<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Latest News | CUHK PAI Research Group</title><link>https://pai-cuhk.github.io/post/</link><atom:link href="https://pai-cuhk.github.io/post/index.xml" rel="self" type="application/rss+xml"/><description>Latest News</description><generator>Hugo Blox Builder (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Thu, 16 Apr 2026 00:00:00 +0000</lastBuildDate><image><url>https://pai-cuhk.github.io/media/icon_hu7988985890593674011.png</url><title>Latest News</title><link>https://pai-cuhk.github.io/post/</link></image><item><title>🏆 Congratulations! Two Projects Won First Prize for Excellent Paper at the 15th Chest Pain Center Conference</title><link>https://pai-cuhk.github.io/post/260416/</link><pubDate>Thu, 16 Apr 2026 00:00:00 +0000</pubDate><guid>https://pai-cuhk.github.io/post/260416/</guid><description>&lt;p>🎉 &lt;strong>Big News!&lt;/strong> We are thrilled to announce that two of our research projects have been awarded the &lt;strong>First Prize for Excellent Paper&lt;/strong> at the &lt;strong>15th Chest Pain Center Conference&lt;/strong>!&lt;/p>
&lt;p>This prestigious conference is co-hosted by the &lt;strong>Hong Kong Hospital Authority&lt;/strong> and the &lt;strong>Oriental Huaxia Cardiovascular Health Research Institute, Suzhou Industrial Park&lt;/strong>. 🏅&lt;/p>
&lt;hr>
&lt;h2 id="-award-winning-projects">🌟 Award-Winning Projects&lt;/h2>
&lt;p>This incredible honor recognizes our team&amp;rsquo;s continuous effort and innovation in leveraging Artificial Intelligence to improve cardiovascular health and patient care. The two awarded projects are:&lt;/p>
&lt;h3 id="1-ai-copilot-of-personalized-risk-education-for-post-revascularization-care-in-patients-with-coronary-artery-disease-cad">1️⃣ AI Copilot of Personalized Risk Education for Post-Revascularization Care in Patients with Coronary Artery Disease (CAD)&lt;/h3>
&lt;p>This project introduces an AI copilot designed to provide personalized risk education and care guidance for CAD patients who have undergone revascularization. It aims to improve long-term patient outcomes through tailored health management and advanced AI reasoning.&lt;/p>
&lt;h3 id="2-clinically-aligned-hierarchical-machine-learning-for-patient-level-classification-of-valvular-regurgitation-severity-and-subtype-from-phonocardiograms-pcg">2️⃣ Clinically Aligned Hierarchical Machine Learning for Patient-Level Classification of Valvular Regurgitation Severity and Subtype from Phonocardiograms (PCG)&lt;/h3>
&lt;p>Focusing on non-invasive diagnostics, this research utilizes a novel clinically aligned hierarchical machine learning framework to accurately classify the severity and subtype of valvular regurgitation using Phonocardiogram (PCG) signals at the patient level.&lt;/p>
&lt;div align="center" style="margin: 2em 0;">
&lt;img src="award2.png" alt="Award Certificate" style="max-width: 100%; border-radius: 12px; box-shadow: 0 8px 24px rgba(0,0,0,0.1);">
&lt;/div>
&lt;hr>
&lt;div style="margin-top: 2.5rem; padding-top: 0.8rem; border-top: 1px solid #eaecef; font-size: 0.85rem; color: #5a626e; text-align: center;">
&lt;i class="fas fa-quote-left" style="margin-right: 6px;">&lt;/i> Advancing AI for Cardiovascular Disease Management.&lt;br>
&lt;/div></description></item><item><title>🏆 Congratulations to Huizi on Winning Outstanding Award at GBA Forum 2025!</title><link>https://pai-cuhk.github.io/post/251101/</link><pubDate>Sat, 01 Nov 2025 00:00:00 +0000</pubDate><guid>https://pai-cuhk.github.io/post/251101/</guid><description>&lt;p>🎉 Our core team member &lt;strong>Yu Huizi&lt;/strong> has been honored with the &lt;strong>Outstanding Award in the Science and Engineering Academic Oral Presentation&lt;/strong> at the &lt;strong>Greater Bay Area Graduate Research Forum 2025&lt;/strong>!&lt;/p>
&lt;p>This prestigious recognition not only celebrates Huizi&amp;rsquo;s exceptional research communication skills but also reflects the growing impact of the AIPatient project. 🏅&lt;/p>
