Abstract WP73: Automatic Classification of Clinical MRI Stroke Datasets With a Recurrent Convolutional Neural Network

To develop a natural language processing system that identifies relations of medications with adverse drug events from clinical narratives. This project is part of the 2018 n2c2 challenge. We developed a novel clinical named entity recognition method based on an recurrent convolutional neural network and compared it to a recurrent neural network implemented using the long-short term memory architecture, explored methods to integrate medical knowledge as embedding layers in neural networks, and investigated 3 machine learning models, including support vector machines, random forests and gradient boosting for relation classification. The performance of our system was evaluated using annotated data and scripts provided by the 2018 n2c2 organizers. Our system was among the top ranked. Our best model submitted during this challenge (based on recurrent neural networks and support vector machines) achieved lenient F1 scores of 0.9287 for concept extraction (ranked third), 0.9459 for relation classification (ranked fourth), and 0.8778 for the end-to-end relation extraction (ranked second). We developed a novel named entity recognition model based on a recurrent convolutional neural network and further investigated gradient boosting for relation classification. The new methods improved the lenient F1 scores of the 3 subtasks to 0.9292, 0.9633, and 0.8880, respectively, which are comparable to the best performance reported in this challenge. This study demonstrated the feasibility of using machine learning methods to extract the relations of medications with adverse drug events from clinical narratives.

This paper proposes a novel deep learning architecture involving combinations of Convolutional Neural Networks (CNN) layers and Recurrent neural networks (RNN) layers that can be used to perform segmentation and classification of 5 cardiac rhythms based on ECG recordings. The algorithm is developed in a sequence to sequence setting where the input is a sequence of five second ECG signal sliding windows and the output is a sequence of cardiac rhythm labels. The novel architecture processes as input both the spectrograms of the ECG signal as well as the heartbeats' signal waveform. Additionally, we are able to train the model in the presence of label noise. The model's performance and generalizability is verified on an external database different from the one we used to train. Experimental result shows this approach can achieve an average F1 scores of 0.89 (averaged across 5 classes). The proposed model also achieves comparable classification performance to existing state-of-the-art approach with considerably less number of training parameters.

In this paper, alternative direction finding methods have been proposed using deep learning techniques. Firstly, Regeression and Classification models have created by using Convolutional Neural Networks (CNNs). In the second Convolutional Neural Networks and Recurrent Neural Networks (RNNs) have been utilized in the proposed methods. Despite having vast amount of direction finding studies, utilization of neural networks is scarce in literature and past works mostly only includes usage of CNNs. In this study, direction finding is performed by learning signals reaching multiple antenna arrays by networks. Created neural networks have been fed with different data formats and their performances against noisy and no-noise data have been shown. In addition, comparative analysis of the developed models were made in the similar Signal-to-Noise Ratio (SNR) range with the subspace based MUSIC algorithm, which is frequently used in direction finding.

Global climate change has become a critical factor influencing marine and coastal safety. According to the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report (AR6), even with global mitigation efforts, the trend of global warming remains irreversible. The increasing intensity of future extreme climate events is expected to affect wave dynamics, storm surges, and coastal hazards. This study aims to assess long-term variations in extreme waves and storm surges around Taiwan under climate change scenarios and to establish design parameters that reflect future conditions for coastal hazard mitigation and engineering applications.Atmospheric data from the EC-Earth3 global climate model under the high-emission scenario (SSP5-8.5) were used to drive wave and ocean models simulating wave fields and storm surge residuals over the Northwest Pacific. To address the coarse resolution of global climate models, an artificial intelligence–based statistical downscaling framework was developed. This approach integrates Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) into a Convolutional Recurrent Neural Network (CRNN) architecture, improving spatial and temporal resolution and generating data representative of coastal-scale processes. The downscaled wave and surge results were analyzed using the Generalized Pareto Distribution (GPD) for extreme-value statistics to estimate design wave heights and storm surge residuals corresponding to return periods of 10, 25, 50, 75, 100, and 200 years. Statistical uncertainties were evaluated using a bootstrap resampling method.Results show that under the high-emission scenario, significant wave heights and storm surge residuals exhibit an increasing trend with longer return periods. Wave analysis reveals marked changes in northern coastal waters, where design wave heights increase from 3.05 m to 5.65 m. The eastern coast shows moderate increases (4.08–4.22 m), while the Taiwan Strait and southern waters remain relatively stable at about 3.6 m and 5.2 m, indicating higher sensitivity of the north to extreme forcing. For storm surges, historical maximum residuals ranged from 0.36 m to 1.49 m, while future projections range from 0.31 m to 1.35 m, showing a slight decrease but similar spatial distribution, with larger deviations along western and island coasts. Design storm surge residuals increase with return period, from about 0.20 m for a 10-year event to 0.32 m for a 200-year event. Under future conditions, increases are projected mainly for western and island coasts, with southern and eastern shores also showing gradual rises.Overall, extreme waves and storm surges around Taiwan exhibit long-term variations under climate change. Although short-term fluctuations remain moderate, both wave and surge intensities increase at longer return periods, implying that future coastal design standards should consider higher thresholds. The AI-based downscaling and extreme-value framework established in this study supports quantitative assessment of coastal hazards, engineering design, and adaptation planning in Taiwan.

