Search Results (47)

Symbolic regression plays a crucial role in machine learning and data science by allowing the extraction of meaningful mathematical models directly from data without imposing a specific structure. This level of adaptability is especially beneficial in scientific and engineering fields, where comprehending and articulating the underlying data relationships is just as important as making accurate predictions. Genetic Programming (GP) has been extensively utilized for symbolic regression and has demonstrated remarkable success in diverse domains. However, GP’s heavy reliance on evolutionary mechanisms makes it computationally intensive and challenging to handle. On the other hand, Particle Swarm Optimization (PSO) has demonstrated remarkable performance in numerical optimization with parallelism, simplicity, and rapid convergence. These attributes position PSO as a compelling option for Automatic Programming (AP), which focuses on the automatic generation of programs or mathematical models. Particle Swarm Programming (PSP) has emerged as an alternative to Genetic Programming (GP), with a specific emphasis on harnessing the efficiency of PSO for symbolic regression. However, PSP remains unsolved due to the high-dimensional search spaces and local optimal regions in AP, where traditional PSO can encounter issues such as premature convergence and stagnation. To tackle these challenges, we introduce Dynamical Sphere Regrouping PSO Programming (DSRegPSOP), an innovative PSP implementation that integrates DSRegPSO’s dynamical sphere regrouping and momentum conservation mechanisms. DSRegPSOP is specifically developed to deal with large-scale, high-dimensional search spaces featuring numerous local optima, thus proving effective behavior for symbolic regression tasks. We assess DSRegPSOP by generating 10 mathematical expressions for mapping points from functions with varying complexity, including noise in position and cost evaluation. Moreover, we also evaluate its performance using real-world datasets. Our results show that DSRegPSOP effectively addresses the shortcomings of PSO in PSP by producing mathematical models entirely generated by AP that achieve accuracy similar to other machine learning algorithms optimized for regression tasks involving numerical structures. Additionally, DSRegPSOP combines the benefits of symbolic regression with the efficiency of PSO. Full article

It is still largely unknown to what extent domestication, ancestry, or recent functional selection are responsible for the behavioral differences in whether dogs look back to a human when presented with a difficult task. Here, we tested whether this ubiquitous human-related response of companion dogs would appear differently in subjects that were selected for either cooperative or independent work tasks. We tested N = 71 dogs from 18 cooperative and 18 independent breeds. Subjects learned in a five-trial warming-up phase that they could easily obtain the reward from a container. In trial six, the reward became impossible to take out from the locked container. When the task was easy, both breed groups behaved similarly, and their readiness to approach the container did not differ between the last ‘solvable’ and the subsequent ‘unsolvable’ trial. Task focus, looking at the container, touching the container for the first time, or interacting with the container with a paw or nose did not differ between the breed groups, indicating that their persistence in problem solving was similar. However, in the ‘unsolvable’ trial, cooperative dogs alternated their gaze more often between the container and the humans than the independent dogs did. The frequency of looking back was also higher in cooperative dogs than in the independent breeds. These are the first empirical results that suggest, in a balanced, representative sample of breeds, that the selection for different levels of cooperativity in working dogs could also affect their human-dependent behavior in a generic problem-solving situation. Full article

Modern convolutional neural networks (CNNs) play a crucial role in computer vision applications. The intricacy of the application scenarios and the growing dataset both significantly raise the complexity of CNNs. As a result, they are often overparameterized and have significant computational costs. One potential solution for optimizing and compressing the CNNs is to replace convolutional layers with low-rank tensor decomposition. The most suitable technique for this is Canonical Polyadic (CP) decomposition. However, there are two primary issues with CP decomposition that lead to a significant loss in accuracy. Firstly, the selection of tensor ranks for CP decomposition is an unsolved issue. Secondly, degeneracy and instability are common problems in the CP decomposition of contractional tensors, which makes fine-tuning the compressed model difficult. In this study, a novel approach was proposed for compressing CNNs by using CP decomposition. The first step involves using the sensitivity of convolutional layers to determine the tensor ranks for CP decomposition effectively. Subsequently, to address the degeneracy issue and enhance the stability of the CP decomposition, two novel techniques were incorporated: optimization with sensitivity constraints and iterative fine-tuning based on sensitivity order. Finally, the proposed method was examined on common CNN structures for image classification tasks and demonstrated that it provides stable performance and significantly fewer reductions in classification accuracy. Full article