&lt;div align="center" style="margin: 2em 0;">
&lt;div style="background: linear-gradient(145deg, #fff9e8 0%, #fff5e0 100%); border-radius: 32px; padding: 24px 20px; max-width: 700px; margin: 0 auto; box-shadow: 0 8px 24px rgba(0,0,0,0.08); border: 1px solid #ffe0a3;">
&lt;div style="font-size: 3.2rem; margin-bottom: 8px;">🏅&lt;/div>
&lt;div style="font-family: 'Times New Roman', serif; font-weight: 600; font-size: 1.8rem; color: #b45f1b; letter-spacing: 2px;">Certificate of Honor&lt;/div>
&lt;hr style="width: 80px; border: 1px solid #e6c8a0; margin: 16px auto;">
&lt;div style="font-size: 1.6rem; font-weight: 600; color: #2c3e2f;">Yu Huizi&lt;/div>
&lt;div style="font-size: 1.1rem; color: #c96f0e; margin: 10px 0 4px;">Science and Engineering Section · Academic Oral Presentation&lt;/div>
&lt;div style="font-size: 1.2rem; font-weight: 500; color: #1f5e3a;">Outstanding Award&lt;/div>
&lt;hr style="width: 60px; border: 1px solid #e6c8a0; margin: 16px auto;">
&lt;div style="font-size: 0.9rem; font-weight: 500; color: #2c3e2f;">Greater Bay Area Graduate Research Forum 2025&lt;/div>
&lt;div style="margin-top: 14px; font-size: 0.85rem;">November 1, 2025&lt;/div>
&lt;div style="margin-top: 12px;">&lt;span style="background: #fef0db; padding: 4px 12px; border-radius: 40px; font-size: 0.75rem; font-weight: 600; color: #aa6f20;">WILEY&lt;/span>&lt;/div>
&lt;/div>
&lt;/div>
&lt;hr>
&lt;h2 id="-behind-the-honor-growing-with-aipatient">🧠 Behind the Honor: Growing with AIPatient&lt;/h2>
&lt;p>This award recognizes Huizi&amp;rsquo;s outstanding oral presentation on our team&amp;rsquo;s groundbreaking work published in &lt;em>Communications Medicine&lt;/em> — &lt;strong>AIPatient: Simulating Patients with EHRs and LLM Powered Agentic Workflow&lt;/strong>. As a core contributor, Huizi delivered a compelling talk on leveraging LLM-powered agentic workflows and electronic health records to build high-fidelity simulated patient systems, earning high praise from the forum judges.&lt;/p>
&lt;ul>
&lt;li>✅ &lt;strong>High-Fidelity Knowledge Graph&lt;/strong> (AIPatient KG, F1 0.89)&lt;/li>
&lt;li>✅ &lt;strong>Six-Agent Reasoning RAG Architecture&lt;/strong> achieving &lt;strong>94.15%&lt;/strong> accuracy in medical QA&lt;/li>
&lt;li>✅ &lt;strong>Superior Readability &amp;amp; Robustness&lt;/strong>, offering new paradigms for medical education and clinical decision-making&lt;/li>
&lt;/ul>
&lt;hr>
&lt;h2 id="-explore-the-original-project-page">🔗 Explore the Original Project Page&lt;/h2>
&lt;p>To dive deeper into the AIPatient framework — including core methodology, technical specifications, performance evaluations, and future directions — please visit the project page or read the full paper published in &lt;em>Communications Medicine&lt;/em>.&lt;/p>
&lt;div class="d-flex flex-wrap" style="display: flex; flex-wrap: wrap; gap: 16px; margin: 1.8em 0;">
&lt;a href="https://huiziy.github.io/aipatient-overview/" target="_blank" style="display: inline-flex; align-items: center; background-color: #1a73e8; color: white; padding: 10px 24px; border-radius: 24px; text-decoration: none; font-weight: 500; font-family: Roboto, Arial, sans-serif; box-shadow: 0 1px 3px rgba(0,0,0,0.12); transition: background-color 0.3s;">
&lt;i class="fas fa-globe" style="margin-right: 8px;">&lt;/i> Project Page
&lt;/a>
&lt;a href="https://www.nature.com/articles/s43856-025-01283-x" target="_blank" style="display: inline-flex; align-items: center; background-color: white; color: #1a73e8; border: 1px solid #dadce0; padding: 10px 24px; border-radius: 24px; text-decoration: none; font-weight: 500; font-family: Roboto, Arial, sans-serif; transition: background-color 0.3s;">
&lt;i class="fas fa-file-alt" style="margin-right: 8px;">&lt;/i> Read Published Paper
&lt;/a>
&lt;/div>
&lt;blockquote>
