Search Results (112)

Clinical notes often contain unstructured text filled with abbreviations, non-standard terminology, and inconsistent phrasing, which pose significant challenges for automated medical information extraction. Named Entity Recognition (NER) plays a crucial role in structuring this data by identifying and categorizing key clinical entities such as symptoms, medications, and diagnoses. However, traditional and even transformer-based NER models often struggle with ambiguity and fail to produce clinically interpretable outputs. In this study, we present a hybrid two-stage framework that enhances medical NER by integrating a fine-tuned BERT model for initial entity extraction with a Dictionary-Infused Retrieval-Augmented Generation (DiRAG) module for terminology normalization. Our approach addresses two critical limitations in current clinical NER systems: lack of contextual clarity and inconsistent standardization of medical terms. The DiRAG module combines semantic retrieval from a UMLS-based vector database with lexical matching and prompt-based generation using a large language model, ensuring precise and explainable normalization of ambiguous entities. The fine-tuned BERT model achieved an F1 score of 0.708 on the MACCROBAT dataset, outperforming several domain-specific baselines, including BioBERT and ClinicalBERT. The integration of the DiRAG module further improved the interpretability and clinical relevance of the extracted entities. Through qualitative case studies, we demonstrate that our framework not only enhances clarity but also mitigates common issues such as abbreviation ambiguity and terminology inconsistency. Full article

With the advancement of artificial intelligence, Agricultural Expert Systems (AESs) show great potential in enhancing agricultural management efficiency and resource utilization. Accurate extraction of semantic features from agricultural short texts is fundamental to enabling key functions such as intelligent question answering, semantic retrieval, and decision support. However, existing single-structure deep neural networks struggle to capture the hierarchical linguistic patterns and contextual dependencies inherent in domain-specific texts. To address this limitation, we propose a hybrid deep learning model—Bidirectional Encoder Recurrent Neural Network (BERNN)—which combines a domain-specific pre-trained Transformer encoder (AgQsBERT) with a Bidirectional Long Short-Term Memory (BiLSTM) network. AgQsBERT generates contextualized word embeddings by leveraging domain-specific pretraining, effectively capturing the semantics of agricultural terminology. These embeddings are then passed to the BiLSTM, which models sequential dependencies in both directions, enhancing the model’s understanding of contextual flow and word disambiguation. Importantly, the bidirectional nature of the BiLSTM introduces a form of architectural symmetry, allowing the model to process input in both forward and backward directions. This symmetric design enables balanced context modeling, which improves the understanding of fragmented and ambiguous phrases frequently encountered in agricultural texts. The synergy between semantic abstraction from AgQsBERT and symmetric contextual modeling from BiLSTM significantly enhances the expressiveness and generalizability of the model. Evaluated on a self-constructed agricultural question dataset with 110,647 annotated samples, BERNN achieved a classification accuracy of 97.19%, surpassing the baseline by 3.2%. Cross-domain validation on the Tsinghua News dataset further demonstrates its robust generalization capability. This architecture provides a powerful foundation for intelligent agricultural question-answering systems, semantic retrieval, and decision support within smart agriculture applications. Full article

The increasing adoption of artificial intelligence across domains presents new opportunities to enhance access to justice. In this paper, we introduce a human-centric AI tool that utilises advances in Automatic Speech Recognition (ASR) and Large Language Models (LLMs) to facilitate semantic linking between written UK Supreme Court (SC) judgements and their corresponding hearing videos. The motivation stems from the critical role UK SC hearings play in shaping landmark legal decisions, which often span several hours and remain difficult to navigate manually. Our approach involves two key components: (1) a customised ASR system fine-tuned on 139 h of manually edited SC hearing transcripts and legal documents and (2) a semantic linking module powered by GPT-based text embeddings adapted to the legal domain. The ASR system addresses domain-specific transcription challenges by incorporating a custom language model and legal phrase extraction techniques. The semantic linking module uses fine-tuned embeddings to match judgement paragraphs with relevant spans in the hearing transcripts. Quantitative evaluation shows that our customised ASR system improves transcription accuracy by 9% compared to generic ASR baselines. Furthermore, our adapted GPT embeddings achieve an F1 score of 0.85 in classifying relevant links between judgement text and hearing transcript segments. These results demonstrate the effectiveness of our system in streamlining access to critical legal information and supporting legal professionals in interpreting complex judicial decisions. Full article

