Neural Methods for Programming: A Comprehensive Survey and Future Directions

Whole-life carbon emissions (WLCE) studies are critical in assessing the environmental impact of buildings and promoting sustainable design practices. However, existing methods for estimating WLCE are time-consuming and data-intensive, limiting their usefulness in the early building design stages. In response to this, this research introduces a novel approach by harnessing various machine learning algorithms to predict WLCE and WLCE intensity (normalised by floor area) for buildings. To evaluate the suitability of machine learning algorithms, we conducted an experiment involving ten algorithms to build the prediction models. These models were trained using data from 150 typical residential properties in Cornwall, UK, along with 28 features obtained from a comprehensive survey, including floor area, heating type, and occupant characteristics. The ten algorithms include Multiple Linear Regression, and non-linear algorithms such as Decision Tree, Random Forest. Performance evaluation metrics, such as coefficient of determination (R2), mean absolute error (MAE), means squared error (MSE), root-mean-square error (RMSE), and elapsed time, were employed. Our research contributes to the field by showcasing the effectiveness of machine learning models in predicting building WLCE. We reveal that all the tested machine learning algorithms have the capability to predict WLCE and WLCE intensity, non-linear models outperform linear ones, and the Random Forest (RF) model demonstrates superior performance in terms of accuracy, stability, and efficiency. This research encourages the integration of life cycle studies into the early design stage, even within tight building design schedules, offering practical guidance to architects and designers. Furthermore, these results also benefit a wide range of stakeholders, not only the architects but also the engineers, policymakers, and life cycle assessment (LCA) researchers, contributing to the advancement of data-driven sustainability approaches within the building sector.

Wireless sensor network (WSN) plays a significant role in various commercial, industrial and agriculture sector in wide ranges of applications. Routing is necessary to maintain the reliable communication between different nodes, cluster heads, and base station; however, performance of routing mechanism is challenging due to network lifetime, dynamic nodes, higher packet drop, scalability issue, limited adaptability and environmental conditions. This paper presents various routing techniques used in WSN that covers the brief overview of flat routing, hierarchical routing, location-based routing and bio-inspired routing techniques such as Ant Colony Optimization (ACO) and Artificial Bee Colony Optimization (ABC) algorithms. It focuses on the routing mechanism, routing scenario, mobility, scalability, energy consumption, data aggregation, performance evaluation metrics, challenges and constraints of the routing in WSN. This comprehensive survey provides the future direction for the improvement in routing mechanisms in WSN.KeywordsBio-inspired routingClusteringFlat routingHierarchical routingRoutingWireless sensor network

Diving deeper into underwater image enhancement: A survey

Performance evaluation metrics for ring-oscillator-based temperature sensors on FPGAs: A quality factor

Automatic image segmentation and quantification are critical steps in medical image analysis. The main challenges in medical image segmentation are due to the imbalance in data distribution and spatial variations of ROI. The ideal segmentation should extract all kinds of segments irrespective of size, shape and position. Commonly used metrics such as accuracy, IOU, Dice similarity coefficient consider all the detected pixels in a similar way. However, the detection of smaller segments is critical in medical analysis since it helps in early treatment of the disease and are also easier to miss. Hence, segmentation evaluation must accord larger weighting to pixels in smaller segments compared to the bigger ones. We propose a novel evaluation metric for segmentation performance, emphasizing smaller segments, by assigning a higher weightage to those pixels. Weighted false positives are also considered in deriving the new metric named, “SSEGEP” (Smatt SEGment Emphasized Performance evaluation metric), (range: 0 (Bad) to 1 (Good)). The proposed approach has been applied to two different publicly available real medical data sets of CT modality consisting of scans of the liver and pancreas of 131 and 107 subjects respectively and the results have been compared with existing evaluation metrics. Statistical significance testing is performed to quantity the relevance of the proposed approach. In comparison to Dice similarity coefficient, SSEGEP resulted in a promising p-value of the order 10-18 for hepatic tumor. The proposed metric is found to perform better for the images having multiple segments for a single label and where the regions of interest are not localized.