&lt;p>The original page covers the complete &lt;strong>Core Methodology&lt;/strong> (AIPatient Knowledge Graph + Reasoning RAG Workflow), &lt;strong>Technical Specifications&lt;/strong>, &lt;strong>Performance Evaluation&lt;/strong>, and &lt;strong>Future Directions&lt;/strong> — an essential resource for understanding this cutting-edge simulated patient system.&lt;/p>
&lt;/blockquote>
&lt;hr>
&lt;p>Warm congratulations to &lt;strong>Huizi&lt;/strong>! 🎊🎈&lt;/p>
&lt;hr>
&lt;div style="margin-top: 2.5rem; padding-top: 0.8rem; border-top: 1px solid #eaecef; font-size: 0.85rem; color: #5a626e; text-align: center;">
&lt;i class="fas fa-quote-left" style="margin-right: 6px;">&lt;/i> Shaping the future of medical simulation with agentic intelligence.&lt;br>
© AIPatient Team · Congratulatory Page | Built in the style of the original project page
&lt;/div></description></item><item><title>🏆 Our Team Wins Best Paper Award for Enhancing Medical Diagnostic Reasoning with Chain-of-Thought in Large Language Models</title><link>https://pai-cuhk.github.io/post/25-sdu/</link><pubDate>Wed, 01 Jan 2025 00:00:00 +0000</pubDate><guid>https://pai-cuhk.github.io/post/25-sdu/</guid><description>&lt;p>🎉 &lt;strong>Congratulations to our research team!&lt;/strong> We are thrilled to announce that our paper &amp;ldquo;Enhancing Medical Diagnostic Reasoning with Chain-of-Thought in Large Language Models&amp;rdquo; has been awarded the &lt;strong>Best Paper Award&lt;/strong> at the 2025 International Conference on Mechatronics, Robotics, and Artificial Intelligence (MRAI)!&lt;/p>
&lt;h2 id="-award-winning-research">🌟 Award-Winning Research&lt;/h2>
&lt;p>Our research team, led by &lt;strong>Guangxin Dai&lt;/strong>, &lt;strong>Xiang Li&lt;/strong>, &lt;strong>Lizhou Fan&lt;/strong>, and &lt;strong>Xin Ma&lt;/strong>, has achieved a significant milestone in advancing medical AI technology. This groundbreaking work represents a major step forward in enhancing the diagnostic reasoning capabilities of Large Language Models (LLMs) in clinical settings.&lt;/p>
&lt;h2 id="-research-innovation">🔬 Research Innovation&lt;/h2>
&lt;p>
&lt;figure >
&lt;div class="d-flex justify-content-center">
&lt;div class="w-100" >&lt;img alt="Chain-of-Thought Workflow" srcset="
/post/25-sdu/2025sdu-workflow_hu14670700543041567280.webp 400w,
/post/25-sdu/2025sdu-workflow_hu16113590882805442106.webp 760w,
/post/25-sdu/2025sdu-workflow_hu8927474483172039582.webp 1200w"
src="https://pai-cuhk.github.io/post/25-sdu/2025sdu-workflow_hu14670700543041567280.webp"
width="760"
height="390"
loading="lazy" data-zoomable />&lt;/div>
&lt;/div>&lt;/figure>
&lt;/p>
&lt;p>To address the critical challenge of improving diagnostic reasoning in AI-assisted healthcare, our team proposed four innovative Chain-of-Thought (CoT) prompting strategies:&lt;/p>
&lt;ol>
&lt;li>&lt;strong>Diagnosis Validation&lt;/strong> - Systematic validation of diagnostic hypotheses&lt;/li>
&lt;li>&lt;strong>Bayesian Diagnosis&lt;/strong> - Probabilistic reasoning approach for medical diagnosis&lt;/li>
&lt;li>&lt;strong>Hierarchical Reasoning&lt;/strong> - Multi-level diagnostic reasoning framework&lt;/li>
&lt;li>&lt;strong>Causal Abduction&lt;/strong> - Root cause analysis for medical conditions&lt;/li>
&lt;/ol>
&lt;p>Each CoT strategy was meticulously designed to emulate distinct modes of clinical reasoning that healthcare professionals use in real-world diagnostic scenarios.&lt;/p>
&lt;h2 id="-experimental-results">📊 Experimental Results&lt;/h2>
&lt;p>Our comprehensive evaluation utilized GPT-3.5 and GPT-4o models, with rigorous testing on the MedQA-USMLE dataset. From an initial pool of 1597 questions, we identified 119 consistently error-prone questions to create a challenging benchmark for diagnostic reasoning.&lt;/p>
&lt;p>&lt;strong>Key Performance Results on GPT-4o:&lt;/strong>&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Causal Abduction CoT&lt;/strong>: Achieved the highest accuracy&lt;/li>
&lt;li>&lt;strong>Diagnosis Validation&lt;/strong>: 31.1% accuracy improvement&lt;/li>
&lt;li>&lt;strong>Hierarchical Reasoning&lt;/strong>: 26.9% accuracy enhancement&lt;/li>
