<p><em>歐洲專利局（EPO）已授予 Ada Health 一項專利，涉及其混合人工智能架構——一種結合大型語言模型優勢的集成方法，具備臨牀精確性、可解釋性和合規性。</em></p>
<p>, /PRNewswire/ -- Ada Health 是一家領先的臨牀人工智能公司，幫助人們理解症狀並安全高效地導航醫療，今天宣佈 EPO 已授予 Ada 的混合臨牀人工智能架構專利。該專利認可了 Ada 的集成方法的新穎性：一個將大型語言模型與其專有的概率圖模型相結合的系統，提供比單獨使用任何一種方法更準確、更用户友好的解決方案。</p>
<p>人工智能現在已成為大規模健康問題的主要入口。但通用大型語言模型的臨牀安全性弱點越來越受到關注：獨立研究已識別出緊急情況的低級別分類、錯誤和幻覺。Ada 的系統提供了現代語言模型的對話靈活性與醫療所需的臨牀嚴謹性，而不是在兩者之間做出選擇。</p>
<p><em>&#34;大型語言模型提供商正在發現規模與醫療安全並不相同。Ada 花費了十多年時間構建使人工智能在醫療中安全部署的臨牀基礎設施——經過驗證的知識庫、監管認證和安全協議。該專利保護了將所有這些與現代語言模型的力量結合在一起的架構。這是行業一直在等待的臨牀層。&#34;</em></p>
<p><strong>— Daniel Nathrath，Ada Health 創始人兼首席執行官</strong></p>
<p><strong>專利保護的內容</strong></p>
<p>Ada 的混合引擎在前端包含一個大型語言模型，允許用户用自己的話描述症狀、病史、藥物和偏好，捕捉到對複雜或罕見病症的預診斷至關重要的臨牀細微差別和上下文細節。該系統動態提示用户與醫療相關的後續問題，限制在 Ada 的經過驗證的知識庫內，防止因在臨牀環境中不受限制使用大型語言模型而產生的無根據輸出。這是內置於架構中的結構性保障。最後，專利混合模型獨立評估症狀，LLM 作為 Ada 的專家策劃醫療知識庫的第二意見。兩個模型是集成的：Ada 的推理引擎充當護欄，同時提高相對準確性超過 7%，並保持醫療級部署所需的嚴謹性。</p>
<p>最終結果是一個完全透明、可審計的臨牀推理系統。自 2016 年以來，Ada 的輸出數據洞察和可解釋性在大規模人羣中運行，繼續成為其全球領先的製藥、健康系統和支付方合作伙伴的重要特徵。</p>
<p><em>&#34;工程挑戰不在於構建一個混合體，而在於構建一個臨牀嚴謹性與對話靈活性相互增強而非相互妥協的混合體。Ada 的架構通過設計解決了這一點：大型語言模型捕捉用户想要表達的內容，Ada 的概率模型和集成解決方案確定其臨牀意義，而 Ada 的推理引擎始終是決策者。這使得它在受監管的環境中安全、可解釋且可部署。&#34;</em></p>
<p><strong>— Graham French，Ada Health 首席技術官</strong></p>
<p>在 EPO 授予專利後，Ada 將把其混合方法擴展到其製藥、健康系統和支付方合作伙伴中，安全性和可解釋性是前提條件。</p>
<p><strong>關於專利</strong></p>
<p>涵蓋 Ada Health 混合架構的專利（EP4679451B1）已由歐洲專利局正式授予，並於 2026 年 3 月 25 日生效，優先日期為 2024 年 7 月 12 日。</p>
<p><strong>關於 Ada Health</strong></p>
<p>Ada Health 結合臨牀專業知識和先進的人工智能，幫助個人和健康組織更早地採取行動。由醫生、科學家和工程師構建，Ada 應用概率推理技術於一個醫療專家策劃的知識庫，將現實世界數據轉化為可操作的指導，賦能個人理解其症狀，並幫助醫療和生命科學領域的合作伙伴提供更及時、個性化的護理。作為歐盟醫療器械法規（MDR）下的 IIa 類醫療器械，Ada 的評估已進行超過 5000 萬次，並有 50 多篇獨立的同行評審出版物，受到全球領先健康系統、保險公司和製藥公司的信任。瞭解更多信息，請訪問 about.ada.com。</p>
<p>Logo - https://mma.prnewswire.com/media/2953902/Ada_Health_Logo.jpg</p>
<p>來源：Ada Health</p>
<p><img src="https://imageproxy.pbkrs.com/https://rt.prnewswire.com/rt.gif/query-RGF0ZUlkPTIwMjYwNDEzJk5ld3NJdGVtSWQ9TE4zMTYzOCZUcmFuc21pc3Npb25fSWQ9MjAyNjA0MTMwODI0UFJfTkVXU19VU1BSX19fX19MTjMxNjM4?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://rt.prnewswire.com/rt.gif/query-RGF0ZUlkPTIwMjYwNDEzJk5ld3NJdGVtSWQ9TE4zMTYzOCZUcmFuc21pc3Npb25fSWQ9MjAyNjA0MTMwODI0UFJfTkVXU19VU1BSX19fX19MTjMxNjM4"/></p>

