Continual learning (CL), a.k.a lifelong learning, is an emerging research topic that has been attracting increasing interest in the field of machine learning. With human activity recognition (HAR) playing a key role in enabling numerous real-world applications, an essential step towards the long-term deployment of such systems is to extend the activity model to dynamically adapt to changes in people’s everyday behavior. Current research in CL applied to HAR domain is still under-explored with researchers exploring existing methods developed for computer vision in HAR. Moreover, analysis has so far focused on task-incremental or class-incremental learning paradigms where task boundaries are known. This impedes the applicability of such methods for real-world systems. To push this field forward, we build on recent advances in the area of continual learning and design a lifelong adaptive learning framework using Prototypical Networks, LAPNet-HAR, that processes sensor-based data streams in a task-free data-incremental fashion and mitigates catastrophic forgetting using experience replay and continual prototype adaptation. Online learning is further facilitated using contrastive loss to enforce inter-class separation. LAPNet-HAR is evaluated on 5 publicly available activity datasets in terms of its ability to acquire new information while preserving previous knowledge. Our extensive empirical results demonstrate the effectiveness of LAPNet-HAR in task-free CL and uncover useful insights for future challenges.

Modern companies accumulate a vast amount of customer data that can be used for creating a personalized experience. Analyzing this data is difficult and most business intelligence tools cannot cope with the volume of the data. One example is churn prediction, where the cost of retaining existing customers is less than acquiring new ones. Several data mining and machine learning approaches can be used, but there is still little information about the different algorithm settings to be used when the dataset doesn't fit into a single computer memory. Because of the difficulties of applying feature selection techniques at a large scale, Incremental Probabilistic Component Analysis (IPCA) is proposed as a data preprocessing technique. Also, we present a new approach to large scale churn prediction problems based on the mini-batch Stochastic Gradient Decent (SGD) algorithm. Compared to other techniques, the new method facilitates training with large data volumes using a small memory footprint while achieving good prediction results.

The growing amount of data demands methods that can gradually learn from new samples. However, it is not trivial to continually train a network. Retraining a network with new data usually results in a known phenomenon, called “catastrophic forgetting.” In a nutshell, the performance of the model drops on the previous data by learning from the new instances. This paper explores this issue in the table detection problem. While there are multiple datasets and sophisticated methods for table detection, the utilization of continual learning techniques in this domain was not studied. We employed an effective technique called experience replay and performed extensive experiments on several datasets to investigate the effects of catastrophic forgetting. Results show that our proposed approach mitigates the performance drop by 15 percent. To the best of our knowledge, this is the first time that continual learning techniques are adopted for table detection, and we hope this stands as a baseline for future research.

Deep learning models are part of the family of artificial neural networks and, as such, it suffers of catastrophic interference when they learn sequentially. In addition, most of these models have a rigid architecture which prevents the incremental learning of new classes. To overcome these drawbacks, in this article we propose the Self-Improving Generative Artificial Neural Network (SIGANN), a type of end-to-end Deep Neural Network system which is able to ease the catastrophic forgetting problem when leaning new classes. In this method, we introduce a novelty detection model to automatically detect samples of new classes, moreover an adversarial auto-encoder is used to produce samples of previous classes. This system consists of three main modules: a classifier module implemented using a Deep Convolutional Neural Network, a generator module based on an adversarial autoencoder; and a novelty detection module, implemented using an OpenMax activation function. Using the EMNIST data set, the model was trained incrementally, starting with a small set of classes. The results of the simulation show that SIGANN is able to retain previous knowledge with a gradual forgetfulness for each learning sequence. Moreover, SIGANN can detect new classes that are hidden in the data and, therefore, proceed with incremental class learning.