&lt;li>&lt;strong>Bayesian Diagnosis&lt;/strong>: Significant performance gains&lt;/li>
&lt;/ul>
&lt;h2 id="-clinical-impact">🎯 Clinical Impact&lt;/h2>
&lt;p>The results demonstrate that clinical reasoning-oriented Chain-of-Thought prompting can effectively improve LLM performance on challenging diagnostic tasks. This breakthrough has significant implications for:&lt;/p>
&lt;ul>
&lt;li>&lt;strong>Enhanced Clinical Decision Support&lt;/strong>: Better AI assistance for healthcare professionals&lt;/li>
&lt;li>&lt;strong>Improved Diagnostic Accuracy&lt;/strong>: Reduced medical errors through advanced reasoning&lt;/li>
&lt;li>&lt;strong>Scalable Healthcare Solutions&lt;/strong>: AI systems that can support medical practice globally&lt;/li>
&lt;/ul>
&lt;h2 id="-future-directions">🚀 Future Directions&lt;/h2>
&lt;p>This award-winning research opens new avenues for advancing AI in healthcare. Our team continues to explore how Chain-of-Thought reasoning can be further refined and applied to various medical specialties, potentially revolutionizing how AI assists in clinical decision-making.&lt;/p>
&lt;p>We extend our gratitude to the conference organizers, reviewers, and the broader research community for recognizing the significance of this work. This achievement motivates us to continue pushing the boundaries of AI in healthcare!&lt;/p></description></item><item><title>🧑🏽‍⚕️ [PUBLISHED] AIPatient: Simulating Patients with EHRs and LLM Powered Agentic Workflow</title><link>https://pai-cuhk.github.io/post/24-09-aaai/</link><pubDate>Fri, 27 Sep 2024 00:00:00 +0000</pubDate><guid>https://pai-cuhk.github.io/post/24-09-aaai/</guid><description>&lt;p>🎉 We present AIPatient, a novel framework that enhances medical education and research through an advanced simulated patient system built upon Electronic Health Records (EHRs) and powered by Large Language Models (LLMs).&lt;/p>
&lt;div class="d-flex flex-wrap" style="gap: 16px; margin: 1.5em 0;">
&lt;a href="https://huiziy.github.io/aipatient-overview/" target="_blank" style="display: inline-flex; align-items: center; background-color: #1a73e8; color: white; padding: 10px 24px; border-radius: 24px; text-decoration: none; font-weight: 500; font-family: Roboto, Arial, sans-serif; box-shadow: 0 1px 3px rgba(0,0,0,0.12), 0 1px 2px rgba(0,0,0,0.24); transition: background-color 0.3s;" onmouseover="this.style.backgroundColor='#1557b0'" onmouseout="this.style.backgroundColor='#1a73e8'">
&lt;i class="fas fa-globe" style="margin-right: 8px;">&lt;/i> Project Page
&lt;/a>
&lt;a href="https://www.nature.com/articles/s43856-025-01283-x" target="_blank" style="display: inline-flex; align-items: center; background-color: white; color: #1a73e8; border: 1px solid #dadce0; padding: 10px 24px; border-radius: 24px; text-decoration: none; font-weight: 500; font-family: Roboto, Arial, sans-serif; transition: background-color 0.3s;" onmouseover="this.style.backgroundColor='#f8f9fa'" onmouseout="this.style.backgroundColor='white'">
&lt;i class="fas fa-file-alt" style="margin-right: 8px;">&lt;/i> Read Published Paper
&lt;/a>
&lt;/div>
&lt;hr>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">🔍 Overview&lt;/h2>
&lt;div style="background-color: #f8f9fa; border-left: 4px solid #1a73e8; padding: 20px; border-radius: 4px; color: #3c4043; font-size: 1.05em; line-height: 1.6; margin-bottom: 2em;">
Simulated patient systems play a crucial role in modern medical education and research, providing safe, integrative learning environments and enabling clinical decision-making simulations. Large Language Models (LLM) could advance simulated patient systems by replicating medical conditions and patient-doctor interactions with high fidelity and low cost. However, ensuring the effectiveness and trustworthiness of these systems remains a challenge, as they require a large, diverse, and precise patient knowledgebase, along with a robust and stable knowledge diffusion to users.