PRTC

阿卡迪亞醫療保健

AHCO

PRTCY

<p>Ada Health 已獲得歐洲專利局授予的專利，專利內容為其混合人工智能架構，該架構將大型語言模型（LLMs）與臨牀精確性和合規性相結合。該創新系統通過將概率圖模型與 LLM 集成，提高了醫療保健的準確性和用户友好性。該專利架構通過動態提示用户相關的醫學問題，並通過 Ada 的專家策劃知識庫提供第二意見，從而確保臨牀安全。此項發展旨在改善人工智能在醫療保健中的應用，解決與通用 LLM 相關的安全問題</p>

<p><em>The European Patent Office (EPO) has granted Ada Health a patent for its hybrid AI architecture — an ensembling approach that combines the strengths of large language models with clinical precision, explainability, and regulatory standing.</em></p>
<div class="lb-trans"><p><em>歐洲專利局（EPO）已授予 Ada Health 一項專利，涉及其混合人工智能架構——一種結合大型語言模型優勢的集成方法，具備臨牀精確性、可解釋性和合規性。</em></p>
</div><p>, /PRNewswire/ -- Ada Health, a leading clinical AI company that helps people understand their symptoms and navigate care safely and efficiently, today announced that the EPO has granted a patent for Ada's hybrid clinical AI architecture. The patent recognizes the novelty of Ada's ensembling approach: a system that combines an LLM with its proprietary probabilistic graphical model to deliver a solution that is more accurate and user-friendly than either approach alone.</p>
<div class="lb-trans"><p>, /PRNewswire/ -- Ada Health 是一家領先的臨牀人工智能公司，幫助人們理解症狀並安全高效地導航醫療，今天宣佈 EPO 已授予 Ada 的混合臨牀人工智能架構專利。該專利認可了 Ada 的集成方法的新穎性：一個將大型語言模型與其專有的概率圖模型相結合的系統，提供比單獨使用任何一種方法更準確、更用户友好的解決方案。</p>
</div><p>AI is now a primary entry point for health questions at scale. But the clinical safety weaknesses of general-purpose LLMs are increasingly well-documented: independent research has identified under-triage of emergencies, errors, and hallucinations. Rather than choosing between the conversational flexibility of modern language models and the clinical rigour that healthcare demands, Ada's system delivers both.</p>
<div class="lb-trans"><p>人工智能現在已成為大規模健康問題的主要入口。但通用大型語言模型的臨牀安全性弱點越來越受到關注：獨立研究已識別出緊急情況的低級別分類、錯誤和幻覺。Ada 的系統提供了現代語言模型的對話靈活性與醫療所需的臨牀嚴謹性，而不是在兩者之間做出選擇。</p>
</div><p><em>&#34;LLM providers are discovering that scale and medical safety are not the same. Ada has spent over a decade building the clinical infrastructure that makes AI safe to deploy in healthcare — the validated knowledge base, the regulatory certification, the safeguarding protocols. This patent protects the architecture that brings all of that together with the power of modern language models. It is the clinical layer the industry has been waiting for.&#34;</em></p>
<div class="lb-trans"><p><em>&#34;大型語言模型提供商正在發現規模與醫療安全並不相同。Ada 花費了十多年時間構建使人工智能在醫療中安全部署的臨牀基礎設施——經過驗證的知識庫、監管認證和安全協議。該專利保護了將所有這些與現代語言模型的力量結合在一起的架構。這是行業一直在等待的臨牀層。&#34;</em></p>
</div><p><strong>— Daniel Nathrath, Founder and CEO, Ada Health</strong></p>