Background/Objectives: This review discusses the resistance mechanisms in the tumor microenvironment (TME) of malignant melanoma that disrupt the efficacy of immune checkpoint inhibitors (ICIs). In this review, we focus on the roles of immune cells, including tumor-infiltrating lymphocytes (TILs), macrophages, dendritic cells, and other signaling pathways. We explore the interplay between innate and adaptive immunity in the TME and tumor intrinsic resistance mechanisms, such as β-catenin, which has future implications for the usage of ICIs in patients with therapy-resistant tumors. Methods: A total of 1052 studies were extracted from the PubMed database searching for keywords and phrases that included [melanoma AND immune checkpoint inhibitor resistance]. After a title/abstract and full-text review, 101 studies were identified that fit the inclusion/exclusion criteria. Results: Cancer-associated fibroblasts (CAFs), M2 macrophages, and myeloid-derived suppressor cells (MDSCs) are significant in remodeling the TME to promote melanoma growth. Melanoma resistance to ICIs is complex and involves TME alterations, tumor intrinsic factors, and immune evasion. Key components of resistance include reduced CD8+ T cell infiltration, decreased host immune response, and immunosuppressive cytokines. Conclusions: Predictive biomarkers and specific models are the future of individualized melanoma management and show great promise in their approach to targeted therapy production. Tumor profiling can be utilized to help predict the efficacy of ICIs, and specific biomarkers predicting therapy responses are instrumental in moving towards personalized and more efficacious medicine. As more melanoma resistance emerges, alternative and combinatorial therapy based on knowledge of existing resistance mechanisms will be needed. Full article

This study investigates syntactic parafoveal processing in Chinese reading using a boundary paradigm with two-character verb–object phrases. Participants (N = 120 undergraduates) viewed sentences with manipulated previews (identity, syntactically consistent, and inconsistent previews). Results showed a selective syntactic preview effect: syntactical violations reduced target word skipping rates, but fixation durations remained unaffected. This dissociation contrasts with robust syntactic preview benefits observed in alphabetic languages, highlighting how Chinese’s lack of morphological markers constrains parafoveal processing. The findings challenge parallel processing models while supporting language-specific modulation of universal cognitive mechanisms. Our results advance understanding of hierarchical information extraction in reading, with implications for developing cross-linguistic reading models. Full article

This paper proposes a KeyBERT-based method for constructing a knowledge graph of the electronic information curriculum system, aiming to enhance the structured representation and relational analysis of educational content. Electronic Information Engineering curricula encompass diverse and rapidly evolving topics; however, existing knowledge graphs often overlook multi-word concepts and more nuanced semantic relationships. To address this gap, this paper presents a KeyBERT-enhanced method for constructing a knowledge graph of the electronic information curriculum system. Utilizing teaching plans, syllabi, and approximately 500,000 words of course materials from 17 courses, we first extracted 500 knowledge points via the Term Frequency–Inverse Document Frequency (TF-IDF) algorithm to build a baseline course–knowledge matrix and visualize the preliminary graph using Graph Convolutional Networks (GCN) and Neo4j. We then applied KeyBERT to extract about 1000 knowledge points—approximately 65% of extracted terms were multi-word phrases—and augment the graph with co-occurrence and semantic-similarity edges. Comparative experiments demonstrate a ~20% increase in non-zero matrix coverage and a ~40% boost in edge count (from 5100 to 7100), significantly enhancing graph connectivity. Moreover, we performed sensitivity analysis on extraction thresholds (co-occurrence ≥ 5, similarity ≥ 0.7), revealing that (5, 0.7) maximizes the F1-score at 0.83. Hyperparameter ablation over n-gram ranges [(1,1),(1,2),(1,3)] and top_n [5, 10, 15] identifies (1,3) + top_n = 10 as optimal (Precision = 0.86, Recall = 0.81, F1 = 0.83). Finally, GCN downstream tests show that, despite higher sparsity (KeyBERT 64% vs. TF-IDF 40%), KeyBERT features achieve Accuracy = 0.78 and F1 = 0.75, outperforming TF-IDF’s 0.66/0.69. This approach offers a novel, rigorously evaluated solution for optimizing the electronic information curriculum system and can be extended through terminology standardization or larger data integration. Full article

