Domain-specific Knowledge Research Articles

RRMSE-enhanced weighted voting regressor for improved ensemble regression.

Ensemble regression methods are widely used to improve prediction accuracy by combining multiple regression models, especially when dealing with continuous numerical targets. However, most ensemble voting regressors use equal weights for each base model's predictions, which can limit their effectiveness, particularly when there is no specific domain knowledge to guide the weighting. This uniform weighting approach doesn't consider that some models may perform better than others on different datasets, leaving room for improvement in optimizing ensemble performance. To overcome this limitation, we propose the RRMSE (Relative Root Mean Square Error) Voting Regressor, a new ensemble regression technique that assigns weights to each base model based on their relative error rates. By using an RRMSE-based weighting function, our method gives more importance to models that demonstrate higher accuracy, thereby enhancing the overall prediction quality. We tested the RRMSE Voting Regressor on six popular regression datasets and compared its performance with several state-of-the-art ensemble regression algorithms. The results show that the RRMSE Voting Regressor consistently achieves lower prediction errors than existing methods across all tested datasets. This improvement highlights the effectiveness of using relative error metrics for weighting in ensemble models. Our approach not only fills a gap in current ensemble regression techniques but also provides a reliable and adaptable method for boosting prediction performance in various machine learning tasks. By leveraging the strengths of individual models through smart weighting, the RRMSE Voting Regressor offers a significant advancement in the field of ensemble learning.

Development of an Automated Construction Contract Review Framework Using Large Language Model and Domain Knowledge

Construction contract review demands specialized expertise, requiring comprehensive understanding of both technical and legal aspects. While AI advancements offer potential solutions, two problems exist: LLMs lack sufficient domain-specific knowledge to analyze construction contracts; existing RAG approaches do not effectively utilize domain expertise. This study aims to develop an automated contract review system that integrates domain expertise with AI capabilities while ensuring reliable analysis. By transforming expert knowledge into a structured knowledge base aligned with the SCF classification, the proposed structured knowledge-integrated RAG pipeline is expected to enable context-aware contract analysis. This enhanced performance is achieved through three key components: (1) integrating structured domain knowledge with LLMs, (2) implementing filtering combined with hybrid dense–sparse retrieval mechanisms, and (3) employing reference-based answer generation. Validation using Oman’s standard contract conditions demonstrated the system’s effectiveness in assisting construction professionals with contract analysis. Performance evaluation showed significant improvements, achieving a 52.6% improvement in Context Recall and a 48.3% improvement in Faithfulness compared to basic RAG approaches. This study contributes to enhancing the reliability of construction contract review by applying a structured knowledge-integrated RAG pipeline that enables the accurate retrieval of expert knowledge, thereby addressing the industry’s need for precise contract analysis.

Applied machine learning in intelligent systems: knowledge graph-enhanced ophthalmic contrastive learning with “clinical profile” prompts

IntroductionThe integration of artificial intelligence (AI) into ophthalmic diagnostics has the potential to significantly enhance diagnostic accuracy and interpretability, thereby supporting clinical decision-making. However, a major challenge in AI-driven medical applications is the lack of transparency, which limits clinicians’ trust in automated recommendations. This study investigates the application of machine learning techniques by integrating knowledge graphs with contrastive learning and utilizing “clinical profile” prompts to refine the performance of the ophthalmology-specific large language model, MeEYE, which is built on the CHATGLM3-6B architecture. This approach aims to improve the model’s ability to capture clinically relevant features while enhancing both the accuracy and explainability of diagnostic predictions.MethodsThis study employs a novel methodological framework that incorporates domain-specific knowledge through knowledge graphs and enhances feature representation using contrastive learning. The MeEYE model is fine-tuned with structured clinical knowledge, enabling it to better distinguish subtle yet significant ophthalmic features. Additionally, “clinical profile” prompts are incorporated to further improve contextual understanding and diagnostic precision. The proposed method is evaluated through comprehensive performance benchmarking, including quantitative assessments and clinical case studies, to ensure its efficacy in real-world ophthalmic diagnosis.ResultsThe experimental findings demonstrate that integrating knowledge graphs and contrastive learning into the MeEYE model significantly improves both diagnostic accuracy and model interpretability. Comparative analyses against baseline models reveal that the proposed approach enhances the identification of ophthalmic conditions with higher precision and clarity. Furthermore, the model’s ability to generate transparent and clinically relevant AI recommendations is substantiated through rigorous evaluation, highlighting its potential for real-world clinical implementation.DiscussionThe results underscore the importance of explainable AI in medical diagnostics, particularly in ophthalmology, where model transparency is critical for clinical acceptance and utility. By incorporating domain-specific knowledge with advanced machine learning techniques, the proposed approach not only enhances model performance but also ensures that AI-generated insights are interpretable and reliable for clinical decision-making. These findings suggest that integrating structured medical knowledge with machine learning frameworks can address key challenges in AI-driven diagnostics, ultimately contributing to improved patient outcomes. Future research should explore the adaptability of this approach across various medical domains to further advance AI-assisted diagnostic systems.