&lt;p>Here, we developed AIPatient, an advanced simulated patient system with AIPatient Knowledge Graph (AIPatient KG) as the input and the Reasoning Retrieval-Augmented Generation (Reasoning RAG) agentic workflow as the generation backbone. AIPatient KG samples data from Electronic Health Records (EHRs) in the Medical Information Mart for Intensive Care (MIMIC)-III database, producing a clinically diverse and relevant cohort of 1,495 patients with high knowledgebase validity (F1 0.89). Reasoning RAG leverages six LLM powered agents spanning tasks including retrieval, KG query generation, abstraction, checker, rewrite, and summarization. This agentic framework reaches an overall accuracy of 94.15% in EHR-based medical Question Answering (QA), outperforming benchmarks that use either no agent or only partial agent integration. Our system also presents high readability (median Flesch Reading Ease 77.23; median Flesch Kincaid Grade 5.6), robustness (ANOVA F-value 0.6126, p&amp;gt;0.1), and stability (ANOVA F-value 0.782, p&amp;gt;0.1). The promising performance of the AIPatient system highlights its potential to support a wide range of applications, including medical education, model evaluation, and system integration.&lt;/p>
&lt;/div>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">⚙️ Core Methodology&lt;/h2>
&lt;p>The methodology of the AIPatient system is primarily structured around two key components:&lt;/p>
&lt;h3 style="color: #202124; margin-top: 1.5em; margin-bottom: 0.8em; font-weight: 500;">1. AIPatient Knowledge Graph (AIPatient KG)&lt;/h3>
&lt;p style="color: #3c4043;">This graph serves as the foundational knowledge base, containing patient data extracted from the Medical Information Mart for Intensive Care (MIMIC-III) database. Key features include:&lt;/p>
&lt;div style="background-color: white; border-left: 4px solid #9c27b0; padding: 16px; border-radius: 4px; box-shadow: 0 1px 2px rgba(0,0,0,0.1); margin-bottom: 2em; margin-top: 1em;">
&lt;ul style="margin: 0; padding-left: 20px; color: #5f6368; font-size: 0.95em;">
&lt;li style="margin-bottom: 8px;">&lt;b>Diverse Patient Cohort:&lt;/b> 1,495 patient profiles offering clinically diverse scenarios.&lt;/li>
&lt;li style="margin-bottom: 8px;">&lt;b>High Accuracy:&lt;/b> F1 score of 0.89, indicating substantial accuracy in capturing medical data through Named Entity Recognition (NER).&lt;/li>
&lt;li style="margin-bottom: 8px;">&lt;b>Structured Organization:&lt;/b> Medical entities such as symptoms, medical histories, allergies, and vital signs organized into nodes, interlinked by edges depicting relationships (e.g., &lt;code style="background-color: #f1f3f4; padding: 2px 4px; border-radius: 4px; font-size: 0.9em; color: #d81b60;">HAS_SYMPTOM&lt;/code>).&lt;/li>
&lt;li style="margin-bottom: 8px;">&lt;b>LLM-based NER:&lt;/b> Utilizes advanced NER techniques to transform unstructured clinical notes into structured data.&lt;/li>
&lt;li>&lt;b>Graph Database:&lt;/b> Efficiently stored and queried using Neo4j for complex relationship analysis.&lt;/li>
&lt;/ul>
&lt;/div>
&lt;h3 style="color: #202124; margin-top: 1.5em; margin-bottom: 0.8em; font-weight: 500;">2. Reasoning Retrieval-Augmented Generation Workflow&lt;/h3>
&lt;p style="color: #3c4043;">This multi-agent system provides a dynamic method for processing natural language queries and generating context-appropriate responses across three primary stages:&lt;/p>
&lt;div style="display: flex; flex-direction: column; gap: 12px; margin-bottom: 2em; margin-top: 1em;">
&lt;div style="background-color: white; border-left: 4px solid #4285f4; padding: 16px; border-radius: 4px; box-shadow: 0 1px 2px rgba(0,0,0,0.1);">
