Evolutionary multitasking optimization (EMTO) is an emerging and highly promising research topic in the evolutionary computation community. It aims to simultaneously solve multiple optimization tasks under shared computational resources and to achieve superior performance compared with single-task optimization through cross-task knowledge transfer. However, EMTO mainly focuses on unconstrained multitask optimization and multitasking-assisted constrained multiobjective optimization, whereas real-world industrial problems often involve constraints, conflicting objectives, and multiple tasks, limiting their applicability. Moreover, most existing knowledge transfer techniques adopt indiscriminate transfer strategies, neglecting the differentiated knowledge demands of various populations or individuals at different evolution stages, which easily leads to negative knowledge transfer. To address these issues, this paper proposes a constrained multiobjective multitask optimization framework, referred to as CMOMT-MSDKT. Within this framework, a multi-stage and diversified demand-oriented knowledge transfer technique (MSDKT) is developed to identify the demands for feasibility, diversity, and convergence knowledge at different evolution process and to perform differentiated knowledge transfer accordingly. In addition, a set of constrained multiobjective multitask optimization problems (MSCO) for multiseries collaboration optimization in copper electrolysis process are constructed, covering multi-level decision-making requirements. Moreover, a new performance metric, multitask convergence score, is proposed to comprehensively evaluate the convergence performance of algorithms for constrained multiobjective multitasking optimization. Experimental results on the constructed MSCO instances and public SOPM instances demonstrate that CMOMT-MSDKT achieves superior overall performance compared with seven representative algorithms. Specifically, CMOMT-MSDKT obtains the best average Friedman ranking among all compared algorithms, and shows statistically significant advantages on the majority of instances in terms of feasibility, diversity, and convergence performance. Finally, ablation experiments further verify the effectiveness of the proposed multi-stage knowledge transfer technique, while parameter sensitivity experiments analyze the performance variations of CMOMT-MSDKT under different parameter settings.

One of the main challenges when developing medical decision support systems for the emergency room is adequately filtering the most relevant information. High workload, stress, and the necessity for urgent decisions require precise answers to the questions posed. Although LLM-based systems can provide abundant information, physicians need concise and relevant data in this particular clinical setting. In this study, we perform a pilot assessment of the transparency of selected LLM-based systems. The comparative analysis includes ChatGPT o1 model, which was asked to produce responses with varying temperatures and a pilot graph-based RAG specializing in cardiovascular diseases. A survey was conducted among 33 clinicians regarding the amount of information contained in the provided prompts. Physicians favored the most readable, specific, and helpful answers in emergency department conditions. Reliable medical data and the form in which answers are delivered are crucial for physicians working in the emergency room. We conclude that physicians have preferences for LLM responses at a specific temperature. Further research should be expanded to enable tailoring responses not only to the clinical situation but also to the experience of the asking physician.

Knowledge graphs are recognized as a valuable format for representing data and information. Their ability to represent semantics using different types of relations between the concepts and denoting information at different levels of abstraction creates a demand for algorithms taking advantage of such data format. In this paper, we propose a method for determining the similarity between concepts in weighted knowledge graphs. The method uses a hierarchical approach to determine the degree of similarity at different levels of 'distance' from the considered graph concepts. The proposed technique employs the T-norm and OWA operator. Similarities between concepts account for edge weights, while OWA aggregates similarities between nodes at different levels of distance from the compared nodes. The method is explained, and its merits are discussed.

A new approach based on the Fuzzy Adaptive Resonance Theory (ART) network, called the Dynamic Recursive Fuzzy ART Classifier (DyRFAC), is presented for incremental supervised learning on multi-label datasets. Whereas many ART-based algorithms rely on a single, unvarying vigilance parameter, our classifier employs a dynamic vigilance mechanism, enabling finer-grained partitioning of the data space. Through recursive splitting guided by a purity measure, DyRFAC iteratively adjusts category boundaries, allowing the category space to more accurately capture the data distribution. In addition, DyRFAC introduces a category merging procedure to control the growth of categories when data scales, preventing excessive proliferation that could degrade model performance. We conduct experiments comparing DyRFAC with other multi-label classification algorithms, demonstrating its competitive performance on complex label distributions.