This paper aims to present multisensory spatial analysis (MSA). The method was designed for the quick, simultaneous identification of concrete cover thickness h, rebar diameter, and alloys of reinforcement in large areas of reinforced concrete (RC) structures, which is a complex and unsolved issue. The main idea is to divide one complex problem into three simple-to-solve and based on separate premises tasks. In the transducers designed with the MSA, sensors are arranged spatially. This arrangement identifies each RC parameter separately based on the different waveforms/attributes. The method consists of three steps. All steps are described in the paper and supported by simulations and statistical analysis of the measurement. The tests were carried out using an Anisotropic Magneto-resistance (AMR) sensor. The AMR sensors can measure strong DC magnetic fields and can be combined in spatial transducers because of their small size. The selection of the sensor was extensively justified in the introduction section. The spatial transducer and the identification’s simplicity can allow for high accuracy in the real-time area testing of all three parameters. The risk of misclassification of discrete parameters was strongly reduced, and the h parameter can be identified with millimeter accuracy. Full article

Crime scene investigation (CSI) is the complex act of reconstructing the dynamics that led to a crime and the circumstances of its perpetration. Crystallizing the CSI is a difficult task for the forensic pathologist; however, it is often requested by the public prosecutor and many judicial cases remain unsolved precisely for this reason. Recent years have seen an improvement in the ability of 3D scanning technology to obtain dense surface scans of large-scale spaces, for surveying, engineering, archaeology, and medical purposes such as forensics. The applications of this new technology are growing every day: forensic measurement of wounds in clinical reports, for example, reconstruction of traffic accidents, bullet trajectory studies in gunshot wounds, and 3D bloodstain pattern analysis. A retrospective analysis was conducted across all crime scene investigations performed by the forensic staff of the Department of Forensic Pathology of the University of Catania from January 2019 to June 2022. Inclusion criteria were the use of a laser scanner (LS), the use of a camera, a full investigative scene, and collection of circumstantial data thanks to the help of the judicial police. Cases in which the LS was not used were excluded. Out of 200 CSIs, 5 were included in the present study. In case number 1, the use of the LS made it possible to create a complete scale plan of the crime scene in a few hours, allowing a ship to be quickly returned to the judicial police officer. In case 2 (fall from a height), the LS clarified the suicidal intent of the deceased. In case number 3 it was possible to reconstruct a crime scene after many years. In case 4, the LS provided a great contribution in making a differential diagnosis between suicide and homicide. In case 5, the LS was fundamental for the COVID team in planning the study of COVID-19 pathways and areas within a hospital with the aim of reduction of nosocomial transmission. In conclusion, the use of the LS allowed the forensic staff to crystallize the investigative scene, making it a useful tool. Full article

Efficient control of tunnel boring machine (TBM) tunneling along the designed tunnel axis in an unknown variable geological environment is a difficult and significant task. At present, the TBM attitude during tunneling is mostly manually controlled based on the deviation between the tunneling axis and the designed tunnel axis and their experiences. The tunneling axis from manual control is often the snakelike motion around the designed tunnel axis, even exceeding the deviation limit, for which this paper analyzed three reasons, the unknown geological environment, the hysteresis of TBM position response, and the unsolved overall optimization of tunneling axis. For these reasons, this paper proposed a real-time optimal control framework of TBM attitude based on reinforcement learning, which contains the geological information predictive model, TBM attitude and position (TBMAP) predictive model, and optimal attitude control policy (OACP). This framework can predict the current geological information in real-time and provide the corresponding real-time optimal attitude control that simultaneously considers the hysteresis of TBM position response and the overall optimization of the tunneling axis. This attitude control framework can be directly deployed to TBM without increasing costs and excessive modifications to the equipment. To verify the effectiveness of this attitude control framework, the Xinjiang Yiner Water Supply Phase II Project, using the TBM method, was adopted as a case study. The results revealed that the accuracy of geological environment recognition reached 94%, and OACP can significantly reduce the accumulated deviation of the tunneling axis from the designed tunnel axis by over 80% compared with the manual control and easily provide real-time decision support for attitude control in actual engineering. Full article