With increasing popularity of wireless local area network (WLAN) and emerging real-time applications, seamless mobility has become one of the primary concerns. Hence, the choice of a proper handover algorithm is of utmost importance. Various performance evaluation metrics for handover algorithms have been proposed in this regard. We argue that in a multiple access point (AP) scenario, the load on each of the APs, the requested data rate of the mobile terminal, and the soft handover have significant impact on these metrics. In this paper, we present an analytical framework to measure various handover performance evaluation metrics under both load condition and soft handover for a specific data rate request. Our approach is based on finding the circular region centered around an AP within which the requested data rate can be satisfied. We have considered WLAN usage efficiency, handover failure probability, and unnecessary handover probability as evaluation metrics. We have analyzed the impact of load and soft handover on these metrics. The impact of velocity on these metrics has also been analyzed. We also have derived lower bounds on both handover failure probability and unnecessary handover probability by formulating the problem into three circle intersection problem. The circular coverage assumption used in the framework is also validated through simulation.

Machine Learning (ML) has been a critical computational paradigm that has shaped contemporary applications in such domains as finance, healthcare, and cybersecurity, such that its performance evaluation cannot be less critical. However, its selection and interpretation of metrics has remained inconsistent, often leading to misleading conclusions. This study presents a systematic analysis of the most commonly used performance evaluation metrics in ML, integrating conceptual taxonomy, mathematical definitions, and empirical assessment under controlled perturbations. There are three dimensions to ML performance evaluation metrics categorization: robustness, discrimination, and calibration. Experiment conducted on classification and regression, and using synthetic datasets and benchmarks, evaluate threshold variation, class imbalance and label noise. Results obtained showed that no single metric captures model performance comprehensively and widely used metrics may yield conflicting or misleading assessments under certain conditions. Also, context-aware selection and multi-dimensional reporting were necessary for reliable evaluation. By empirically linking metric behaviour to data characteristics, this study provides guidance for context-aware metric selection and reporting that is not only standardized but also evidence-based.

Graph edit distance (GED) is an error tolerant graph matching paradigm whose methods are often evaluated in a classification context and less deeply assessed in terms of the accuracy of the found solution. To evaluate the accuracy of GED methods, low level information is required not only at the classification level but also at the matching level. Most of the publicly available repositories with associated ground truths are dedicated to evaluating graph classification or exact graph matching methods and so the matching correspondences as well as the distance between each pair of graphs are not directly evaluated. This paper consists of two parts. First, we provide a graph database repository annotated with low level information like graph edit distances and their matching correspondences. Second, we propose a set of performance evaluation metrics to assess the performance of GED methods.

In Intelligent Video Systems, most of the recent advanced performance evaluation metrics perform a stage of mapping data between the system results and ground truth. This paper aims to review these metrics using a proposed framework. It will focus on metrics for events detection, objects detection and objects tracking systems.

Early detection of sexually transmitted infections from skin lesions with deep learning: a systematic review and meta-analysis.

This research provides an in-depth evaluation of big data analysis methods using the EDAS (Estimation Based on Distance from Average Solutions) framework. As data volumes continue to rapidly expand, choosing the most appropriate analytics approach has become increasingly important for organizations looking to extract valuable and actionable insights. This research analyzes five widely used clustering algorithms on four key evaluation metrics - Expectation Maximization (EM), Farthest-First (FF), Filtered Clustering (FC), Hierarchical Clustering (HC), and Make Density-Based Clustering (MD). By using the EDAS method, normalized decision ranks and preference scores were generated to evaluate the performance. Among the methods, despite the varying individual metric performance, EM ranked the highest overall (preference score: 0.56843). FC followed with balanced and consistent performance (0.54032), and FF came in third place (0.51761), outperforming RI and Jacquard scores. Interestingly, HC recorded the highest scores in ARI (98.23) and almost perfect Jacquard values (96.34), but ranked the lowest in overall preference (0.46742), emphasizing the need for a complete set of evaluation criteria. The findings suggest that method selection should be aligned with the specific priorities of a use case: EM and FC are best for balanced performance, while HC and MD are best suited for situations requiring high clustering accuracy. This research provides a comparative view of clustering methods in the EDAS context, which assists practitioners and researchers in making data-driven decisions and encourages further exploration into domain-specific or hybrid clustering strategies. Keywords: Big Data Analytics, EDAS Methodology, Clustering Algorithms, Performance Evaluation Metrics, Multi-criteria Decision Analysis