Named Entity Recognition in Online Medical Consultation Using Deep Learning

Named entity recognition in online medical consultation aims to address the challenge of identifying various types of medical entities within complex and unstructured social text in the context of online medical consultations. This can provide important data support for constructing more powerful online medical consultation knowledge graphs and improving virtual intelligent health assistants. A dataset of 26 medical entity types for named entity recognition for online medical consultations is first constructed. Then, a novel approach for deep named entity recognition in the medical field based on the fusion context mechanism is proposed. This approach captures enhanced local and global contextual semantic representations of online medical consultation text while simultaneously modeling high- and low-order feature interactions between local and global contexts, thereby effectively improving the sequence labeling performance. The experimental results show that the proposed approach can effectively identify 26 medical entity types with an average F1 score of 85.47%, outperforming the state-of-the-art (SOTA) method. The practical significance of this study lies in improving the efficiency and performance of domain-specific knowledge extraction in online medical consultation, supporting the development of virtual intelligent health assistants based on large language models and enabling real-time intelligent medical decision-making, thereby helping patients and their caregivers access common medical information more promptly.

Medical foundation large language models for comprehensive text analysis and beyond

Recent advancements in large language models (LLMs) show significant potential in medical applications but are hindered by limited specialized medical knowledge. We present Me-LLaMA, a family of open-source medical LLMs integrating extensive domain-specific knowledge with robust instruction-following capabilities. Me-LLaMA is developed through continual pretraining and instruction tuning of LLaMA2 models using diverse biomedical and clinical data sources (e.g., biomedical literature and clinical notes). We evaluated Me-LLaMA on six text analysis tasks using 12 benchmarks (e.g., PubMedQA and MIMIC-CXR) and assessed its clinical utility in complex case diagnosis through automatic and human evaluations. Me-LLaMA outperforms existing open medical LLMs in zero-shot and supervised settings and surpasses ChatGPT and GPT-4 after task-specific instruction tuning for most text analysis tasks. Its performance is also comparable to ChatGPT and GPT-4 for diagnosing complex clinical cases. Our findings highlight the importance of combining domain-specific continual pretraining with instruction tuning to enhance performance in medical LLMs.

Hallucination Mitigation for Retrieval-Augmented Large Language Models: A Review

Retrieval-augmented generation (RAG) leverages the strengths of information retrieval and generative models to enhance the handling of real-time and domain-specific knowledge. Despite its advantages, limitations within RAG components may cause hallucinations, or more precisely termed confabulations in generated outputs, driving extensive research to address these limitations and mitigate hallucinations. This review focuses on hallucination in retrieval-augmented large language models (LLMs). We first examine the causes of hallucinations from different sub-tasks in the retrieval and generation phases. Then, we provide a comprehensive overview of corresponding hallucination mitigation techniques, offering a targeted and complete framework for addressing hallucinations in retrieval-augmented LLMs. We also investigate methods to reduce the impact of hallucination through detection and correction. Finally, we discuss promising future research directions for mitigating hallucinations in retrieval-augmented LLMs.

