MQRLD: A multimodal data retrieval platform with query-aware feature representation and learned index based on data lake

Information Processing & Management, Год журнала: 2025, Номер 62(4), С. 104101 - 104101

Опубликована: Фев. 17, 2025

Язык: Английский

---

Can LLMs Revolutionize Text Mining in Chemistry? A Comparative Study with Domain-Specific Tools

Computer Standards & Interfaces, Год журнала: 2025, Номер 94, С. 103997 - 103997

Опубликована: Март 2, 2025

Язык: Английский

---

An intent recognition pipeline for conversational AI

International Journal of Information Technology, Год журнала: 2023, Номер 16(2), С. 731 - 743

Опубликована: Дек. 29, 2023

Abstract Natural Language Processing (NLP) is one of the Artificial Intelligence applications that entitled to allow computers process and understand human language. These models are utilized analyze large volumes text also support aspects like summarization, language translation, context modeling, sentiment analysis. language, a subset Understanding (NLU), turns natural into structured data. NLU accomplishes intent classification entity extraction. The paper focuses on pipeline maximize coverage conversational AI (chatbot) by extracting maximum meaningful intents from data corpus. A can best answer queries with respect dataset if it trained number be gathered which what we focus getting in this paper. higher gather dataset, more cover training AI. modularized three broad categories - Gathering corpus, finding misspellings synonyms intents, finally deciding order picked up for any classifier ML model. Several heuristic machine-learning approaches have been considered optimum results. For synonyms, they extracted through vector neural network-based algorithms. Then system concludes suggestive priority list should fed In end, an example corpus picked, their suggested functioning pipeline. This attempts pick descending most optimal way possible.

Язык: Английский

---

Potential Applications and Ethical Considerations for Artificial Intelligence in Traumatic Brain Injury Management

Biomedicines, Год журнала: 2024, Номер 12(11), С. 2459 - 2459

Опубликована: Окт. 26, 2024

Artificial intelligence (AI) systems have emerged as promising tools for rapidly identifying patterns in large amounts of healthcare data to help guide clinical decision making, well assist with medical education and the planning research studies. Accumulating evidence suggests AI techniques may be particularly useful aiding diagnosis management traumatic brain injury (TBI)—a considerably heterogeneous neurologic condition that can challenging detect treat. However, important methodological ethical concerns use medicine necessitate close monitoring regulation these advancements continue. The purpose this narrative review is provide an overview common describe recent studies on possible applications context TBI. Finally, describes challenges medicine, guidelines from White House, Department Defense (DOD), National Academies Sciences, Engineering, Medicine (NASEM), other organizations appropriate uses research.

Язык: Английский

---

Enhancing Medical Decision Making: A Semantic Technology-Based Framework for Efficient Diagnosis Inference

Mathematics, Год журнала: 2024, Номер 12(4), С. 502 - 502

Опубликована: Фев. 6, 2024

In the dynamic landscape of healthcare, decision support systems (DSS) confront continuous challenges, especially in era big data. Background: This study extends a Q&A-based medical DSS framework that utilizes semantic technologies for disease inference based on patient’s symptoms. The inputs “evidential symptoms” (symptoms experienced by patient) and outputs ranked list hypotheses, comprising an ordered pair characteristic symptom. Our focus is advancing introducing ontology integration to semantically enrich its knowledgebase refine outcomes, offering three key advantages: Propagation, Hierarchy, Range Expansion Additionally, we assessed performance fully implemented Python. During evaluation, inspected framework’s ability infer from subset reported symptoms evaluated effectiveness ranking it prominently among hypothesized diseases. Methods: We conducted expansion using dedicated algorithms. For evaluation process, defined various metrics applied them across our knowledge base, encompassing 410 patient records 41 different Results: presented outcomes toy problem, highlighting advantages. Furthermore, process yielded promising results: With third as evidence, successfully identified 94% cases, achieving top-ranking accuracy 73%. Conclusions: These results underscore robust capabilities framework, enrichment enhances efficiency experts, enabling provide more precise informed diagnostics.