Cat welfare is a topic of growing interest in the scientific literature. Although previous studies have focused on the effects of living style (i.e., indoor/outdoor) on cat welfare, there has been a noticeable dearth of analysis regarding the impact of lifestyle on cats’ inclination and mode of communication with humans. Our research aimed to analyze the possible effect of lifestyle (e.g., living indoors only or indoor/outdoor) on cat–human communication. The cats were tested using the impossible task paradigm test, which consists of some solvable trials in which the subject learns to obtain a reward from an apparatus, followed by an impossible trial through blocking the apparatus. This procedure triggers a violation of expectations and is considered a useful tool for assessing both the decision-making process and the tendency to engage in social behaviors towards humans. A specific ethogram was followed to record the behavioral responses of the cats during the unsolvable trial. Our results show the effects of lifestyle and age on domestic cats, providing valuable insights into the factors that influence their social behaviors. Cats that can roam freely outdoors spent less time interacting with the apparatus compared to indoor-only cats. Additionally, roaming cats showed stress behaviors sooner following the expectancy of violation compared to indoor cats. The lifestyle of cats can influence their problem-solving approach while not affecting their willingness to interact with humans or their overall welfare. Future studies on this topic can be useful for improving the welfare of domestic cats. Full article

Although procrastination has been extensively studied, precrastination remains an unsolved puzzle. Precrastination is the tendency to start tasks as soon as possible, even at the cost of extra effort. Using the near bucket paradigm with 81 undergraduate students, this study examined the relationship between precrastination and time perspective, proactive personality, and subjects’ differential performance in intertemporal decision-making. The results confirmed the cognitive-load-reduction (CLEAR) hypothesis. Precrastination was found to be positively predicted by the future time dimension of time perspective and negatively predicted by proactive personality. In addition, there is a significant positive correlation between precrastination and delay discounting of intertemporal decision-making, which exists only for the loss situation. Full article

Question Classification (QC) is the fundamental task for Question Answering Systems (QASs) implementation, and is a vital task, as it helps in identifying the question category. It plays a big role in predicting the answer to a question while building a QAS. However, classifying medical questions is still a challenging task due to the complexity of medical terms. Many researchers have proposed different techniques to solve these problems, but some of these problems remain partially solved or unsolved. With the help of deep learning technology, various text-processing problems have become much easier to solve. In this paper, an improved deep learning-based model for Medical Forum Question Classification (MFQC) is proposed to classify medical questions. In the proposed model, feature representation is performed using Word2Vec, which is a word embedding model. Additionally, the features are extracted from the word embedding layer based on Convolutional Neural Networks (CNNs). Finally, a Bidirectional Long Short Term Memory (BiLSTM) network is used to classify the extracted features. The BiLSTM model analyzes the target information of the representation and then outputs the question category via a SoftMax layer. Our model achieves state-of-the-art performance by effectively capturing semantic and syntactic features from the input questions. We evaluate the proposed CNN-BiLSTM model on two benchmark datasets and compare its performance with existing methods, demonstrating its superiority in accurately categorizing medical forum questions. Full article

Coherence change detection (CCD) is a remote sensing technique used to map phenomena that, under certain conditions, can be directly related to changes in Interferometric SAR (InSAR) coherence. Mapping the areas affected by sediment transport events in arid environments is one of the most common applications of CCD. However, the reliability of these maps remains an unsolved issue. This paper focuses on verifying that InSAR coherence is indeed able to detect all the fluvial sediment transport events that have actually mobilised sediments in arid environments by building a classification model and validating its results. The proposed methodology is tested in three study areas in Salar de Atacama, Chile, using three years of Sentinel data plus a fourth year for validation, and meteorological records of rainfall, the relative humidity of the air and snow cover. The results prove that InSAR coherence can be used to remotely detect sediment transport events related to flash floods in arid environments, that it might have a greater detection ability than meteorological records and that the perpendicular baseline does have a relevant effect on the InSAR coherence that needs to be considered. All these findings will increase the reliability of maps based on InSAR coherence. In addition, the proposed method will allow focusing the mapping tasks only on the relevant dates and, once calibrated, the classification model will enable the automatised remote detection of new events. Full article

Currently, numerous machine learning (ML) techniques are being applied in the field of renewable energy (RE). These techniques may not perform well if they do not have enough training data. Additionally, the main assumption in most of the ML algorithms is that the training and testing data are from the same feature space and have similar distributions. However, in many practical applications, this assumption is false. Recently, transfer learning (TL) has been introduced as a promising machine-learning framework to mitigate these issues by preparing extra-domain data so that knowledge may be transferred across domains. This learning technique improves performance and avoids the resource expensive collection and labeling of domain-centric datasets; furthermore, it saves computing resources that are needed for re-training new ML models from scratch. Lately, TL has drawn the attention of researchers in the field of RE in terms of forecasting and fault diagnosis tasks. Owing to the rapid progress of this technique, a comprehensive survey of the related advances in RE is needed to show the critical issues that have been solved and the challenges that remain unsolved. To the best of our knowledge, few or no comprehensive surveys have reviewed the applications of TL in the RE field, especially those pertaining to forecasting solar and wind power, load forecasting, and predicting failures in power systems. This survey fills this gap in RE classification and forecasting problems, and helps researchers and practitioners better understand the state of the art technology in the field while identifying areas for more focused study. In addition, this survey identifies the main issues and challenges of using TL for REs, and concludes with a discussion of future perspectives. Full article