&lt;h4 style="color: #1a73e8; margin-top: 0; margin-bottom: 8px; font-size: 1.1em;">&lt;i class="fas fa-search" style="margin-right: 8px;">&lt;/i>Retrieval Stage&lt;/h4>
&lt;ul style="margin: 0; padding-left: 20px; color: #5f6368; font-size: 0.95em;">
&lt;li>&lt;b>Retrieval Agent:&lt;/b> Selects relevant nodes and relationships from the AIPatient KG based on queries.&lt;/li>
&lt;li>&lt;b>KG Query Generation Agent:&lt;/b> Formulates Cypher queries to extract pertinent information.&lt;/li>
&lt;/ul>
&lt;/div>
&lt;div style="background-color: white; border-left: 4px solid #fbbc05; padding: 16px; border-radius: 4px; box-shadow: 0 1px 2px rgba(0,0,0,0.1);">
&lt;h4 style="color: #d93025; margin-top: 0; margin-bottom: 8px; font-size: 1.1em;">&lt;i class="fas fa-brain" style="margin-right: 8px;">&lt;/i>Reasoning Stage&lt;/h4>
&lt;ul style="margin: 0; padding-left: 20px; color: #5f6368; font-size: 0.95em;">
&lt;li>&lt;b>Abstraction Agent:&lt;/b> Simplifies and rephrases user queries into more general forms.&lt;/li>
&lt;li>&lt;b>Checker Agent:&lt;/b> Verifies alignment between retrieved information and user inquiries.&lt;/li>
&lt;/ul>
&lt;/div>
&lt;div style="background-color: white; border-left: 4px solid #34a853; padding: 16px; border-radius: 4px; box-shadow: 0 1px 2px rgba(0,0,0,0.1);">
&lt;h4 style="color: #188038; margin-top: 0; margin-bottom: 8px; font-size: 1.1em;">&lt;i class="fas fa-comment-dots" style="margin-right: 8px;">&lt;/i>Generation Stage&lt;/h4>
&lt;ul style="margin: 0; padding-left: 20px; color: #5f6368; font-size: 0.95em;">
&lt;li>&lt;b>Rewrite Agent:&lt;/b> Transforms KG results into natural language with simulated patient personality traits.&lt;/li>
&lt;li>&lt;b>Summarization Agent:&lt;/b> Updates and maintains context for multi-turn interactions.&lt;/li>
&lt;/ul>
&lt;/div>
&lt;/div>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">💻 Technical Specifications&lt;/h2>
&lt;div style="background-color: white; border: 1px solid #e8eaed; border-radius: 8px; overflow: hidden; box-shadow: 0 1px 3px rgba(0,0,0,0.06); margin-bottom: 2em;">
&lt;div style="padding: 20px; border-bottom: 1px solid #e8eaed; background-color: #f8f9fa;">
&lt;h3 style="color: #1a73e8; margin: 0 0 10px 0; font-weight: 500; font-size: 1.2em;">&lt;i class="fas fa-project-diagram mr-2">&lt;/i>Named Entity Recognition (NER) &amp; Database&lt;/h3>
&lt;ul style="margin-bottom: 0; color: #3c4043; padding-left: 20px;">
&lt;li>LLM-based NER system for extracting medical entities from unstructured discharge summary data.&lt;/li>
&lt;li>Establishes distinct relationships (e.g., "HAS_FAMILY_MEMBER") within the graph.&lt;/li>
&lt;li>Hosted in &lt;b>Neo4j&lt;/b> for complex multi-faceted relationship queries instead of linear retrieval.&lt;/li>
&lt;li>Utilizes &lt;b>Cypher Query Language&lt;/b> to navigate graph functionalities efficiently.&lt;/li>
&lt;/ul>
&lt;/div>
&lt;div style="padding: 20px;">
&lt;h3 style="color: #1a73e8; margin: 0 0 10px 0; font-weight: 500; font-size: 1.2em;">&lt;i class="fas fa-chart-line mr-2">&lt;/i>Performance Evaluation&lt;/h3>
&lt;ul style="margin-bottom: 0; color: #3c4043; padding-left: 20px;">
&lt;li>&lt;b>Ablation Studies:&lt;/b> Configurations using Retrieval &amp; Abstraction Agents significantly outperformed simpler setups.&lt;/li>
&lt;li>&lt;b>Overall QA Accuracy:&lt;/b> 94.15% in EHR-based medical Question Answering.&lt;/li>
&lt;li>&lt;b>Readability Metrics:&lt;/b> Median Flesch Reading Ease (77.23) &amp; M. Flesch-Kincaid Grade Level (5.6).&lt;/li>
&lt;li>&lt;b>Robustness &amp; Stability:&lt;/b> ANOVA F-value 0.6126 (p>0.1) and 0.782 (p>0.1) respectively.&lt;/li>
&lt;/ul>
&lt;/div>
&lt;/div>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">🌟 Key Contributions&lt;/h2>