Computers have evolved over the years and as the evolution continues, we have been ushered into an era where high-speed internet has made it possible for devices in our homes, hospital, energy and industry to communicate with each other. This era is what is known as the Internet of Things (IoT). IoT has several benefits in the health, energy, transportation and agriculture sectors of a country’s economy. These enormous benefits coupled with the computational constraint of IoT devices which makes it difficult to deploy enhanced security protocols on them make IoT devices a target of cyber-attacks. One approach that has been used in traditional computing over the years to fight cyber-attacks is Intrusion Detection System (IDS). However, it is practically impossible to deploy IDS meant for traditional computers in IoT environments because of the computational constraint of these devices. In this regard, this study proposes a lightweight IDS for IoT devices using an incremental ensemble learning technique. We used Gaussian Naive Bayes and Hoeffding tree to build our incremental ensemble model. The model was then evaluated on the TON IoT dataset. Our proposed model was compared with other state-of-the-art methods proposed and evaluated using the same dataset. The experimental results show that the proposed model achieved an average accuracy of 99.98\%. We also evaluated the memory consumption of our model which showed that our model achieved a lightweight model status of 650.11KB as the highest memory consumption and 122.38KB as the lowest memory consumption.

The use of biodegradable materials has a growing field of application due to environmental concerns, however, scientific research on incremental forming using biomaterials is scarce. Thus, this study focuses on the single point incremental forming (SPIF) process applied to a composite sheet that combines a biodegradable thermoplastic matrix (Solanyl) reinforced with natural fibres (flax). The influence of the process parameters on the final geometry is determined, evaluating the effect of the following factors: step depth, wall angle and temperature reached during the process. Additionally, a heated aqueous medium is incorporated which facilitates the formability of the composite sheets. This method is especially useful for materials that have poor formability at room temperature. The benefits of using controlled heat include the reduction of formation forces applied to the plate, improved accuracy due to the reduction of elastic recovery, and the manipulation of the samples remarkably close to the glass transition temperatures. Through this experimental study with the variables analysed, a maximum shaping depth of 310 mm is obtained. These results confirm that the single point shaping used with bioplastic materials is possible and has positive outcomes for incremental forming.

We investigated effects of molecular hydrogen (H2) supplementation on acid-base status, pulmonary gas exchange responses, and local muscle oxygenation during incremental exercise. Eighteen healthy, trained subjects in a randomized, double-blind, crossover design received H2-rich calcium powder (HCP) (1500 mg/day, containing 2.544 µg/day of H2) or H2-depleted placebo (1500 mg/day) for 3 consecutive days. They performed cycling incremental exercise starting at 20-watts work rate, increasing by 20 watts/2 min until exhaustion. Breath-by-breath pulmonary ventilation (VE) and CO2 output (VCO2) were measured and muscle deoxygenation (deoxy[Hb + Mb]) was determined via time-resolved-NIRS in the vastus lateralis (VL) and rectus femoris (RF). Blood gases' pH, lactate, and HCO3− concentrations were measured at rest and 120-, 200-, and 240-watt work rates. At rest, the HCP group had significantly lower VE, VCO2, and higher HCO3−, PCO2 versus placebo. During exercise, a significant pH decrease and greater HCO3− continued until 240-watts work rate in HCP. The VE was significantly lower in HCP versus placebo, but HCP did not affect the gas exchange status of VCO2 or oxygen uptake (VO2). HCP increased absolute values of deoxy[Hb + Mb] at the RF but not VL. Thus, HCP-induced hypoventilation would lead to lower pH and secondarily impaired balance between O2 delivery and utilization in the local RF during exercise, suggesting that HCP supplementation, which increases the at-rest antioxidant potential, affects the lower ventilation and pH status during incremental exercise. HPC induced a significantly lower O2 delivery/utilization ratio in the RF but not the VL, which may be because these regions possess inherently different vascular/metabolic control properties, perhaps related to fiber-type composition.

Advanced manufacturing techniques, aimed at implants with high dependability, flexibility, and low manufacturing costs, are crucial in meeting the growing demand for high-quality products like biomedical implants. Incremental sheet forming is a promising flexible manufacturing approach for rapidly prototyping sheet metal components using low-cost tools. Titanium and its alloys are used to shape most biomedical implants because of its superior mechanical qualities, biocompatibility, low weight, and great structural strength. The poor formability of titanium sheets at room temperature, however, limits their widespread use. The goal of this research is to show that gradual sheet formation of a titanium biomedical implant is possible. The possibility of creative and cost-effective concepts for the manufacturing of such complicated shapes with significant wall angles is explored in this study. A numerical simulation based on finite element modeling as well as a design process tailored to metal forming is used to complete the development. The mean of uniaxial tensile tests with a constant strain rate was used to study the flow behavior of the studied material. To forecast the crack, the obtained flow behavior was modeled using the behavior model and failure model.