Given a graph G and a query node q , the goal of community search (CS) is to find a structurally cohesive subgraph from G that contains q. Significant progress has been made in community search using deep learning in recent years. To the best of our knowledge, no existing work has provided a comprehensive review of learning-based community search methods. Additionally, we find that: (1) Existing methods offer diverse definitions or descriptions of communities, which require systematic summarization. (2) The methods rely on distinct metrics for limited community assessment. (3) Overhead evaluations of the methods vary and exhibit certain biases. Therefore, a comprehensive survey and experimental study are essential to achieve four key objectives: designing a unified pipeline, clarifying community definitions, enriching community evaluation, and establishing overhead assessment. In this paper, we first propose a unified pipeline for these methods, highlighting techniques. We categorize community definitions and analyze the relationships between identified communities. Beyond that, we proposed several community metrics to evaluate the communities comprehensively. Moreover, we introduce a more detailed overhead evaluation approach that considers resource consumption during both the training and search phases. Finally, we employ the proposed community evaluation metrics and overhead assessment framework to evaluate and analyze the methods, examine correlations among metrics, and explore the effects of several commonly used techniques.

Fire hazards pose a significant risk to human lives, infrastructure, and the environment, necessitating the development of efficient fire detection and localization systems. Traditional methods, including smoke and heat sensors, suffer from high false alarm rates and delayed response times. In recent years, deep learning has emerged as a transformative approach, leveraging computer vision to enhance accuracy in fire detection and localization. This paper provides a comprehensive survey of deep learning techniques employed for fire detection, including Convolutional Neural Networks (CNNs), object detection models such as YOLO and Faster R-CNN, and hybrid approaches integrating multimodal data. The study further explores publicly available fire datasets, preprocessing techniques, and performance evaluation metrics. Additionally, key challenges such as real-time processing constraints, environmental variability, and model generalization are discussed. The paper concludes with an outlook on future advancements, including lightweight AI models, multi-sensor fusion, and synthetic dataset generation for robust fire detection systems.

Single-view 3D object reconstruction (SVOR) aims to recover the 3D shape of an object from a single 2D image. Despite advances in deep learning (DL), challenges such as incomplete image information, scarce 3D data annotation, and highly variable object shapes still limit the performance of SVOR. Meanwhile, with the rapid development of novel view synthesis (NVS) techniques, the SVOR field has received significant advancements. However, existing reviews have not comprehensively covered the rapid developments in NVS-based approaches. This article aims to fill this gap by highlighting the latest progress in SVOR, particularly advancements related to NVS-based methods. Additionally, we observed discrepancies between existing quality evaluation metrics in SVOR and human visual perception. This is because some critical object parts are essential to consider during the evaluation. For example, when reconstructing airplanes, critical parts like the empennage and wings are often overlooked in evaluation metrics due to their smaller size compared to the fuselage. Consequently, poor reconstruction of these parts may not significantly affect overall evaluation scores. To address this issue, we propose a more comprehensive evaluation method that reflects human visual perception accurately. To achieve this, we introduce a weighted evaluation method that considers part saliency and proposes a novel technique for automatically perceiving reconstruction discrepancies. This study effectively enhances the accuracy and consistency of evaluations through these approaches, offering new insights and methodologies, filling a void in the existing literature, and providing valuable contributions to both research and practical applications in SVOR.

Image synthesis with adversarial networks: A comprehensive survey and case studies