We present ConvEQ as a tool for discriminating seismic phases, leveraging artificial intelligence technique (Convolutional Neural Network) for short-time Frequency Transform of the seismic signal. Timely detection of the vertical (P) wave from an earthquake can generate a warning several tens of precious seconds before the more destructive waves strike. We propose a train-for-each-station approach for an Internet-of-Things-based Smart Earthquake Early Warning System, where lightweight neural networks trained for the seismic data belonging to each station are implemented on edge devices directly interfaced with seismometers. The approach has the potential to get the most from the sparse seismic network for Pakistan and other third-world countries. We train networks for multi-station and single-station data and achieve 96% and 99% accuracy, respectively, proving that train-for-each-station maximizes accuracy. The total processing time (including preprocessing and inference) is about 30ms for each event, thus suitable for real-time deployment. We further compare the performance of ConvEQ on simulated real-time data with several state-of-the-art contemporary algorithms. Our proposed approach demonstrates a robust response on diverse metrics. The ConvEQZ classifies the vertical seismic signal component with high accuracy and the ConvEQX can classify any seismic data component, inculcating robustness against connectivity issues.

Today’s modern world has seen massive population growth in the past few decades, causing the automobile industry to expand rapidly. The exponential increase has disrupted the seamless flow of traffic globally and has fashioned significant problems such as traffic congestion and its mismanagement. To tackle these issues, some hardware and data-driven-based solutions have been previously proposed. The hardware-based solutions utilize different types of sensors installed at relevant sites to monitor the status of parking slots which makes them less scalable, complex, costly to install and maintain. In contrast, the data-driven solutions utilize the already available infrastructure of surveillance cameras installed at parking lots, thus overcoming the limitations of sensor-based solutions. However, currently only classification-based approaches are prevalent and adopted to monitor the status of the parking slots which makes the systems less generalized for scalability with slower performance. This paper proposes an intelligent parking management system which employs deep learning to alleviate the limitations in the data driven solutions by leveraging the high performance and fast inference capability of YOLO v5 for vehicles detection instead of parking slot classification. The model was evaluated using the PKLot dataset which is a benchmark for identifying the status of the parking lot with state of the art performance achieved having an accuracy of 99.5%. To augment the performance of the algorithm in real-time, a pretrained model of YOLO v5 on MS COCO dataset was employed to detect and assign vacant parking slots and generate vehicle statistics. The performance of the proposed system was evaluated on our custom dataset with an accuracy of 96.8% achieved and a nearly real-time performance of 45 Fps which makes it more efficient, scalable, and generalized.

Surveillance videos record malicious events in a locality utilizing various machine learning algorithms for detection.Deep-learning algorithms being the most prominent AI algorithms are data-hungry as well as computationally expensive.These algorithms perform better when trained over a diverse and huge set of examples.These modern AI methods have a dire need of utilizing human intelligence to pamper the problem in such a way as to reduce the ultimate effort in terms of computational cost.In this research work, a novel methodology termed Bag of Focus (BoF) based training methodology has been proposed.BoF is based on the concept of selecting motionintensive blocks in a long video, for training different deep neural networks (DNN's).The methodology reduced the computational overhead by 90% (ten times) in comparison to when full-length videos are entertained.It has been observed that training networks using BoF are equally effective in terms of performance for the same network trained over the full-length dataset.In this research work, firstly, a fine-grained annotated dataset including instance and activity information has been developed for real-world volume crimes.Secondly, a BoFbased methodology has been introduced for effective training of the state-of-the-art 3D, and 2D Convolutional Neural Networks (CNNs).Lastly, a comparison between the state-of-the-art networks have been presented for malicious event recognition in videos.It has been observed that 2D CNN even with lesser parameters achieved a promising classification accuracy of 98.7% and Area under the curve (AUC) of 99.7%.