The study is to identify motivational factors that lead to collective intelligence and to understand how these factors relate to each other and to innovation capabilities in enterprises. The relationships between each of the sub-factors of the collective intelligence construct with the sub-factor of incremental innovation were examined. The study used the convenience sampling of corporate employees who use collective intelligence from corporate panel members (n=1500). Collective intelligence was found to affect work process, operations, and service innovation. This suggests that as work processes are made more innovative, the more actively collective intelligence is pursued, the greater the improvement in the performance of work processes, work procedures, work efficiency, customer satisfaction, and services. This study provides significant implications for corporations operating collective intelligence services such as online communities. First, such corporations vitalize their services by raising the quality of information and knowledge shared in their communities. Additionally, contribution motivations that take the characteristics of knowledge and information contributors into consideration require further development.The sample for this study was identified through convenience sampling of corporate employees who use collective intelligence from corporate panel members (n=1500). Collective intelligence was found to affect work process, operations, and service innovation. This suggests that as work processes are made more innovative, the more actively collective intelligence is pursued, the greater the improvement in the performance of work processes, work procedures, work efficiency, customer satisfaction, and services. This study provides significant implications for corporations operating collective intelligence services such as online communities. First, such corporations vitalize their services by raising the quality of information and knowledge shared in their communities. Additionally, contribution motivations that take the characteristics of knowledge and information contributors into consideration require further development.

Fast 3D reconstruction with semantic information on road scenes is of great requirements for autonomous navigation. It involves issues of geometry and appearance in the field of computer vision. In this work, we propose a method of fast 3D semantic mapping based on the monocular vision. At present, due to the inexpensive price and easy installation, monocular cameras are widely equipped on recent vehicles for the advanced driver assistance and it is possible to acquire semantic information and 3D map. The monocular visual sequence is used to estimate the camera pose, calculate the depth, predict the semantic segmentation, and finally realize the 3D semantic mapping by combination of the techniques of localization, mapping and scene parsing. Our method recovers the 3D semantic mapping by incrementally transferring 2D semantic information to 3D point cloud. And a global optimization is explored to improve the accuracy of the semantic mapping in light of the spatial consistency. In our framework, there is no need to make semantic inference on each frame of the sequence, since the mesh data with semantic information is corresponding to sparse reference frames. It saves amounts of the computational cost and allows our mapping system to perform online. We evaluate the system on naturalistic road scenes, e.g., KITTI and observe a significant speed-up in the inference stage by labeling on the mesh.

Optimism and self-efficacy have been associated with psychological health. Empathy has also been found to promote positive functioning and to have a unique role in community health volunteering. This study investigated whether self-efficacy and optimism were associated with psychological and subjective well-being in a group of healthcare volunteers and whether empathy added incrementally to these associations. A sample of 160 Italian clown doctors volunteering in various hospitals completed self-report measures of self-efficacy, optimism, empathy, psychological well-being, and subjective well-being. Results indicated that self-efficacy and optimism were associated with both outcomes and that aspects of empathy, such as others’ perspective-taking and personal distress for others’ difficulties, incrementally added to these associations, although with opposite effects. The present study adds to previous research on the role of self-efficacy, optimism, and empathy for community health volunteers’ psychological health and offers suggestions regarding the training of this type of volunteer.