Hierarchy analysis of the knowledge graphs aims to discover the latent structure inherent in knowledge base data. Drawing inspiration from topic modeling, which identifies latent themes and content patterns in text corpora, our research seeks to adapt these analytical frameworks to the hierarchical exploration of knowledge graphs. Specifically, we adopt a non-parametric probabilistic model, the nested hierarchical Dirichlet process, to the field of knowledge graphs. This model discovers latent subject-specific distributions along paths within the tree. Consequently, the global tree can be viewed as a collection of local subtrees for each subject, allowing us to represent subtrees for each subject and reveal cross-thematic topics. We assess the efficacy of this model in analyzing the topics and word distributions that form the hierarchical structure of complex knowledge graphs. We quantitatively evaluate our model using four common datasets: Freebase, Wikidata, DBpedia, and WebRED, demonstrating that it outperforms the latest neural hierarchical clustering techniques such as TraCo, SawETM, and HyperMiner. Additionally, we provide a qualitative assessment of the induced subtree for a single subject.

This paper proposes a Quantum Computational Intelligence (QCI) model integrated with Generative Artificial Intelligence (GAI) for Taiwanese/English language co-learning applications within human-machine interactions. The QCI model comprises two main phases: human-machine interaction and data processing for quantum circuit generation and real-world applications. During the human-machine interaction phase, a synergy between Human Intelligence (HI) and Machine Intelligence (MI) enables young students to gain familiarity with CI that converges with QCI. The second phase involves data processing, which encompasses stages of data preprocessing, analysis, and evaluation. The methodology is applied to two distinct applications: 1) Application 1 focuses on constructing a knowledge graph using the Ollama platform and the TAIDE model—a Trustworthy AI Dialogue Engine developed by the Taiwanese government based on the LLaMa 2 model. 2) Application 2 addresses the GAI images to text/voice, and text/voice to GAI images, depending on the type of Taiwanese/English data collected. Subsequently, the QCI model is refined through Particle Swarm Optimization (PSO) and Genetic Algorithm Neural Networks (GANN). Moreover, a Quantum Fuzzy Inference Mechanism (QFIM) is integrated to enhance the QCI model’s capability in creating a quantum circuit. The experimental results suggest that the QCI model significantly enhances human-machine collaboration. Looking forward, we plan to extend the QCI model to reach more young learners.

Due to aging infrastructure, technical issues, increased demand, and environmental developments, the reliability of power systems is of paramount importance. Utility companies aim to provide uninterrupted and efficient power supply to their customers. To achieve this, they focus on implementing techniques and methods to minimize downtime in power networks and reduce maintenance costs. In addition to traditional statistical methods, modern technologies such as machine learning have become increasingly common for enhancing system reliability and customer satisfaction. The primary objective of this study is to review parametric and nonparametric machine learning techniques and their applications in relation to maintenance-related aspects of power distribution system assets, including (1) distribution lines, (2) transformers, and (3) insulators. Compared to other reviews, this study offers a unique perspective on machine learning algorithms and their predictive capabilities in relation to the critical components of power distribution systems.

Measuring the kyphotic angle (KA) and lordotic angle (LA) on lateral radiographs is important to truly diagnose children with adolescent idiopathic scoliosis. However, it is a time-consuming process to measure the KA because the endplate of the upper thoracic vertebra is normally difficult to identify. To save time and improve measurement accuracy, a machine learning algorithm was developed to automatically extract the KA and LA. The accuracy and reliability of the T1-T12 KA, T5-T12 KA, and L1-L5 LA were reported. A convolutional neural network was trained using 100 radiographs with data augmentation to segment the T1-L5 vertebrae. Sixty radiographs were used to test the method. Accuracy and reliability were reported using the percentage of measurements within clinical acceptance (≤9°), standard error of measurement (SEM), and inter-method intraclass correlation coefficient (ICC2,1). The automatic method detected 95 % (57/60), 100 %, and 100 % for T1-T12 KA, T5-T12 KA, and L1-L5 LA, respectively. The clinical acceptance rate, SEM, and ICC2,1 for T1-T12 KA, T5-T12 KA, and L1-L5 LA were (98 %, 0.80°, 0.91), (75 %, 4.08°, 0.60), and (97 %, 1.38°, 0.88), respectively. The automatic method measured quickly with an average of 4 ± 2 s per radiograph and illustrated how measurements were made on the image, allowing verifications by clinicians.