Evaluating Serious Slow Game Jams as a Mechanism for Co-Designing Serious Games to Improve Understanding of Cybersecurity

We present a first evaluation of a Serious Slow Game Jam (SSGJ) methodology as a mechanism for co-designing serious games in the application domain of cybersecurity to assess how the SSGJ contributed to improving the understanding of cybersecurity. To this end, we engaged 13 participants with no experience in cybersecurity in a multidisciplinary SSGJ involving domain-specific, pedagogical and game design knowledge and encouraged engagement in between scheduled days of the SSGJ. Findings show improved confidence of participants in their knowledge of cybersecurity (from 12.5% to 62.5%) after undertaking the SSGJ, with free-text answers specifically indicating an improved understanding in terms of vulnerabilities, attacks and defences for three-quarters of the participants. Also, confidence in knowledge of game design improved (12.5% to 75%), and the SSGJ successfully engaged participants in between scheduled days. Finally, a serious game is presented that was co-designed with participants during our SSGJ and produced as an output of the SSGJ methodology.

Engaging Undergraduate Students in Research: From Cross-Discipline Programs to Published Articles

This paper discusses the effective integration of undergraduate students into research, ensuring mutually beneficial outcomes for students, faculty, higher education institutes, and society at large. Student candidates are identified through a screening process leveraging the existing institutional resources and programs. Selection criteria, including minimum grade point average (GPA), strong writing skills, and achievements in challenging quantitative courses, are employed. Once a suitable candidate is identified, research collaboration commences. We work with the student to identify a topic aligned with our projects that piques their interest and curiosity, while ensuring feasibility within the timeframe. Subsequently, we invest a month or two in guiding the student through relevant literature, building a comprehensive understanding of the chosen subject matter. As the research question evolves, we assist the student in mastering the necessary tools and methods, providing relevant programming code or directing the student to specific knowledge domains. The structured approach mirrors the collaboration with graduate students. We deliberately avoid limiting our selection to specific disciplines, promoting diversity in student engagement. This exposure broadens their understanding of applied economics research and enriches the educational experience for both students and faculty mentors. The result is the cultivation of cross-disciplinary programs that contribute to the growth and development of undergraduate research.

OpenFOAMGPT: A retrieval-augmented large language model (LLM) agent for OpenFOAM-based computational fluid dynamics

This work presents a large language model (LLM)-based agent OpenFOAMGPT tailored for OpenFOAM-centric computational fluid dynamics (CFD) simulations, leveraging two foundation models from OpenAI: the GPT-4o (GPT means Generative Pre-trained Transformer) and a chain-of-thought–enabled o1 preview model. Both agents demonstrate success across multiple tasks. While the price of token with o1 model is six times as that of GPT-4o, it consistently exhibits superior performance in handling complex tasks, from zero-shot/few-shot case setup to boundary condition modifications, zero-shot turbulence model adjustments, and zero-shot code translation. Through an iterative correction loop, the agent efficiently addressed single-phase and multiphase flow, heat transfer, Reynolds-averaged Navier–Stokes modeling, large eddy simulation, and other engineering scenarios, often converging in a limited number of iterations at low token costs. To embed domain-specific knowledge, we employed a retrieval-augmented generation pipeline, demonstrating how preexisting simulation setups can further specialize the agent for subdomains such as energy and aerospace. Despite the great performance of the agent, human oversight remains crucial for ensuring accuracy and adapting to shifting contexts. Fluctuations in model performance over time suggest the need for monitoring in mission-critical applications. Although our demonstrations focus on OpenFOAM, the adaptable nature of this framework opens the door to developing LLM-driven agents into a wide range of solvers and codes. By streamlining CFD simulations, this approach has the potential to accelerate both fundamental research and industrial engineering advancements.

Nursing practices on sexual health for people with neurological disorders: A narrative review.