The frequent incidents of oil spills and other forms of pollution arising from crude oil exploration and exploitation (OEE) in the Niger Delta have caused several investigations on Polycyclic Aromatic Hydrocarbons (PAHs) pollution. This study aimed at developing a comprehensive report on PAH pollution and its human health risks recorded in the Niger Delta. Studies were extracted from Google Scholar, PubMed, and ResearchGate using a defined selection criterion. The quality of each study was assessed using the Newcastle – Ottawa Scale. Thirty-eight studies were selected with the majority reporting on PAH pollution in aquatic environments. Across all the selected studies, the total number of PAHs recorded ranged from 7 to 28 PAH congeners. Also, PAH potential sources reported in the studies were of pyrogenic and petrogenic sources. PAH concentrations recorded in water, sediment, aquatic organisms (fish and shrimp), soil, dust, and crop samples ranged from below detection limit (BDL) to 450 ± 117.9 mg/L, BDL to 1821.5 mg/kg, 0.005 to 1.098 mg/kg, ND to 4154 ± 3461 mg/kg, 165.1 to 1012 mg/kg, and 0.020 to 3.37 mg/kg, respectively. Majority of the selected studies reported PAH levels which were higher than the permissible limits. Incremental Lifetime Cancer Risk (ILCR) assessment of PAHs in samples ranged from low to high via ingestion and dermal routes of exposure to humans. It is recommended that the Federal Government of Nigeria promotes environmentally friendly operations of OEE. Future studies should focus on PAH pollution in farmlands, ambient air and the associated human and ecological health risks.

Polycyclic aromatic hydrocarbons (PAHs) in street dust pose a serious problem threatening both environment and human health. Street dust were collected from five different land use patterns (traffic areas TRA, urban area URA, residential areas REA, mixed residential commercial areas MCRA and suburban areas SUA) in a Saudi coastal city, Jeddah, and one in rural area (RUA) in Hada Al Sham. This study aimed to investigate the status, profile, sources of PAHs and estimate their human health risk. The results revealed an average concentration of total PAHs of 3320 ng/g in street dust of Jeddah and 223 ng/g in RUA dust. PAHs with high molecular weight represented 83.38% of total PAHs in street dust of Jeddah, while the carcinogenic PAH compounds accounted 57.84%. The highest average concentration of total PAHs in street dust of Jeddah was found in TRA (4980 ng/g) and the lowest in REA (1660 ng/g). PAHs ratios indicated that the principal source of PAHs in street dust of Jeddah is pyrogenic, mainly traffic emission. Benzo(a)anthracene/ chrysene (BaA/CHR) ratio suggests that PAHs in street dusts of Jeddah come mainly from emission of local sources, while PAHs in RUA might be transported from the surrounding urban areas. The estimated Incremental Lifetime Cancer Risk (ILCR) associated with exposure to PAHs in street dusts indicated that both dermal contact and ingestion pathways are major contributed to cancer risk for both children and adults. Based on BaP_equivalence concentrations of total PAHs, ILCR_Ingestion, ILCR_dermal and cancer risk values for children and adults exposed to PAHs in street dust of different areas in Jeddah were found between 10⁻⁶ and 10⁻⁴, indicating potential risk. The sequence of cancer risk was TRA > URA > MCRA > SUA > REA. Only exposure to BaP and DBA compounds had potential risk for both children and adults.

This work presents a control scheme to control the grid-connected single-phase photovoltaic (PV) system. The considered system has four 250W solar panels, a non-inverting buck-boost DC-DC converter, and DC-AC inverter with LCL filter. The control system aims to track and operate at the maximum power point (MPP) of PV panels, regulate the voltage of DC link, and supply the grid with a unity power factor. To well achieve these goals, the proposed control system consists of three parts, that are MPP tracking controller module with a fuzzy-based modified incremental conductance (INC) algorithm, a DC-link voltage regulator with a hybrid fuzzy proportional-integral (PI) controller, and a Current Controller module using the linear quadratic regulator (LQR) for grid-connected power. Based on fuzzy control and LQR, this work introduces a full control solution for grid-connected single-phase PV systems. The key novelty of this research is to analyze and prove that the newly proposed method is more successful in numerous aspects by comparing and evaluating the previous and present control methods. The designed control system settles quickly, which is critical for output stability. In addition, as compared to backstepping approach used in our past study, the LQR technique is more resistant to sudden changes and disturbances. Furthermore, backstepping method produces the larger overshoot, which has a detrimental impact on efficiency. Simulation findings under various weather conditions were compared to theoretical ones to indicate that the system can deal with variations in weather parameters.