Large-language-model (LLM) APIs demonstrate impressive reasoning capabilities, but their size, cost, and closed weights limit the deployment of knowledge-aware AI within biomedical research groups. At the other extreme, standard attention-based neural language models (SANLMs)—including encoder–decoder architectures such as Transformers, Gated Recurrent Units (GRUs), and Long Short-Term Memory (LSTM) networks—are computationally inexpensive. However, their capacity for semantic reasoning in noisy, open-vocabulary knowledge bases (KBs) remains unquantified. Therefore, we investigate whether compact SANLMs can (i) reason over hybrid OpenIE-derived KBs that integrate commonsense, general-purpose, and non-communicable-disease (NCD) literature; (ii) operate effectively on commodity GPUs; and (iii) exhibit semantic coherence as assessed through manual linguistic inspection. To this end, we constructed four training KBs by integrating ConceptNet (600k triples), a 39k-triple general-purpose OpenIE set, and an 18.6k-triple OpenNCDKB extracted from 1200 PubMed abstracts. Encoder–decoder GRU, LSTM, and Transformer models (1–2 blocks) were trained to predict the object phrase given the subject + predicate. Beyond token-level cross-entropy, we introduced the Meaning-based Selectional-Preference Test (MSPT): for each withheld triple, we masked the object, generated a candidate, and measured its surplus cosine similarity over a random baseline using word embeddings, with significance assessed via a one-sided t-test. Hyperparameter sensitivity (311 GRU/168 LSTM runs) was analyzed, and qualitative frame–role diagnostics completed the evaluation. Our results showed that all SANLMs learned effectively from the point of view of the cross entropy loss. In addition, our MSPT provided meaningful semantic insights: for the GRUs (256-dim, 2048-unit, 1-layer): mean similarity $\left({\mu }_{sts}\right)$ of 0.641 to the ground truth vs. 0.542 to the random baseline (gap 12.1%; $p<{10}^{-180}$). For the 1-block Transformer: ${\mu }_{sts}=0.551$ vs. $0.511$ (gap 4%; $p<{10}^{-25}$). While Transformers minimized loss and accuracy variance, GRUs captured finer selectional preferences. Both architectures trained within <24 GB GPU VRAM and produced linguistically acceptable, albeit over-generalized, biomedical assertions. Due to their observed performance, LSTM results were designated as baseline models for comparison. Therefore, properly tuned SANLMs can achieve statistically robust semantic reasoning over noisy, domain-specific KBs without reliance on massive LLMs. Their interpretability, minimal hardware footprint, and open weights promote equitable AI research, opening new avenues for automated NCD knowledge synthesis, surveillance, and decision support. Full article

Deciphering the structural hierarchy of musical compositions is indispensable for a range of music analysis applications, encompassing feature extraction, data compression, interpretation, and visualization. In this paper, we introduce a quantitative model grounded in fractal theory to evaluate the significance of individual notes within a musical piece. To analyze the quantized note importance, we adopt a rhythm-based approach and propose a series of detection operators informed by fundamental rhythmic combinations. Employing the Mamba model, we carry out recursive detection operations that offer a hierarchic understanding of musical structures. By organizing the composition into a tree data structure, we achieve an ordered layer traversal that highlights the music piece’s multi-dimensional features. Musical compositions often exhibit intrinsic symmetry in their temporal organization, manifested through repetition, variation, and self-similar patterns across scales. Among these symmetry properties, fractality stands out as a prominent characteristic, reflecting recursive structures both rhythmically and melodically. Our model effectively captures this property, providing insights into the fractal-like regularities within music. It also proves effective in musical phrase boundary detection tasks, enhancing the clarity and visualization of musical information. The findings illustrate the model’s potential to advance the quantitative analysis of music hierarchy, promoting novel methodologies in musicological research. Full article