To identify and describe nursing practices on the sexual health of people with neurological disorders. Narrative review. Data were extracted from 1 January 2002, to 20 May 2021. Inclusion criteria were nursing practices, sexual health and people with neurological disorders. The main outcome measures were: context of nursing practice implementation (assumptions, knowledge, strategies and skills), facilitators of and barriers to addressing and treating the sexual health of people with neurological disorders, and benefits of nursing practices in sexual health. PRISMA reporting guidelines were used. PubMed, Embase, ScienceDirect and CINAHL. In total, 926 articles were identified and nine were included. The involvement of nurses was recommended in most studies. Assumptions about the impact of neurology on sexuality and nurse's role in sexual healthcare, biopsychosocial knowledge, and skills (ethical, interpersonal, and technical) were highlighted. We found that the modes of knowledge proposed by Carper were mobilized in an unequal way. Sexual difficulties were the key focus and eroticization concerns were not addressed in any of the articles. Several studies advocate nursing intervention; however, few accurately present, detail and evaluate sexual health nursing practices of patients with neurological pathologies. Literature describes practices structured around disorders rather than the potentials, fails to address the brake of eroticism and provides little information on the results of interventions. Developing teaching programs on sexual health in nursing programs may be necessary if nurses are to support a diverse range of patients in an inclusive and positive manner. These programs should highlight the domain-specific knowledge that is mobilized. Sexual health is a fundamental human right. Alterations in the nervous system have shown to affect sexual health, however, it is not often discussed among patients with neurological disorders, who are rarely provided with sexual health counselling. Our findings may impact healthcare professionals engaged in care with these patients. PRISMA. No patient or public contribution.

AI in the cosmos

Artificial intelligence (AI) is revolutionizing research by enabling the efficient analysis of large datasets and the discovery of hidden patterns. In astrophysics, AI has become essential, transforming the classification of celestial sources, data modeling, and the interpretation of observations. In this review, I highlight examples of AI applications in astrophysics, including source classification, spectral energy distribution modeling, and discuss the advancements achievable through generative AI. However, the use of AI introduces challenges, including biases, errors, and the “black box” nature of AI models, which must be resolved before their application. These issues can be addressed through the concept of Human-Guided AI (HG-AI), which integrates human expertise and domain-specific knowledge into AI applications. This approach aims to ensure that AI is applied in a robust, interpretable, and ethical manner, leading to deeper insights and fostering scientific excellence.

ChatGPT performance in assessing musculoskeletal MRI scan appropriateness based on ACR appropriateness criteria

Large Language Models (LLMs) hold potential as clinical decision support tools, particularly when integrated with domain-specific knowledge. In radiology, there is limited research on LLMs for assessing imaging appropriateness. This study evaluates a contextualized GPT-4-based LLM’s performance in assessing the appropriateness of musculoskeletal MRI scan requests with standard models and different versions of optimization. The LLMs’ performances was also compared against human clinicians with varying experience (two radiology residents, two subspecialist attendings, an orthopaedic surgeon). Using a retrieval-augmented generation framework, the LLM was provided with a domain-specific knowledge base from 33 American College of Radiology Appropriateness Criteria guidelines. A test dataset of 70 fictional case scenarios was created, including cases with insufficient clinical information. Quantitative analysis using the McNemar mid-P test revealed that the optimized LLM achieved 92.86% accuracy, significantly outperforming the baseline model (61.43%, P < .001) and the standard GPT-4 model (51.29%, P < .001). The optimized model also excelled in identifying cases with insufficient clinical information. In comparison to human clinicians, the optimized LLM performed better than all but one radiologist. This study demonstrates that with contextualization and optimization, GPT-4-based LLMs can improve performance in assessing imaging appropriateness and show promise as clinical decision support tools in radiology.

LLM을 활용한 검색 증강 생성(RAG) 모델 구현: 래피드마이너 기반으로

Generative AI technology, driven by Large Language Models (LLMs), is being increasingly utilized to overcome existing limitations. Retrieval-Augmented Generation (RAG) has emerged as an effective approach to reduce hallucination in LLMs by leveraging up-to-date and domain-specific knowledge beyond training data. However, most studies propose programming-based implementations. This research introduces a GUI-based RAG framework using RapidMiner, to construct RAG systems without programming proficiency. The methodology includes storing and retrieving embeddings with the Qdrant vector database and generating question-and-answer pairs via the OpenAI API. Practical demonstrations confirm the system’s effectiveness in real-world scenarios, offering a simpler and more efficient method for developing generative AI services with LLMs.