To implement renewable energy resources, microgrid systems have been adopted and developed into the technology of choice to assure mass electrification in the next decade. Microgrid systems have a number of advantages over the conventional utility grid systems, however, it faces severe instability issues due to continually increasing constant power loads. To improve the stability of the entire system, load side compensation technique is chosen because of its robustness and cost effectiveness. In this particular occasion, a sliding mode controller is developed for microgrid system in the presence of CPL to assure certain control objective of keeping the output voltage constant at 480V. After that, the robustness analysis of the sliding mode controller against parametric uncertainties is presented. The sliding mode controller robustness against parametric uncertainties, frequency variations, and additive white Gaussian noise (AWGN) are illustrated in this paper. Later, the performance of the PID and sliding Mode controller is compared in case of nonlinearity, parameter uncertainties, and noise rejection to justify the selection of Sliding Mode controller over PID controller. All the necessary calculations are reckoned mathematically and results are verified in the virtual platform such as MATLAB/Simulink with the appreciable outcome.

Studies employing Arellano-Bond and Blundell-Bond GMM estimation for single linear dynamic panel data models are growing exponentially in number. However, for researchers it is hard to make a reasoned choice between many different possible implementations of these estimators and associated tests. By simulation the effects are examined of many options regarding: (i) reducing, extending or modifying the set of instruments; (ii) specifying the weighting matrix in relation to the type of heteroskedasticity; (iii) using (robustified) 1-step or (corrected) 2-step variance estimators; (iv) employing 1-step or 2-step residuals in Sargan-Hansen overall or incremental overidentification restrictions tests. This is all done for models in which some regressors may be either strictly exogenous, predetermined or endogenous. Surprisingly, particular asymptotically optimal and relatively robust weighting matrices are found to be superior in finite samples to ostensibly more appropriate versions. Most of the variants of tests for overidentification restrictions show serious deficiencies. A recently developed modification of GMM is found to have great potential when the cross-sectional heteroskedasticity is pronounced and the time-series dimension of the sample not too small. Finally all techniques are employed to actual data and lead to some profound insights.

The In this paper, a universal method has been developed, which is a combination of an optimization method and a method for assessing the marginal utility. At present, the problem of optimal load distribution in the power system between a hydropower plant (HPP) and thermal power plants (TPP) is solved using the equality of the differential incremental rate characteristics of fuel consumption at TPP and water consumption at HPP by the Lagrangian multiplier method. In this case, the number of iterations can be five or more. The proposed approach is based, first of all, on the correct representation of the differential characteristics and calculation of a hydro resource price for the operational control of the HPP. Based on the comparison of water volume at a HPP and fuel amount at combined heat and power plants (CHPP) used for generation of 1 kW power, it is possible to determine a water price for a HPP. As a result of implementing the developed method for the HPP, a price of sold electricity in the flexible energy market will be comparable with the price for sold electricity produced at CHPPs, being equal to approximately 120 rubles/MW∙h.

A large number of bridges are designed and built without considering seismic actions and, differently from buildings, there are currently no comprehensive guidelines to evaluate existing bridges without performing, as in the well known incremental dynamic analysis (IDA), complex non linear dynamic analyses (RHA). Bridges are structurally very different from building but, at the same time, are sensitive to higher modes as well as many multi-storey buildings that inspired innovative pushover procedures such as the well known modal pushover analysis (MPA). In the present study the incremental modal pushover analysis (IMPA), a pushover based approach already proposed and applied on buildings by the same authors, is revised and proposed for bridges (IMPAβ). IMPAβ accounts for the effects of higher modes in order to accurately estimate the seismic response of bridges; the effect of higher modes is considered by introducing a suitable number of modes to ensure the participation of a predefined total effective modal mass. The efficiency of the proposed method is demonstrated by conducting a study on two bridges, one regular and one irregular, and the IDA analysis is employed as reference solution. Numerical results indicate good accuracy of the proposed method in assessing the seismic response and a very good accuracy if compared to other available pushover procedures available in the literature.