Large-language models (LLMs) show promise for automating evidence synthesis, yet head-to-head evaluations remain scarce. We benchmarked five state-of-the-art LLMs—openai/o1-mini, x-ai/grok-2-1212, meta-llama/Llama-3.3-70B-Instruct, google/Gemini-Flash-1.5-8B, and deepseek/DeepSeek-R1-70B-Distill—on extracting protocol details from transcranial direct-current stimulation (tDCS) trials enrolling older adults. A multi-LLM ensemble pipeline ingested ClinicalTrials.gov records, applied a structured JSON schema, and generated comparable outputs from unstructured text. The pipeline retrieved 83 aging-related tDCS trials—roughly double the yield of a conventional keyword search. Across models, agreement was almost perfect for the binary field brain stimulation used (Fleiss κ ≈ 0.92) and substantial for the categorical primary target (κ ≈ 0.71). Numeric parameters such as stimulation intensity and session duration showed excellent consistency when explicitly reported (ICC 0.95–0.96); secondary targets and free-text duration phrases remained challenging (κ ≈ 0.61; ICC ≈ 0.35). An ensemble consensus (majority vote or averaging) resolved most disagreements and delivered near-perfect reliability on core stimulation attributes (κ = 0.94). These results demonstrate that multi-LLM ensembles can markedly expand trial coverage and reach expert-level accuracy on well-defined fields while still requiring human oversight for nuanced or sparsely reported details. The benchmark and open-source workflow set a solid baseline for future advances in prompt engineering, model specialization, and ensemble strategies aimed at fully automated evidence synthesis in neurostimulation research involving aging populations. Overall, the five-model multi-LLM ensemble doubled the number of eligible aging-related tDCS trials retrieved versus keyword searching and achieved near-perfect agreement on core stimulation parameters (κ ≈ 0.94), demonstrating expert-level extraction accuracy. Full article

Background: Lactoferrin is a major functional protein involved in maintaining human health, which possesses antioxidant, anti-inflammatory, antibacterial, and antiviral properties. Therefore, it can be used to support the treatment of viral and bacterial diseases, as well as in cancer prevention. Lactoferrin-manufacturing processes may compromise its protein structure and function, so it is necessary to establish reliable analytical methods for production efficiency and quality control purposes. This paper reviews the lactoferrin production processes, summarising the methods using various matrices (milk, milk powder, infant formula, whey, bovine lactoferrin lyophilised powder, yoghurt, colostrum, and human milk), the most popular purification methods, and sample preparation. Material and methods: The Medline and Embase databases were searched using the following phrases: ”lactoferrin” and “purification” or “isolation” or “extraction” or “separation”. The search was limited to recent studies from the last five years published in English up until 12 March 2025. Of the 573 articles identified, 17 were reviewed. Results: Lactoferrin purification and determination methods depend on the matrix used. The latest research focuses on improving parameters of lactoferrin determination, shortening time, improving efficiency or limiting costs, and even reducing toxicity by changing the reagents. The method of separating lactoferrin using magnetic beads or nanoparticles has been developed, as well as the determination parameters using high-performance liquid chromatography (HPLC). Conclusions: The current lactoferrin production techniques are characterised by increased efficiency and quality, but they require standardisation of the purification process depending on the matrix. The latest Lf determination methods are highly precise, and most of them produce high-quality Lf. This allows to introduce on the market a higher quality product, which can significantly improve standard approaches. Full article

The purpose of this bibliometric analysis is to understand what skill sets are needed for the entrepreneurial mindset, how the entrepreneurial mindset is practically operationalized, and where opportunities can be identified using the entrepreneurial mindset. The entrepreneurial mindset is crucial in entrepreneurship and innovation, leading to value creation, business development, and competitive advantage. The methodological approach involves a bibliometric analysis utilizing seven databases and a total of 478 articles that were selected based on the phrase “entrepreneurial mindset”. Data were extracted on 6 July 2024, and the bibliometric analysis consisted of four separate steps in the methodological approach. The findings identified six different clusters in which the entrepreneurial mindset adopted a process-oriented perspective, a concept that is underexplored in the current literature. The novelty in this study involves a cluster in the findings, labeled “the missing dynamics”, which warrants attention. Overall, the missing dynamics cluster in this bibliometric analysis offers originality and further research suggestions. By continuing to explore the process-oriented views of the entrepreneurial mindset, new value opportunities can be created, while the missing dynamics can be better understood. Full article

Companies’ marketing decision-making effectiveness depends on the quality of actions and time. In the current digital era, any marketing decision making must be timely in response to customers’ feedback, and implementing artificial intelligence (AI) technology is one significant option. This paper focuses on designing an Algorithm for Marketing Strategy Decision Making (AMSDM) that employs AI services to process online feedback from customers regarding products and services from companies’ websites or other e-commerce and social media platforms. For this research, 1200 texts containing customer feedback were analyzed by Azure Text Analytics service, which identifies the types of customers’ online feedback, domains, subdomains, and keywords it refers to and understands the emotional tone and attitudes conveyed in customer responses through sentiment analysis techniques. The model performance was underlined by computing the Accuracy, Precision, Recall, and F₁-Score metrics for both short and long phrases feedback. Furthermore, Azure Text Analytics was integrated into a C# script to extract the frequency of occurrence of domains, subdomains, and keywords. After that, the process of AMSDM and its advantages were detailed. The AMSDM eliminates the necessity for manual intervention and conserves both time and resources. Moreover, the real-time nature of the analysis allows companies to respond promptly to changing market dynamics and customer preferences. Full article