&lt;div style="background-color: white; border: 1px solid #e8eaed; border-radius: 8px; padding: 20px; box-shadow: 0 1px 3px rgba(0,0,0,0.06); margin-bottom: 2em;">
&lt;ul style="list-style: none; padding-left: 0; margin: 0;">
&lt;li style="margin-bottom: 12px; display: flex; align-items: flex-start;">
&lt;span style="color: #34a853; margin-right: 12px; font-size: 1.2em;">✓&lt;/span>
&lt;div>&lt;b>Advanced Knowledge Graph:&lt;/b> AIPatient KG provides a robust foundation with high validity (F1 0.89) derived from real EHR data.&lt;/div>
&lt;/li>
&lt;li style="margin-bottom: 12px; display: flex; align-items: flex-start;">
&lt;span style="color: #34a853; margin-right: 12px; font-size: 1.2em;">✓&lt;/span>
&lt;div>&lt;b>Multi-Agent Architecture:&lt;/b> Six specialized LLM agents working in concert for comprehensive patient simulation.&lt;/div>
&lt;/li>
&lt;li style="margin-bottom: 12px; display: flex; align-items: flex-start;">
&lt;span style="color: #34a853; margin-right: 12px; font-size: 1.2em;">✓&lt;/span>
&lt;div>&lt;b>Superior Performance:&lt;/b> 94.15% accuracy outperforming existing benchmarks.&lt;/div>
&lt;/li>
&lt;li style="margin-bottom: 12px; display: flex; align-items: flex-start;">
&lt;span style="color: #34a853; margin-right: 12px; font-size: 1.2em;">✓&lt;/span>
&lt;div>&lt;b>High Readability:&lt;/b> Ensures accessible communication for educational purposes.&lt;/div>
&lt;/li>
&lt;li style="display: flex; align-items: flex-start;">
&lt;span style="color: #34a853; margin-right: 12px; font-size: 1.2em;">✓&lt;/span>
&lt;div>&lt;b>Robustness and Stability:&lt;/b> Consistent performance across varied scenarios and question formulations.&lt;/div>
&lt;/li>
&lt;/ul>
&lt;/div>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">🚀 Applications and Impact&lt;/h2>
&lt;p style="color: #3c4043; font-size: 1.05em; margin-bottom: 1.5em;">The AIPatient framework offers transformative potential for:&lt;/p>
&lt;div class="d-flex flex-wrap" style="gap: 12px; margin-bottom: 2em;">
&lt;span style="background-color: #e8f0fe; color: #1967d2; padding: 8px 16px; border-radius: 20px; font-size: 0.95em; font-weight: 500;">&lt;i class="fas fa-user-md mr-1">&lt;/i> Medical Education&lt;/span>
&lt;span style="background-color: #fce8e6; color: #c5221f; padding: 8px 16px; border-radius: 20px; font-size: 0.95em; font-weight: 500;">&lt;i class="fas fa-stethoscope mr-1">&lt;/i> Clinical Decision-Making&lt;/span>
&lt;span style="background-color: #e6f4ea; color: #137333; padding: 8px 16px; border-radius: 20px; font-size: 0.95em; font-weight: 500;">&lt;i class="fas fa-microscope mr-1">&lt;/i> Research &amp; Evaluation&lt;/span>
&lt;span style="background-color: #fef7e0; color: #b06000; padding: 8px 16px; border-radius: 20px; font-size: 0.95em; font-weight: 500;">&lt;i class="fas fa-book-open mr-1">&lt;/i> Curriculum Development&lt;/span>
&lt;/div>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">🔮 Future Directions&lt;/h2>
&lt;div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(250px, 1fr)); gap: 16px; margin-bottom: 2em;">
&lt;div style="background-color: #f8f9fa; border-radius: 8px; padding: 16px; border-left: 4px solid #1a73e8;">
&lt;h4 style="margin-top: 0; color: #202124; font-size: 1.05em;">&lt;i class="fas fa-database" style="color: #1a73e8; margin-right: 8px;">&lt;/i> Broader Data Scope&lt;/h4>
&lt;p style="margin-bottom: 0; font-size: 0.95em; color: #5f6368;">Incorporating wider variety of healthcare scenarios for more extensive knowledge bases.&lt;/p>
&lt;/div>
&lt;div style="background-color: #f8f9fa; border-radius: 8px; padding: 16px; border-left: 4px solid #34a853;">
&lt;h4 style="margin-top: 0; color: #202124; font-size: 1.05em;">&lt;i class="fas fa-photo-video" style="color: #34a853; margin-right: 8px;">&lt;/i> Multimodal Integration&lt;/h4>