This research focuses on load-forecasting using Cartesian Genetic Programming evolved Artificial Neural Networks (CGPANN) and load-balancing using Genetic Algorithm in an electrical system. An unbalanced load in a distribution feeder has adverse effects on the system. All the transformer units connected to the feeder have different operating loads, and the system’s overall behaviour depends on them. Even if the transformers are not overloaded, any feeder phase can become overloaded due to excessive load contributed by individual transformers on that phase which results in a system-wide blackout. A custom-built monitoring device is installed on each transformer to monitor real-time electrical load data. A switching mechanism introduced at the transformer level can interchange the load between phases. This combination helps eliminate excessive load conditions and minimize unbalanced load conditions. Considering the involved parameters, i.e. transformers, phases, switching possibilities, and operating load, this constitutes a search problem within an available solution set. An optimum solution could be searched for and identified by interchanging the loads. This study aims to develop a feasible algorithm for such a search problem. The developed Genetic Algorithm can arrive at an optimum solution in minimum iterations. The load-forecasting model is used to predict load and identify system anomalies while the load-balancing model can re-adjust the system by shifting loads on individual transformers so as to balance the feeder’s overall load with no excessive load condition.

Standard method of assessing breast cancer is a triple test assessment. In this method, initially a thorough medical examination and patient history is evaluated, secondly imaging of the breast using x-rays and/or ultrasound is done and finally a preoperative cytodiagnosis is done that is either Fine Needle Aspiration Cytology (FNAC) or Core Needle Biopsy (CNB) or both. FNAC being a minimally invasive and rapidly performed test is preferred in many cases over CNB that is more invasive. If a triple test gives positive result in any one of the three steps then the result is taken positive. FNAC involves determining the cell size and shape parameters and based on their values a case is classified as benign or malignant. To automate the process of decision making a novel technique has been proposed. In this technique a digital logic circuit was evolved using Cartesian Genetic Programming (CGP). A CGP network was trained and then tested with FNAC feature data from the Breast Cancer Wisconsin Dataset. The dataset consists of 669 samples, of which 350 samples were used for training purposes and then the trained system was evaluated with 349 test samples. A number of experiments were performed, each with a different set of network parameters. The best evolved network classified the samples with an accuracy of 99.42%, which is higher than that produced with most of the contemporary techniques. The network so produced can be implemented on re-configurable hardware.

A smart city ensures the safety of its citizens by the reduction of crime and terror threats. Despite intensive efforts to prevent and control anomalous human activities, they still pose a major risk and challenge to the society. This paper presents an automatic recognition of unusual human behavior captured by a CCTV camera in public areas, using spatio-temporal 3D convolutional neural networks. The weakly labeled benchmark dataset has been properly annotated to remove noise for accurately localizing anomalies within videos. This human-related dataset with real crime scenes is then compared to other state-of-the-art techniques such as Pseudo 3D and ResNet 3D. Our experimental results on the newly developed dataset outperforms most competing models in terms of area under the curve (AUC), obtaining 97.39% AUC.

We present the 3DGAN for the simulation of a future high granularity calorimeter output as three-dimensional images. We prove the efficacy of Generative Adversarial Networks (GANs) for generating scientific data while retaining a high level of accuracy for diverse metrics across a large range of input variables. We demonstrate a successful application of the transfer learning concept: we train the network to simulate showers for electrons from a reduced range of primary energies, we then train further for a five times larger range (the model could not train for the larger range directly). The same concept is extended to generate showers for other particles depositing most of their energies in electromagnetic interactions (photons and neutral pions). In addition, the generation of charged pion showers is also explored, a more accurate effort would require additional data from other detectors not included in the scope of the current work. Our further contribution is a demonstration of using GAN-generated data for a practical application. We train a third-party network using GAN-generated data and prove that the response is similar to a network trained with data from the Monte Carlo simulation. The showers generated by GAN present accuracy within 10% of Monte Carlo for a diverse range of physics features, with three orders of magnitude speedup. The speedup for both the training and inference can be further enhanced by distributed training.

Predicting the Covid-19 spread and its impact on the stock market is an important research challenge these days. In order to obtain the best forecasting model, we have exploited neuro-evolutionary technique Cartesian genetic programming evolved artificial neural network (CGPANN) based solution to predict the future cases of COVID-19 up to 6-days in advance. This helps authorities and paramedical staff to take precautionary measures on time which helps in counteracting the spreading of the virus. The rising number of COVID cases has caused a significant impact on the stock market. CGPANN being the best performer for the time series prediction model seems ideal for the case under consideration. The proposed model achieved an accuracy as high as 98% predicting COVID-19 cases for the next six days. When compared with other contemporary models CGPANN seems to perform well ahead in terms of accuracy.