This study aimed to develop and evaluate predictive models for injury prevention in athletes using advanced artificial intelligence (AI) techniques. The research focused on harnessing AI to identify injury risk factors, predict potential injuries, and ultimately enhance athlete safety and performance. We initiated the study by collecting extensive data from 500 athletes over a two-year period, amassing more than 1,000,000 data points. The dataset included diverse variables such as physiological metrics (heart rate, VO2 max), biomechanical data (joint angles, muscle activation patterns), training loads (frequency, intensity, and duration), historical injury records, and environmental conditions (temperature, humidity). Data preprocessing involved cleaning datasets to remove noise, addressing 5% of missing values through mean imputation, and normalizing data to a consistent range. We then employed a suite of AI techniques for predictive modeling. Random Forests and Gradient Boosting Machines (GBM) were used for their robustness in handling diverse data types and providing initial performance benchmarks. More complex models, including Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) with LSTM units, were applied for time-series data analysis to capture temporal patterns in training and injury data. We divided the data into training (60%), validation (20%), and test (20%) sets, and used 10-fold cross-validation to ensure model reliability. Our results demonstrated high efficacy in predicting injury risks. The Random Forest model achieved an accuracy of 0.89, while the GBM model attained an AUC-ROC score of 0.92, indicating excellent discriminatory power. CNNs and RNNs achieved mean squared errors of 0.015 and 0.012 on the validation set, respectively. In practical applications, including a partnership with the FC Barcelona Innovation Hub, our models contributed to a 20% reduction in injury rates and showed a correlation coefficient of 0.75 between predicted risks and actual injuries. Continuous learning mechanisms improved the models’ predictive accuracy by 5% over six months. Thus this study validated that AI-driven predictive modeling is a powerful tool for injury prevention in athletes. By integrating sophisticated AI techniques with comprehensive data analysis, we developed models that significantly enhanced injury prevention strategies. The practical applications and positive outcomes from the study underscore the transformative potential of AI in sports medicine, offering a new standard for creating safer and more effective athletic training environments.

In this digital era, there is an exponential growth of text-based content in the electronic world. Data as texts exist in the form of documents, social media posts on Facebook, Twitter, etc., logs, sensor data, and emails. Twitter is a social platform where users express their views on various aspects in a day to day life. Twitter produces over 500 million tweets daily that is 6000 tweets per second. Twitter data is, by definition, very noisy and unstructured in nature. Text classifications based on the machine learning techniques have problems like poor generalization ability and sparsity dimension explosion. Classifiers based on deep learning techniques are implemented to improve accuracy to overcome shortcomings of machine learning techniques and to avoid feature extraction processes and have high prediction accuracy and strong learning ability. In this work, the classification of tweets is performed on Covid-19 dataset by implementing deep learning techniques namely Convolution Neural Network (CNN), Recurrent Neural Network (RNN), Recurrent Convolution Neural Network (RCNN), Recurrent Neural Network with Long Short Term Memory (RNN+LSTM), and Bidirectional Long Short Term Memory with Attention (BI-LSTM + Attention). The algorithms are implemented using two-word embedding techniques namely Global Vectors for Word Representation (GloVe) and Word2Vec. RNN with Bidirectional LSTM model has performed better than all the classifiers considered. It has classified the text with an accuracy of 93% and above when used with GloVe and Word2Vec.