Язык: Английский

---

Electronic Health Record–Oriented Knowledge Graph System for Collaborative Clinical Decision Support Using Multicenter Fragmented Medical Data: Design and Application Study

Journal of Medical Internet Research, Год журнала: 2024, Номер 26, С. e54263 - e54263

Опубликована: Июль 5, 2024

Background The medical knowledge graph provides explainable decision support, helping clinicians with prompt diagnosis and treatment suggestions. However, in real-world clinical practice, patients visit different hospitals seeking various services, resulting fragmented patient data across hospitals. With security issues, fragmentation limits the application of graphs because single-hospital cannot provide complete evidence for generating precise support comprehensive explanations. It is important to study new methods systems integrate into multicenter, information-sensitive environments, using records while maintaining privacy security. Objective This aims propose an electronic health record (EHR)–oriented system collaborative reasoning multicenter data, all preserving privacy. Methods introduced EHR framework a novel process utilizing information. was deployed each hospital used unified semantic structure Observational Medical Outcomes Partnership (OMOP) vocabulary standardize local set. transforms formats performs generate intermediate findings. generated findings hypernym concepts isolate original data. hash-encrypted identities were synchronized through blockchain network. collaborated final without gathering Results underwent evaluation involving utilization alert non-nephrology about overlooked chronic kidney disease (CKD). covered 1185 nonnephrology departments from 3 visited at least two Of these, 124 identified as meeting CKD criteria whereas individual alone could not facilitate identification these patients. assessment by indicated that 78/91 (86%) positive. Conclusions proposed able effectively utilize application. showed benefits support.

Язык: Английский

---

A Strategy-Driven Semantic Framework for Precision Decision Support in Targeted Medical Fields

Applied Sciences, Год журнала: 2025, Номер 15(3), С. 1561 - 1561

Опубликована: Фев. 4, 2025

Healthcare 4.0 addresses modernization and digital transformation challenges, such as home-based care precision treatments, by leveraging advanced technologies to enhance accessibility efficiency. Semantic technologies, particularly knowledge graphs (KGs), have proven instrumental in representing interconnected medical data improving clinical decision-support systems. We previously introduced a semantic framework assist experts during patient interactions. Operating iteratively, the prompts with relevant questions based on input, progressing toward accurate diagnoses time-constrained settings. It comprises two components: (a) KG symptoms, diseases, their relationships, (b) algorithms that generate prioritize hypotheses—a ranked list of symptom–disease pairs. An earlier extension enriched symptom ontology, incorporating hierarchical structures inheritance relationships improve accuracy question-generation capabilities. This paper further extends introducing strategies tailored specific domains. Strategies integrate domain-specific algorithms, refining decision making while maintaining iterative nature expert–patient demonstrate this approach using an emergency medicine case study, focusing life-threatening conditions. The is attributes contexts supported dedicated algorithms. Boolean rules attached graph edges evaluate TRUE or FALSE at runtime patient-specific data. These enhancements optimize embedding goal-oriented inference processes, providing scalable adaptable solution for diverse contexts.

Язык: Английский

---

Multiple disease diagnoses using heterogeneous EHR curated knowledge graph and machine learning models

Applied Intelligence, Год журнала: 2025, Номер 55(7)

Опубликована: Апрель 3, 2025

Язык: Английский

---

A disambiguation method for potential ambiguities in Chinese based on knowledge graphs and large language model

Alexandria Engineering Journal, Год журнала: 2025, Номер 126, С. 293 - 302

Опубликована: Май 1, 2025

Язык: Английский

---

XLR-KGDD: leveraging LLM and RAG for knowledge graph-based explainable disease diagnosis using multimodal clinical information

Knowledge and Information Systems, Год журнала: 2025, Номер unknown

Опубликована: Май 13, 2025

Язык: Английский

---