech enhancement methods based on the deep neural network have been proposed. Mostly, these networks, such as SEGAN, Wave-U-Net, adopt the autoencoder structure. In this paper, we propose the cross conditional network for speech enhancement based on SEGAN architecture. The proposed network has two Auto-Encoder, where the mutual latent vector is composed of the concatenated vector of these encoder outputs. In the experiments, we show that the proposed method exceeds SEGAN in terms of the objective evaluation measure by PESQ.
A speaker embeddings framework achieves state-of-the-art speaker recognition performance by modeling speaker discriminant information directly using deep neural networks ( DNNs). After the introduction of neural netwo...
详细信息
ISBN:
(纸本)9781479981311
A speaker embeddings framework achieves state-of-the-art speaker recognition performance by modeling speaker discriminant information directly using deep neural networks ( DNNs). After the introduction of neural network based speaker embeddings, researchers have explored the requirements for training an effective embeddings network. However, the domain of the data used for system development should match the domain of operation for optimal performance. In this paper, we investigate the sensitivity of domain mismatch in the embeddings space. Specifically, degradation in performance is observed when back-end scoring with embeddings is performed with out-domain data. To compensate for the domain mismatch, we propose two novel deep domain adaptation techniques based on autoencoder architectures trained on embeddings in an unsupervised fashion. The results show that domain mismatch can be compensated effectively using autoencoders to adapt the out-domain data to in-domain.
暂无评论