Floods are the world's most damaging natural disasters, which not only claim thousands of human lives but also result in huge damage to infrastructure. Floods if forecasted in advance can help in the reduction of damages. Flood prediction especially long term is a complex task as it involves many hydrological and metrological parameters. For the short and medium-term, machine learning methods seem to have contributed to a great extent in simulating mathematical modelling of the physical flow processes of floods. However, these developed model's performance lacks generalization. Such systems trained on one geographical location's data have degraded performance when exploited for another location. In this paper, Long Short-Term Memory (LSTM) machine learning algorithm was applied where the hourly river level, river flow, and rainfall data from Brooklyn station was used as input data to the model and test for one hour, two hours, four hours, six hours, eight hours, and twelve hours in advance for river level prediction at Hoppers Crossing station. The developed algorithm achieved an accuracy of 98% for one hour and 97.2 %, 96.14 %, 94.67%,94.61 %, and 93.55% for two, four, six, eight, and twelve hours respectively. These systems not only forecast the future water level but also help in estimating the water level in case of a sensor failure. Multivariate modelling is utilized to predict the unknown parameter from the given other parametric values, thus not only predicting the forecasted water level but also reporting the sensor failure.

Identification of exact malicious instant in lengthy CCTV recordings depends solely on Auto activity cognizance. The 3D CNN has previously been explored for the analysis of motion in video streams. Studies exhibit that, using separate filters for encoding spatial and temporal information has the same level of efficiency as that of 3D convolution filters. This study presents a novel approach through introduction of independent filters for event recognition in videos. This aims at learning extended Spatio-temporal features utilizing modified ResNet architecture. A novel 2D block termed as Quasi-3D (Q3D) decouples 3D information by combining 2D filters. The proposed Quasi-3D block encodes not only the spatial information in each frame but also the relative motion of objects along the x -axis and y -axis in a set of frames. Three variations of Quasi-3D block have been introduced to emphasize more on the features for further enhancing performance. A multi-class malicious activity recognition video dataset CrimesScene (drive:google:com/file/d/1omiQG9sxx375HjL97DqXxIX9nnfW3oQ/view?usp=sharing) inclusive of annotated video segments from 4 different classes of volume crimes has been developed. Results exhibit that the proposed Q3D ResNet model outperforms all other variants by achieving the overall detection accuracy of 94.9% and 93.07% on Hockey Fight and CrimesScene datasets, respectively.

Financial market volatility has become one of the most difficult applications for stock price forecasting in ongoing situations. The current statistical models for stock price forecasting are too rigid and inefficient to appropriately deal with the uncertainty and volatility inherent in stock data. CGPANN-CGP based ANNs and LSTM are the most common methods used these days to predict such dynamics in time series data. In comparison to other methodologies, studies have demonstrated that the application of Cartesian genetic programming evolved Artificial Neural Networks (CGPANNs) to time series forecasting problems produces better results, and LSTM can be competitive at times. CGPANN provides the ability to train both structure, topology, and weights of network to achieve the global optimum solution. The prediction model is trained on the behavior of stock exchange patterns and is based on trends in historical daily stock prices. The proposed CGPANN and LSTM models produced competitive results of 98.86% and 98.52% respectively. However, CGPANN architecture is capable computationally efficient than LSTM and its ability of quick predictions makes it ideal for real-time applications.

Particulate matter is one of the key contributors of air pollution and climate change. Long-term exposure to constituents of air pollutants has exerted serious health implications in both humans and plants leading to a detrimental impact on economy. Among the pollutants contributing to air quality determination, particulate matter has been linked to serious health implications causing pulmonary complications, cardiovascular diseases, growth retardation and ultimately death. In agriculture, crop yield is also negatively impacted by the deposition of particulate matter on stomata of the plant which is alarming and can cause food security concerns. The deleterious impact of air pollutants on human health, agricultural and economic well-being highlights the importance of quantifying and forecasting particulate matter. Several deterministic and deep learning models have been employed in the recent years to forecast the concentration of particulate matter. Among them, deep learning models have shown promising results when it comes to modeling time series data and forecasting it. We have explored recurrent neural networks with LSTM model which shows potential to predict the particulate matter (PM2.5) based on multi-step multi-variate data of two of the most polluted regions of South Asia, Beijing, China and Punjab, Pakistan effectively. The LSTM model is tuned using Bayesian optimization technique to employ the appropriate hyper-parameters and weight initialization strategies based on the dataset. The model was able to predict PM2.5 for the next hour with root-mean-square error (RMSE) of 0.1913 (91.5% accuracy) and this error gradually increases with the number of time steps with next 24 hours steps prediction having RMSE of 0.7290. While in case of Punjab dataset with data recorded once a day, the RMSE for the next day fore-cast is 0.2192. These multi-step short-term forecasts would play a pivotal role in establishing an early warning system based on the air quality index (AQI) calculated and enable the government in enacting policies to contain it.