This study examines the role of linguistic regional variations in synthetic dataset generation and their impact on emotion detection performance. Emotion detection is essential for natural language processing (NLP) applications such as social media analysis, customer service, and mental health monitoring. To explore this, synthetic datasets were generated using a state-of-the-art language model, incorporating English variations from the United States, United Kingdom, and India, alongside a general baseline dataset. Two levels of prompt specificity were employed to assess the influence of regional linguistic nuances. Statistical analyses—including frequency distribution, term frequency-inverse document frequency (TF-IDF), type–token ratio (TTR), hapax legomena, pointwise mutual information (PMI) scores, and key-phrase extraction—revealed significant linguistic diversity and regional distinctions in the generated datasets. To evaluate their effectiveness, classification experiments were conducted with two models using bidirectional encoder representations from transformers (BERT) and its de-noising sequence to sequence variation (BART), beginning with zero-shot classification on the contextualized affect representations for emotion recognition (CARER) dataset, followed by fine-tuning with both baseline and region-specific datasets. Results demonstrated that region-specific datasets, particularly those generated with detailed prompts, significantly improved classification accuracy compared to the baseline. These findings underscore the importance of incorporating global linguistic variations in synthetic dataset generation, offering insights into how regional adaptations can enhance emotion detection models for diverse NLP applications. Full article

Vision–language pre-training (VLP) faces challenges in aligning hierarchical textual semantics (words/phrases/sentences) with multi-scale visual features (objects/relations/global context). We propose a hierarchical VLP model (HieVLP) that addresses such challenges through semantic decomposition and progressive alignment. Textually, a semantic parser deconstructs captions into word-, phrase-, and sentence-level components, which are encoded via hierarchical BERT layers. Visually, a Swin Transformer extracts object- (local), relation- (mid-scale), and global-level features through shifted window hierarchies. During pre-training, a freezing strategy sequentially activates text layers (sentence→phrase→word), aligning each with the corresponding visual scales via contrastive and language modeling losses. The experimental evaluations demonstrate that HieVLP outperforms hierarchical baselines across various tasks, with the performance improvements ranging from approximately 3.2% to 11.2%. In the image captioning task, HieVLP exhibits an average CIDEr improvement of around 7.2% and a 2.1% improvement in the SPICE metric. For image–text retrieval, it achieves recall increases of 4.7–6.8%. In reasoning tasks, HieVLP boosts accuracy by 2.96–5.8%. These results validate that explicit multi-level alignment enables contextually coherent caption generation and precise cross-modal reasoning. Full article

Background/Objectives: Oral cancer, the sixth most common cancer worldwide, is linked to smoke, alcohol, and HPV. This scoping analysis summarized early-onset oral cancer diagnosis applications to address a gap. Methods: A scoping review identified, selected, and synthesized AI-based oral cancer diagnosis, screening, and prognosis literature. The review verified study quality and relevance using frameworks and inclusion criteria. A full search included keywords, MeSH phrases, and Pubmed. Oral cancer AI applications were tested through data extraction and synthesis. Results: AI outperforms traditional oral cancer screening, analysis, and prediction approaches. Medical pictures can be used to diagnose oral cancer with convolutional neural networks. Smartphone and AI-enabled telemedicine make screening affordable and accessible in resource-constrained areas. AI methods predict oral cancer risk using patient data. AI can also arrange treatment using histopathology images and address data heterogeneity, restricted longitudinal research, clinical practice inclusion, and ethical and legal difficulties. Future potential includes uniform standards, long-term investigations, ethical and regulatory frameworks, and healthcare professional training. Conclusions: AI may transform oral cancer diagnosis and treatment. It can develop early detection, risk modelling, imaging phenotypic change, and prognosis. AI approaches should be standardized, tested longitudinally, and ethical and practical issues related to real-world deployment should be addressed. Full article