This paper proposes a quantum computational intelligence (QCI) model integrated with generative artificial intelligence (GAI) for Taiwanese/English language co-learning applications within human–machine interactions, focusing on Trustworthy AI Dialogue Engine (TAIDE)-based knowledge graph construction and multimodal data transformation. The QCI model comprises two main phases: human–machine interaction and data processing for quantum circuit generation and real-world applications. During the human–machine interaction phase, a synergy between human intelligence (HI) and machine intelligence (MI) enables young students to gain familiarity with CI that converges with QCI. The second phase involves data processing, which encompasses stages of data preprocessing, analysis, and evaluation. The methodology is applied to two distinct applications: Application 1 focuses on constructing a knowledge graph using the Ollama platform and the TAIDE model developed by the Taiwanese government based on the LLaMa 2 model. Application 2 addresses the GAI images to text/voice and text/voice to GAI images, depending on the type of Taiwanese/English data collected. Subsequently, the QCI model is refined through particle swarm optimization (PSO) and genetic algorithm neural networks (GANN). Moreover, a quantum fuzzy inference mechanism (QFIM) is integrated into the developed QCI AI-FML learning platform to generate quantum circuits for the QCI model, which helps teach young students and facilitate their learning of QCI. The experimental results indicate that the QCI model significantly enhances human–machine collaboration. Looking forward, we plan to extend the QCI model to reach more young learners.

OBJECTIVE:To develop and validate machine learning algorithms to automatically extract the rod length of the magnetically controlled growing rod from ultrasound images (US) in a pilot study. METHODS:Two machine-learning (ML) models, called the "Boundary model" and "Rod model," were developed to identify specific rod segments on ultrasound images. The models were developed utilizing Mask Regional Convolutional Neural Networks (Mask RCNN). Ninety US images were acquired from 23 participants who had early onset scoliosis (EOS) surgeries; among those, 70 were used for model development, including training and validation, and 20 were used for testing by comparing the AI-based vs. manual measurements. RESULTS:The average precision (AP) of the ML models was 88.5% and 60.2%, respectively. The inter-method correlation coefficient (ICC) was 0.98, and the mean absolute difference ± standard deviation (MAD ± SD) between AI and manual measurements was 0.86 ± 1.0 mm. The Bland-Altman analysis showed no bias, and 90% of the data were within the 95% confidence interval. The automated method was reliable, accurate, and fast. Measurements were displayed in 4.6 seconds after the US image was inputted. CONCLUSION:This was the first AI-based method to measure the MCGR rod length on US images automatically.

This paper proposes a transformer-based semantic robot with a computational intelligence (CI) mechanism designed for use in an educational co-learning environment, where teachers, teaching assistants, and students interact with the CI robot and attention ontology to enhance the learning process. The approach is applied in two distinct applications. The first, focusing on student-machine co-learning with writing performance evaluation, involves an attention-based mechanism for curating learning content from students, which is further refined by a preprocessing mechanism with expert-based fuzzy numbers. The second, concentrating on student-machine co-learning with speaking performance evaluation, introduces a Meta AI Universal Speech Translator (UST) Taiwanese/English agent that translates content into English and Taiwanese speeches, as well as into English and Chinese texts. This transformer-based robot for computing semantic similarities employs a trained semantic Sentence-BERT (SBERT) model to analyze student-machine co-learning contents. Given the large size of the co-learning content with the ontology model, we implement a chunk-based approach for processing. This method enables effective comparison of the extensive student-provided learning content with the evaluative content from teachers and teaching assistants. Additionally, a Human Intelligence (HI)-based robot, equipped with a CI assessment mechanism based on fuzzy numbers, evaluates performance and adjusts the evaluation content of teachers and teaching assistants based on HI fuzzy numbers. Experimental results indicate that the proposed CI robot can reduce teachers' burden and objectively evaluate student-machine co-learning performance, thereby narrowing the gap in actual student-machine co-learning performance. Furthermore, it aids in assessing student-machine co-learning performance and understanding, creating a more personalized and effective learning environment.