Self-attention based recurrent convolutional neural network for disease prediction using healthcare data

Optical character recognition is a technology that turns texts and scanned documents into digital formats. The practical applications of OCR systems face a lot of challenges because of heterogeneity in scripts, font styles, and different quality images arising during their usage. The proposed work attempts to improve the robustness of OCR models built for the Hindi language by addressing the deficiencies inherent in the widely used IIIT-HW-Dev dataset. Toward this end, two methods of data augmentation are proposed: 1) synthetic images with half characters and conjuncts for complex words, and 2) various image degradations to simulate the real world. The augmented training set is used to train the two deep architectures, namely Convolutional Recurrent Neural Networks (CRNNs) and ResNet-50. Comparisons and in-depth experiments further reveal the merits of the suggested data augmentation techniques: the top performance obtained by the CRNN model exceeds current approaches. On the test set, CRNN achieved a Character Error Rate (CER) of 2.14% and Word Error Rate (WER) of 7.96%, much beyond the CER of 3.27% and WER of 11.82% achieved by the ResNet-50 model. Qualitatively, there was robustness in the case of complex words and images that were also deteriorated. The augmented dataset and trained models can be downloaded by anyone to promote more research in Hindi OCR. So, developing robust OCR systems for complex scripts starts from variability-rich and representative training data along with specific architectures, such as the CRNN architecture.

In existing convolutional neural networks (CNNs), both convolution and pooling are locally performed for image regions separately, no contextual dependencies between different image regions have been taken into consideration. Such dependencies represent useful spatial structure information in images. Whereas recurrent neural networks (RNNs) are designed for learning contextual dependencies among sequential data by using the recurrent (feedback) connections. In this work, we propose the convolutional recurrent neural network (C-RNN), which learns the spatial dependencies between image regions to enhance the discriminative power of image representation. The C-RNN is trained in an end-to-end manner from raw pixel images. CNN layers are firstly processed to generate middle level features. RNN layer is then learned to encode spatial dependencies. The C-RNN can learn better image representation, especially for images with obvious spatial contextual dependencies. Our method achieves competitive performance on ILSVRC 2012, SUN 397, and MIT indoor.

Automatic sleep stage monitoring is an essential tool for the diagnosis and treatment of sleep-related disorders effectively. Although extensive studies focusing on single source data or information, such as single channel electroencephalogram (EEG), electrooculogram (EOG), and electromyogram (EMG), with machine learning (ML) and deep learning (DL) have been presented in this field of automatic sleep staging, there are very limited literature's that have investigated the impact of multisource information fusion on ML or DL based schemes for the sleep stage classification. In addition, exploiting recurrent neural network (RNN) to learn temporal information from the sequences of multi-source inputs can bring significant improvement in classifying the sleep stage. Therefore, we aim to develop an multi-source data fusion, more specifically two source's signal fusion, enabled end-to-end convolutional recurrent neural network (CRNN) in order to perform automatic sleep stage classification in this article. In the proposed scheme, no hand crafted features is used to train the model and hence the proposed model can be called an end-to-end approach. The proposed CRNN model can extract time-invariant features from raw data of two input sources and can fuse them to learn temporally correlated features. Sleep-EDF expanded benchmark dataset is used in this study and we have employed Fpz-Cz EEG channel signal and EOG signal as input with 3 variants of RNN layer in model architecture separately in order to evaluate and investigate the performance. The proposed CRNN model with standard RNN layers can classify the Sleep-EDF dataset with maximum average accuracy of 90.30% and maximum Cohen's kappa coefficient of 86.86 compared other two RNN variant's layer (e.g., gated recurrent unit and long short term memory) based models which is very promising in the studies on classification of sleep stages as well as can be regarded as an alternative to conventional sleep staging.