The goal of this paper is to examine how the development of attention networks has left many important issues unsolved and to propose possible directions for solving them by combining human and animal studies. The paper starts with evidence from citation mapping that indicates attention has played a central role in integrating cognitive and neural studies into Cognitive Neuroscience. The integration of the fields depends in part upon similarities and differences in performance over a wide variety of animals. In the case of exogenous orienting of attention primates, rodents and humans are quite similar, but this is not so with executive control. In humans, attention networks continue to develop at different rates during infancy and childhood and into adulthood. From age four on, the Attention Network Test (ANT) allows measurement of individual differences in the alerting, orienting and executive networks. Overt and covert orienting do overlap in their anatomy, but there is evidence of some degree of functional independence at the cellular level. The attention networks frequently work together with sensory, memory and other networks. Integration of animal and human studies may be advanced by examining common genes involved in individual attention networks or their integration with other brain networks. Attention networks involve widely scattered computation nodes in different brain areas, both cortical and subcortical. Future studies need to attend to the white matter that connects them and the direction of information flow during task performance. Full article

Robotic manipulation challenges, such as grasping and object manipulation, have been tackled successfully with the help of deep reinforcement learning systems. We give an overview of the recent advances in deep reinforcement learning algorithms for robotic manipulation tasks in this review. We begin by outlining the fundamental ideas of reinforcement learning and the parts of a reinforcement learning system. The many deep reinforcement learning algorithms, such as value-based methods, policy-based methods, and actor–critic approaches, that have been suggested for robotic manipulation tasks are then covered. We also examine the numerous issues that have arisen when applying these algorithms to robotics tasks, as well as the various solutions that have been put forth to deal with these issues. Finally, we highlight several unsolved research issues and talk about possible future directions for the subject. Full article

The sensor–weapon–target allocation (S-WTA) is a typical collaborative task allocation problem involved in network-centric warfare (NCW). The existing related studies have a limitation to the nature of cooperation and uncertainty in an air defense battle scenario, and most existing models have the assumption that they are determinate, i.e., the parameters in them are known certainly. For the actual battlefield environment, the asymmetric information in it could lead to the failure of the above assumption, and there are many uncertainties whose frequency can not be evaluated objectively. Based on uncertainty theory, this paper studied the S-WTA problem in an indeterminate battlefield environment. First, we analyze the uncertain factors existing in the actual battlefield environment and their influence on the S-WTA problem, and by considering the threat value of the target, the deviation parameters of the sensor tracking performance and weapon interception performance as uncertain variables, we then establish an uncertain S-WTA (USWTA) model, where the destruction value to targets is regarded as an objective function and four categories of typical constraints are set. Further, an equivalent transformation is presented to convert the unsolvable model into a determinate one by the expected value principle. To solve the proposed model efficiently, a permutation-based representation for the allocation scheme of the USWTA problem is introduced firstly, which can construct a feasible solution efficiently, and on this basis, a constructive heuristic algorithm based on maximum marginal return rule (MMRCH) is designed to construct a feasible solution with high quality. Additionally, a local search (LS) operation is proposed to explore for the better solution locally and further improve the quality of solution obtained by MMRCH. Finally, a set of instances are set to be solved by the designed algorithm, and the simulation experiment demonstrates the superiority of the designed algorithm and the feasibility of the proposed model. Full article

Modern crack detection algorithms based on deep learning have unsolved issues, such as an abundance of parameters in the resulting models and lack of context information. Such issues may lower the efficiency of feature extraction and lead to unexpected task performance. Based on two semantic segmentation models, U-Net and the dual-attention network (DANet), an efficient mobile-attention X-network (MA-Xnet) is proposed for crack detection. For performance evaluation, segmentation experiments were performed on concrete crack images from an internationally recognized dataset, which were collected from various campus buildings of Middle East Technical University. The experimental results demonstrated that, compared with U-Net, the proposed method parameters were reduced by 82.33%, and improved by 11.32% and 12.37% in the key indices of the F1-Score and the mean intersection of union (mIoU), respectively, providing a reference for subsequent related lightweight crack-segmentation research. Full article