This paper proposes a Siamese motion-aware Spatio-temporal network ( SiamMAST ) for video action recognition. The SiamMAST is designed based on the fusion of four features via processing video frames: spatial features, temporal features, spatial dynamic features, and temporal dynamic features of a moving target. The SiamMAST comprises AlexNets as the backbone, LSTMs, and the spatial motion-awareness and temporal motion-awareness sub-modules. RGB images are fed into the network, where AlexNets extract spatial features. Further, they are fed into LSTMs to generate temporal features. Additionally, spatial motion-awareness and temporal motion-awareness sub-modules are proposed to capture spatial and temporal dynamic features. Finally, all features are fused and fed into the classification layer. The final recognition result is produced by averaging the test label probabilities across a fixed number of RGB frames and selecting the label of the highest probability. The whole network is trained offline using an end-to-end approach with large-scale image datasets using the standard SGD algorithm with back-propagation. The proposed network is evaluated on two challenging datasets UCF101 (93.53%) and HMDB51 (69.36%). The experiments have demonstrated the effectiveness and efficiency of our proposed SiamMAST .

Families of individuals with neurodevelopmental disabilities or differences (NDDs) often struggle to find reliable health information on the web. NDDs encompass various conditions affecting up to 14% of children in high-income countries, and most individuals present with complex phenotypes and related conditions. It is challenging for their families to develop literacy solely by searching information on the internet. While in-person coaching can enhance care, it is only available to a minority of those with NDDs. Chatbots, or computer programs that simulate conversation, have emerged in the commercial sector as useful tools for answering questions, but their use in health care remains limited. To address this challenge, the researchers developed a chatbot named CAMI (Coaching Assistant for Medical/Health Information) that can provide information about trusted resources covering core knowledge and services relevant to families of individuals with NDDs. The chatbot was developed, in collaboration with individuals with lived experience, to provide information about trusted resources covering core knowledge and services that may be of interest. The developers used the Django framework (Django Software Foundation) for the development and used a knowledge graph to depict the key entities in NDDs and their relationships to allow the chatbot to suggest web resources that may be related to the user queries. To identify NDD domain–specific entities from user input, a combination of standard sources (the Unified Medical Language System) and other entities were used which were identified by health professionals as well as collaborators. Although most entities were identified in the text, some were not captured in the system and therefore went undetected. Nonetheless, the chatbot was able to provide resources addressing most user queries related to NDDs. The researchers found that enriching the vocabulary with synonyms and lay language terms for specific subdomains enhanced entity detection. By using a data set of numerous individuals with NDDs, the researchers developed a knowledge graph that established meaningful connections between entities, allowing the chatbot to present related symptoms, diagnoses, and resources. To the researchers’ knowledge, CAMI is the first chatbot to provide resources related to NDDs. Our work highlighted the importance of engaging end users to supplement standard generic ontologies to named entities for language recognition. It also demonstrates that complex medical and health-related information can be integrated using knowledge graphs and leveraging existing large datasets. This has multiple implications: generalizability to other health domains as well as reducing the need for experts and optimizing their input while keeping health care professionals in the loop. The researchers' work also shows how health and computer science domains need to collaborate to achieve the granularity needed to make chatbots truly useful and impactful.