&lt;p style="margin-bottom: 0; font-size: 0.95em; color: #5f6368;">Adding medical imaging and other data modalities to simulate full clinical inputs.&lt;/p>
&lt;/div>
&lt;div style="background-color: #f8f9fa; border-radius: 8px; padding: 16px; border-left: 4px solid #fbbc05;">
&lt;h4 style="margin-top: 0; color: #202124; font-size: 1.05em;">&lt;i class="fas fa-hospital" style="color: #fbbc05; margin-right: 8px;">&lt;/i> Real-time Clinical&lt;/h4>
&lt;p style="margin-bottom: 0; font-size: 0.95em; color: #5f6368;">Adapting environments for live clinical settings and interactive deployment.&lt;/p>
&lt;/div>
&lt;div style="background-color: #f8f9fa; border-radius: 8px; padding: 16px; border-left: 4px solid #ea4335;">
&lt;h4 style="margin-top: 0; color: #202124; font-size: 1.05em;">&lt;i class="fas fa-chalkboard-teacher" style="color: #ea4335; margin-right: 8px;">&lt;/i> Enhanced Training&lt;/h4>
&lt;p style="margin-bottom: 0; font-size: 0.95em; color: #5f6368;">Further enriching educational landscapes and outcomes for healthcare professionals.&lt;/p>
&lt;/div>
&lt;/div>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">📝 Conclusion&lt;/h2>
&lt;p>The AIPatient framework exemplifies a transformative approach to patient simulation through EHR data and LLM-powered interactions, emphasizing structured retrieval and knowledge dissemination in a manner conducive to enhancing medical training and decision-making processes. By embedding these systems within medical curricula and research protocols, we propose significant advancements in how interactions with simulated patients can impact educational outcomes and clinical preparedness.&lt;/p>
&lt;hr>
&lt;h2 style="color: #1a73e8; border-bottom: 2px solid #e8eaed; padding-bottom: 10px; margin-top: 2em; margin-bottom: 1em; font-weight: 500;">📖 Welcome to Cite Our Article&lt;/h2>
&lt;p>If you find our work helpful, please consider citing it:&lt;/p>
&lt;div style="background-color: #f8f9fa; border: 1px solid #e1e4e8; border-radius: 8px; margin-top: 1.5em; overflow: hidden; box-shadow: 0 2px 8px rgba(0,0,0,0.05); max-width: 100%;">
&lt;div style="background-color: #f1f3f4; padding: 10px 16px; border-bottom: 1px solid #e1e4e8; display: flex; justify-content: space-between; align-items: center;">
&lt;span style="color: #5f6368; font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Helvetica, Arial, sans-serif; font-size: 0.85em; font-weight: 600; text-transform: uppercase; letter-spacing: 0.5px;">BibTeX&lt;/span>
&lt;/div>
&lt;div style="padding: 20px; overflow-x: auto;">
&lt;pre style="margin: 0; background: transparent; border: none; padding: 0;">&lt;code style="font-family: 'SFMono-Regular', Consolas, 'Liberation Mono', Menlo, Courier, monospace; font-size: 0.95em; line-height: 1.5; color: #24292e;">@ARTICLE{Yu2025-tb,
title = "Simulated patient systems powered by large language model-based
{AI} agents offer potential for transforming medical education",
author = "Yu, Huizi and Zhou, Jiayan and Li, Lingyao and Chen, Shan and
Gallifant, Jack and Shi, Anye and Sun, Jie and Li, Xiang and He,
Jingxian and Hua, Wenyue and Jin, Mingyu and Chen, Guang and
Zhou, Yang and Li, Zhao and Gupte, Trisha and Chen, Ming-Li and
Azizi, Zahra and Dou, Qi and Yan, Bryan P and Xing, Yanqiu and
Zhang, Yongfeng and Assimes, Themistocles L and Bitterman,
Danielle S and Ma, Xin and Lu, Lin and Fan, Lizhou",
journal = "Commun. Med. (Lond.)",
publisher = "Springer Science and Business Media LLC",
volume = 6,
number = 1,
pages = 27,
month = dec,
year = 2025,
language = "en"
}&lt;/code>&lt;/pre>
&lt;/div>
&lt;/div></description></item></channel></rss>