Knowledge Management in a World of Generative AI: Impact and Implications

The management of organizational and worker knowledge is a critical part of any organization. Knowledge management recognizes the explicit knowledge that workers contribute to an organization, as well as the organizational knowledge and skills acquired over time. The potential impact of generative AI (GenAI) raises many issues and opportunities related to effective knowledge management as organizations explore the impact and implications of its adoption. This research explores issues related to an increasing world of GenAI and proposes a Framework for Knowledge Management in the GenAI Era . Knowledge management workers who are novices might be more inclined to use generative AI for performing their tasks. Expert knowledge workers might focus on quality evaluation and quality assessment. Adoption of GenAI may help with the progression from novice to expert, while freeing up time and resources for other work. Challenges include the potential loss of critical organizational knowledge, definition of job-related requirements, expectations, and tasks, as well as over-reliance, and loss of tacit or unarticulated knowledge. These challenges must be balanced against available access to large amounts of general and domain-specific knowledge from GenAI to capture the value realized from its adoption.

Sparse knowledge sharing (SKS) for privacy-preserving domain incremental seizure detection.

Objective. Epilepsy is a neurological disorder that affects millions of patients worldwide. Electroencephalogram-based seizure detection plays a crucial role in its timely diagnosis and effective monitoring. However, due to distribution shifts in patient data, existing seizure detection approaches are often patient-specific, which requires customized models for different patients. This paper considers privacy-preserving domain incremental learning (PP-DIL), where the model learns sequentially from each domain (patient) while only accessing the current domain data and previously trained models. This scenario has three main challenges: (1) catastrophic forgetting of previous domains, (2) privacy protection of previous domains, and (3) distribution shifts among domains.Approach. We propose a sparse knowledge sharing (SKS) approach. First, Euclidean alignment is employed to align data from different domains. Then, we propose an adaptive pruning approach for SKS to allocate subnet for each domain adaptively, allowing specific parameters to learn domain-specific knowledge while shared parameters to preserve knowledge from previous domains. Additionally, supervised contrastive learning is employed to enhance the model's ability to distinguish relevant features.Main Results. Experiments on two public seizure datasets demonstrated that SKS achieved superior performance in PP-DIL.Significance. SKS is a rehearsal-free privacy-preserving approach that effectively learns new domains while minimizing the impact on previously learned domains, achieving a better balance between plasticity and stability.

An Unsupervised Moving Object Detection Network for UAV Videos

UAV moving object detection focuses on identifying moving objects in images captured by UAVs, with broad applications in regional surveillance and event reconnaissance. Compared to general moving object detection scenarios, UAV videos exhibit unique characteristics, including foreground sparsity and varying target scales. The direct application of conventional background modeling or motion segmentation methods from general settings may yield suboptimal performance in UAV contexts. This paper introduces an unsupervised UAV moving object detection network. Domain-specific knowledge, including spatiotemporal consistency and foreground sparsity, is integrated into the loss function to mitigate false positives caused by motion parallax and platform movement. Multi-scale features are fully utilized to address the variability in target sizes. Furthermore, we have collected a UAV moving object detection dataset from various typical scenarios, providing a benchmark for this task. Extensive experiments conducted on both our dataset and existing benchmarks demonstrate the superiority of the proposed algorithm.

Research on machine learning-based anomaly detection techniques in biomechanical big data environments

Anomaly detection is critical in identifying abnormal patterns in big data environments, where traditional techniques often struggle with scalability and efficiency. This paper explores machine learning-based anomaly detection techniques, focusing on their effectiveness in large-scale biomechanical data contexts. The study investigates three prominent methods: K-means clustering, autoencoders, and One-Class Support Vector Machine (SVM), each known for distinct strengths in handling biomechanical data. Through comprehensive simulations and experiments, precision, recall, F1-score, Area Under Curve (AUC), and time efficiency metrics are analyzed. The results highlight the trade-offs between accuracy and computational efficiency, offering insights into model performance in various biomechanical big data scenarios. The discussion emphasizes the suitability of autoencoders for detecting anomalies in complex biomechanical signals (e.g., gait analysis or joint kinematics) and the application of One-Class SVM in high-dimensional biomechanical datasets (e.g., muscle activation patterns or force plate data). The study concludes with recommendations for future research directions, including the integration of domain-specific biomechanical knowledge into machine learning models and the development of hybrid approaches for improved anomaly detection in biomechanics.