Identification/recognition of assault, fighting, shooting, and vandalism from video sequence using deep 2D and 3D convolutional neural networks (CNNs) is explored in this paper. Recent wave of extensive unrestricted urbanization has not only uplifted the standard of living, but has also threatened the safety of a common man leading to an extraordinary rise in crime rate. Although Closed-circuit television (CCTV) footage provides a monitoring framework, yet, it’s useless without an auto volume crime detection system. The system proposed in this work is an effort to eradicate volume crimes through accurate detection in real-time. Firstly, a fine-grained annotated dataset including instance and activity information has been developed for real-world volume crimes. Secondly, a comparison between 3D CNN and 2D CNN network has been presented to identify the malicious event from the video sequence. This is carried out to explore the significance of spatial and temporal information present in the video for event recognition. It has been observed that 2D CNN even with lesser parameters achieved a promising classification accuracy of 91.2%and Area under the curve (AUC) of 95.2%on four classes. The system also reduces false alarm rate in comparison to state-of-the-art approaches.

Cloud computing use is exponentially increasing with the advent of industrial revolution 4.0 technologies such as the Internet of Things, artificial intelligence, and digital transformations. These technologies require cloud data centers to process massive volumes of workloads. As a result, the data centers consume gigantic amounts of electrical energy, and a large portion of data center electrical energy comes from fossil fuels. It causes greenhouse gas emissions and thus ensuing in global warming. An adaptive resource utilization mechanism of cloud data center resources is vital to get by with this huge problem. The adaptive system will estimate the resource utilization and then adjust the resources accordingly. Cloud resource utilization estimation is a two-fold challenging task. First, the cloud workloads are sundry, and second, clients’ requests are uneven. In the literature, several machine learning models have estimated cloud resources, of which artificial neural networks (ANNs) have shown better performance. Conventional ANNs have a fixed topology and allow only to train their weights either by back-propagation or neuroevolution such as a genetic algorithm. In this paper, we propose Cartesian genetic programming (CGP) neural network (CGPNN). The CGPNN enhances the performance of conventional ANN by allowing training of both its parameters and topology, and it uses a built-in sliding window. We have trained CGPNN with parallel neuroevolution that searches for global optimum through numerous directions. The resource utilization traces of the Bitbrains data center is used for validation of the proposed CGPNN and compared results with machine learning models from the literature on the same data set. The proposed method has outstripped the machine learning models from the literature and resulted in 97% prediction accuracy.

Systems based on physiological biometrics are ubiquitous but requires subject cooperation or high resolution to capture. Gait recognition is a great avenue for identification and authentication due to uniqueness of individual stride in an un-intrusive manner. Machine vision systems have been designed to capture the uniqueness of stride of a specific person but factors such as change in speed of stride, view point, clothes and carrying accessories make gait recognition challenging and open to innovation. Our proposed approach attempts to tackle these problems by capturing the spatio-temporal features of a gait sequence by training a 3D convolutional deep neural network (3D CNN). The proposed 3D CNN architecture tackles gait identification by employing holistic approach in the form of gait energy images (GEI) which is a condensed representation capturing the shape and motion characteristics of the the human gait. The network was evaluated on two of the largest publicly available datasets with substantial gender and age diversity; OULP and CASIA-B. Optimization strategies were explored to tune the hyper-parmeters and improve the performance of the 3D CNN network. The optimized 3D CNN and the GEI were effectively able to capture the unique characteristics of the gait cycle of an individual irrespective of the challenging covariates. State of the art results achieved on the multi-views and multiple carrying conditions of the subjects belonging to CASIA-B dataset demonstrating the efficacy of our proposed algorithm.