The study presents the results of the development and testing of deep learning neural network architectures, which demonstrate high accuracy rates in classifying neurophysiological data, in particular hemodynamic brain activation patterns obtained by functional near-infrared spectroscopy, during solving mathematical problems on spatial-numerical associations. The analyzed signal represents a multidimensional time series of oxyhemoglobin and deoxyhemoglobin dynamics. Taking the specificity of the fNIRS signal into account, a comparative analysis of 2 types of neural network architectures was carried out: (1) architectures based on recurrent neural networks: recurrent neural network with long short-term memory, recurrent neural network with long short-term memory with fully connected layers, bidirectional recurrent neural network with long short-term memory, convolutional recurrent neural network with long short-term memory; (2) architectures based on convolutional neural networks with 1D convolutions: convolutional neural network, fully convolutional neural network, residual neural network. Trained long short-term memory recurrent neural network architectures showed worse results in accuracy in comparison with 1D convolutional neural network architectures. Residual neural network (model_Resnet) demonstrated the highest accuracy rates in three experimental conditions more than 88% in detecting age-related differences in brain activation during spatial-numerical association tasks considering the individual characteristics of the respondents’ signal.

Neural network models have recently been applied to the task of automatic essay scoring, giving promising results. Existing work used recurrent neural networks and convolutional neural networks to model input essays, giving grades based on a single vector representation of the essay. On the other hand, the relative advantages of RNNs and CNNs have not been compared. In addition, different parts of the essay can contribute differently for scoring, which is not captured by existing models. We address these issues by building a hierarchical sentence-document model to represent essays, using the attention mechanism to automatically decide the relative weights of words and sentences. Results show that our model outperforms the previous state-of-the-art methods, demonstrating the effectiveness of the attention mechanism.

The presented paper proposes a new speech command recognition model for novel engineering applications with limited resources. We built the proposed model with the help of a Convolutional Recurrent Neural Network (CRNN). The use of CRNN instead of Convolutional Neural Network (CNN) helps us to reduce the model parameters and memory requirement as per resource constraints. Furthermore, we insert transmute and curtailment layer between the layers of CRNN. By doing this we further reduce model parameters and float number of operations to half of the CRNN requirement. The proposed model is tested on Google’s speech command dataset. The obtained result shows that the proposed CRNN model requires 1/3 parameters as compared to the CNN model. The number of parameters of the CRNN model is further reduced by 45% and the float numbers of operations between 2% to 12 % in different recognition tasks. The recognition accuracy of the proposed model is 96% on Google’s speech command dataset, and on laboratory recording, its recognition accuracy is 89%.

Accurate and effective mapping of soil properties is regarded as a critical task in environmental and agricultural management. The evaluation of properties of soil is a daunting task while monitoring and sensing the environment. Existing sampling methods is a time-consuming and laborious job and they are limited based on the regions. However, the need of soil analysis and its properties is essential at landscape level. In this paper, we use Recurrent Convolution Neural Network (RCNN) to assess the soil properties via its classification task. The model in turn is compared with conventional geostatistical spatial interpolation methods. The utilization of Recurrent Neural Network (RNN) aims at studying the spatial and temporal variability of the properties of soil that adopts Kriging interpolation technique. The simulation is conducted to study the efficacy of the model under different soil conditions and the efficacy of RCNN is reported. The results of simulation shows that the proposed method achieves higher rate of classification accuracy than other models.

Mobile robot localization has been considered to be an important task in the field of robotics research. It is known that it is difficult to estimate the self-position in dynamic environments where the positions of objects used as landmarks change. In this paper, we propose a robust method to estimate self-position from the first person view captured by a camera on a robot using Recurrent Convolutional Neural Networks (RCNN), which is a neural network model that has a convolutional architecture known as CNN with recurrent nodes. The RCNN receives images and directly estimates the positions of the robot. Our proposed method is evaluated in simulated environments. Our experiments show that RCNN model can estimate the selfposition of the robot with high accuracy even if some objects move to different positions, that is, it has a robustness against objects obstructing visibility.