Fine-Tuning Large Language Models for Ontology Engineering: A Comparative Analysis of GPT-4 and Mistral

Ontology engineering (OE) plays a critical role in modeling and managing structured knowledge across various domains. This study examines the performance of fine-tuned large language models (LLMs), specifically GPT-4 and Mistral 7B, in efficiently automating OE tasks. Foundational OE textbooks are used as the basis for dataset creation and for feeding the LLMs. The methodology involved segmenting texts into manageable chapters, generating question–answer pairs, and translating visual elements into description logic to curate fine-tuned datasets in JSONL format. This research aims to enhance the models’ abilities to generate domain-specific ontologies, with hypotheses asserting that fine-tuned LLMs would outperform base models, and that domain-specific datasets would significantly improve their performance. Comparative experiments revealed that GPT-4 demonstrated superior accuracy and adherence to ontology syntax, albeit with higher computational costs. Conversely, Mistral 7B excelled in speed and cost efficiency but struggled with domain-specific tasks, often generating outputs that lacked syntactical precision and relevance. The presented results highlight the necessity of integrating domain-specific datasets to improve contextual understanding and practical utility in specialized applications, such as Search and Rescue (SAR) missions in wildfire incidents. Both models, despite their limitations, exhibited potential in understanding OE principles. However, their performance underscored the importance of aligning training data with domain-specific knowledge to emulate human expertise effectively. This study, based on and extending our previous work on the topic, concludes that fine-tuned LLMs with targeted datasets enhance their utility in OE, offering insights into improving future models for domain-specific applications. The findings advocate further exploration of hybrid solutions to balance accuracy and efficiency.

Building Strategic Partnerships for Generative AI Adoption in the Enterprise

Enterprises are increasingly exploring the transformative potential of generative AI across various functions such as marketing, R&amp;D, and customer service. However, successfully integrating these advanced models requires strategic partnerships that leverage both internal capabilities and external expertise. By aligning with specialized technology providers, enterprises can efficiently access domain-specific knowledge, robust infrastructure, and continuous model improvements. On the other hand, forging collaborative relationships with consultancies and research institutions facilitates the development of tailored AI strategies, ensuring that adoption roadmaps align with unique organizational goals and regulatory frameworks. In addition to technology-centric alliances, cross-sector partnerships can further accelerate generative AI adoption. Collaborative efforts between domain experts, industry consortia, and government agencies foster the exchange of best practices while addressing concerns related to ethical deployment, data privacy, and transparency. These partnerships not only mitigate potential risks but also promote adherence to emerging standards in responsible AI governance. By proactively establishing partnerships, organizations can optimize resource allocation and hasten the realization of generative AI’s full potential. Joint ventures enable ongoing support, knowledge-sharing, and the co-creation of innovative solutions that drive tangible value. Moreover, the synergy derived from these strategic alliances helps organizations maintain a competitive advantage by rapidly adapting to evolving AI trends. Ultimately, building and nurturing diverse partnerships serve as a cornerstone for successful generative AI adoption, fostering sustainable growth and ensuring that enterprises harness AI-driven insights responsibly, ethically, and profitably. Ultimately, such collaborations also unlock opportunities for workforce upskilling, driving cultural transformation that underpins long-term success in an increasingly AI-driven business landscape.

Glycan mixture analysis by kernel component composition for matrix factorization.

A major challenge in structural glycomics is the presence of isomeric glycan structures, which may not be fully resolved by separation techniques such as liquid chromatography (LC) and ion mobility spectrometry (IMS). Tandem mass spectrometry (MS/MS) can be employed following on-line separation to distinguish unresolved features, as the temporal profiles of various fragment ions reflect different combinations of those from their respective precursor ions. However, traditional principal component analysis can produce negative signals that are unrealistic for real data, and classic non-negative matrix factorization (NMF) methods may result in factors that include contributions from multiple components. This paper introduces a new variation of NMF, termed kernel component composition (KCC), which enables users to impose domain-specific prior knowledge about the components as parametric kernels. These kernel parameters are then learned directly from the data. We developed a theoretically guaranteed algorithm based on proximal gradient descent to solve the optimization problem posed by KCC and derived detailed parameter update rules when using Gaussian kernels. The effectiveness of the KCC algorithm is demonstrated through simulation tests and its application to deconvoluting chemical datasets, including LC- and IM-MS/MS analysis of isomeric glycan mixtures.