Emotion recognition from facial expression is an exciting field of research with applications like safety, security, personal information and marketing. Researchers want to develop techniques that can interpret, and extract facial expressions so that computers can make better emotional predictions. In recent years, different types of architectures have been used in machine learning to improve facial expression performance. In this paper, machine learning techniques are used to study facial emotion recognition. We present various machine learning techniques to identify the best methodology for the test at hand. Support Vector Machine (SVM), Convolution Neural Network (CNN) and Artificial Neural Network (ANN) along with face detection and preprocessing techniques for the expressions in Japanese Female Facial Expression (JAFFE) dataset and the Extended Cohn-Kanade (CK+) dataset are exploited to achieve best accuracy of 98.47% on CK+ dataset using CNN, and 89.18% accuracy for JAFFE dataset using ANN.

A multi-chromosomal structure for neuro-evolution of recurrent neural networks (RNNs) with Cartesian genetic programming (CGP) architecture is presented to develop a signal reconstruction model. The group behaviour of multi-chromosomes evolved together is explored in terms of computational and structural complexity, and accurate reconstruction of the corrupted signal. Signal reconstruction modelling is done through prediction of missing samples, and a sliding window mechanism is utilized to recover long segments of missing samples in periodic and aperiodic corrupted signals. The proposed model is compared with the conventional single chromosome-based Cartesian Genetic Programming-evolved Recurrent Neural Network (CGPRNN) architecture as well as state-of-the-art machine learning algorithms, i.e. linear regression, random forest and long short-term memory deep learning neural network. Diverse network architectures of the proposed Multi-Chromosomal Cartesian Genetic Programming-evolved Recurrent Neural Network (MC-CGPRNN) are evolved to explore the robustness of algorithm in terms of optimization of hyper parameters, evolutionary speed and generalization ability. The proposed algorithm is trained and tested on speech signal with both periodic and aperiodic noise. The best trained network is also evaluated on standard music signals achieving remarkable results. Experimental results show that the proposed MC-CGPRNN algorithm achieved SNR of 25.23 dB for speech signal, 29.92 dB for guitar signal and 28.45 dB for flute signal in case of 50% missing samples showing best SNR improvement in comparison with its counterparts for time domain signal reconstruction.

Safe cities initiative promises the eradication of crimes specially the volume crimes which constitute fighting, shooting, and vandalism. This paper focuses on the detection of real-world volume crimes in a video using spatio-temporal convolutional neural network. Firstly, a benchmark dataset has been annotated on the fine-grained level to ensure reliable training. Secondly, a comparison is made between two state of the art networks recently used for spatio temporal analysis, over the developed dataset. Also, detailed analysis has been performed for detection and classification over the two networks. From the comparison it is concluded that the performance of C3D model is better than ResNet3D for binary classification. Whereas, ResNet3D outperforms C3d incase of multi-class classification.

Landslides are the calamities responsible for the losses of lives and property at large. Early warning systems to predict the upcoming landslide can be of great advantage in developing a guide to control and alleviate these disasters. In this research, we have used artificial neural network (ANN) and long short term neural network (LSTM) for landslide prediction in Pakistan. The data has been collected from various regions of Pakistan including Jammu and Kashmir, Chitral, Gilgit Baltistan and Hazara division. A total of 12 parameters: surface pressure, specific humidity, surface runoff, rainfall, soil moisture at root level, soil moisture at 10cm, average temperature, transpiration, wind speed, aspect, slope and elevation are extracted for each location. The developed models have achieved accuracy greater than 95%.

Generative Adversarial Networks (GANs) have gained notoriety by generating highly realistic images. The present work explores GAN for simulating High Energy Physics detectors, interpreting detector output as three-dimensional images. The demands and requirements of a scientific simulation are quite stringent, as compared to the domain of visual images. Image characteristics such as pixel intensity and sparsity, for example, have very different distributions. Moreover, detector simulation requires conditioning on physics inputs, and domain knowledge becomes essential. We, therefore, adjust the pre-processing and incorporate physics-based constraints in the loss function. We also introduce a multi-step training process based on transfer learning by breaking up the task complexity. Validation of the results primarily consists of a detailed comparison to full Monte Carlo in terms of several physics quantities where a high level of agreement is found (ranging from a few percent up to 10% across a large particle energy range). In addition, we assess the performance by physics unrelated metrics, thereby proving further the variability and pertinence through diverse standpoints. We have demonstrated that an image generation technique from vision can successfully simulate highly complex physics processes while achieving a speedup of more than three orders of magnitude in comparison to the standard Monte Carlo.