This paper proposes a Content Attention Ontology (CAO) robot for constructing Taiwanese/English Knowledge Graphs (KGs) by prompting audio or texts to Large Language Models (LLMs), including TAIDE, Zephyr, and Llama 3.1. The collected data includes lecture videos from the IEEE WCCI 2024 in Japan and the 2024 National Language Development Forum in Taiwan, along with students' learning data from the 2024 Summer School on Taiwanese/English Human and Robot Co-Learning at Rende Elementary School (RDES). In addition, the fundamental concepts of Computational Intelligence (CI) and Quantum CI (QCI) learning were incorporated into the study. The generative KGs highlight important concepts, relations, and communities within the collected teaching and learning data. Additionally, we utilized data from subjects wearing braincomputer interface (BCI) devices while speaking Taiwanese/English to generate KGs. We also compared the differences in these KGs and analyzed the similarities between the transcribed texts of lectures and learners. In the future, we plan to expand the CAO robot to more validation fields across Taiwan, aiming to engage young students in speaking Taiwanese while concurrently enhancing their English language skills through interaction with the robot.

The observed growth in Artificial Intelligence sparks an increased expectation of constructing intelligent systems. Equipping such systems with a semantically rich representation of data and information is essential. This paper briefly introduces the use of knowledge graphs and category theory to create a framework for ‘clever’ data processing. It focuses on the key role of category theory mechanisms in data synthesis, particularly in concept validation and construction.

Distribution grids are complex networks containing multiple pieces of equipment. These components are interconnected, and each of them is described by various attributes. A knowledge graph is an interesting data format that represents pieces of information as nodes and relations between the pieces as edges. In this paper, we describe the proposed vocabulary used to build a distribution system knowledge graph. We identify the concepts used in such graphs and a set of relations to represent links between concepts. Both provide a semantically rich representation of a system. Additionally, we offer a few illustrative examples of how a distributed system knowledge graph can be utilized to gain more insight into the operations of the grid. We show a simplified analysis of how outages can influence customers based on their locations and how adding DERs can influence/change it. These demonstrative use cases show that the graph-based representation of a distribution grid allows for integrating information of different types and how such a repository can be efficiently utilized. Based on the experiments with distribution system knowledge graphs presented in this article, we postulate that graph-based representation enables a novel way of storing information about power grids and facilitates interactive methods for their visualization and analysis.

To validate a fast 3D biplanar spinal radiograph reconstruction method with automatic extract curvature parameters using artificial intelligence (AI). Three-hundred eighty paired, posteroanterior and lateral, radiographs from the EOS X-ray system of children with adolescent idiopathic scoliosis were randomly selected from the database. For the AI model development, 304 paired images were used for training; 76 pairs were employed for testing. The validation was evaluated by comparing curvature parameters, including Cobb angles (CA), apical axial vertebral rotation (AVR), kyphotic angle (T1–T12 KA), and lordotic angle (L1–L5 LA), to manual measurements from a rater with 8 years of scoliosis experience. The mean absolute differences ± standard deviation (MAD ± SD), the percentage of measurements within the clinically acceptable errors, the standard error of measurement (SEM), and the inter-method intraclass correlation coefficient ICC[2,1] were calculated. The average reconstruction speed of the 76 test images was recorded. Among the 76 test images, 134 and 128 CA were exported automatically and measured manually, respectively. The MAD ± SD for CA, AVR at apex, KA, and LA were 3.3° ± 3.5°, 1.5° ± 1.5°, 3.3° ± 2.6° and 3.5° ± 2.5°, respectively, and 98

The magnetically controlled growing rod technique is an effective surgical treatment for children who have early-onset scoliosis. The length of the instrumented growing rods is adjusted regularly to compensate for the normal growth of these patients. Manual measurement of rod length on posteroanterior spine radiographs is subjective and time-consuming. A machine learning (ML) system using a deep learning approach was developed to automatically measure the adjusted rod length. Three ML models—rod model, 58 mm model, and head-piece model—were developed to extract the rod length from radiographs. Three-hundred and eighty-seven radiographs were used for model development, and 60 radiographs with 118 rods were separated for final testing. The average precision (AP), the mean absolute difference (MAD) ± standard deviation (SD), and the inter-method correlation coefficient (ICC[2,1]) between the manual and artificial intelligence (AI) adjustment measurements were used to evaluate the developed method. The AP of the 3 models were 67.6