<div class="lb-trans"><p><strong>— Daniel Nathrath，Ada Health 創始人兼首席執行官</strong></p>
</div><p><strong>What the Patent Protects</strong></p>
<div class="lb-trans"><p><strong>專利保護的內容</strong></p>
</div><p>Ada's hybrid engine includes an LLM on the front end, which allows users to describe symptoms, medical history, medications, and preferences in their own words, capturing the clinical nuance and contextual detail crucial to pre-diagnosing complex or rare conditions. The system dynamically prompts users with medically relevant follow-up questions constrained to Ada's validated knowledge base, preventing ungrounded outputs that arise from unconstrained LLM use in clinical settings. This is a structural safeguard built into the architecture. Finally, the patented hybrid model assesses symptoms independently, with the LLM functioning as a second opinion alongside Ada's expert-curated medical knowledge base. The two models are ensembled: Ada's reasoning engine acts as a guardrail, while improving relative accuracy by more than 7% and preserving the rigour that medical-grade deployments demand.</p>
<div class="lb-trans"><p>Ada 的混合引擎在前端包含一個大型語言模型，允許用户用自己的話描述症狀、病史、藥物和偏好，捕捉到對複雜或罕見病症的預診斷至關重要的臨牀細微差別和上下文細節。該系統動態提示用户與醫療相關的後續問題，限制在 Ada 的經過驗證的知識庫內，防止因在臨牀環境中不受限制使用大型語言模型而產生的無根據輸出。這是內置於架構中的結構性保障。最後，專利混合模型獨立評估症狀，LLM 作為 Ada 的專家策劃醫療知識庫的第二意見。兩個模型是集成的：Ada 的推理引擎充當護欄，同時提高相對準確性超過 7%，並保持醫療級部署所需的嚴謹性。</p>
</div><p>The result is a fully white-box, auditable clinical reasoning system. The data insight and explainability of Ada's outputs, which have been running at population scale since 2016, continues to be a crucial feature for its industry-leading pharma, health systems, and payer partners around the world.</p>
<div class="lb-trans"><p>最終結果是一個完全透明、可審計的臨牀推理系統。自 2016 年以來，Ada 的輸出數據洞察和可解釋性在大規模人羣中運行，繼續成為其全球領先的製藥、健康系統和支付方合作伙伴的重要特徵。</p>
</div><p><em>&#34;The engineering challenge is not building a hybrid, it's making one where clinical rigour and conversational flexibility reinforce each other rather than trade off. Ada's architecture resolves that by design: the LLM captures what users want to say, Ada's probabilistic model and ensemble solution determine what it means clinically, and Ada's reasoning engine is always the decision-maker. That is what makes it safe, explainable, and deployable in regulated environments.&#34;</em></p>
<div class="lb-trans"><p><em>&#34;工程挑戰不在於構建一個混合體，而在於構建一個臨牀嚴謹性與對話靈活性相互增強而非相互妥協的混合體。Ada 的架構通過設計解決了這一點：大型語言模型捕捉用户想要表達的內容，Ada 的概率模型和集成解決方案確定其臨牀意義，而 Ada 的推理引擎始終是決策者。這使得它在受監管的環境中安全、可解釋且可部署。&#34;</em></p>
</div><p><strong>— Graham French, Chief Technology Officer, Ada Health</strong></p>
<div class="lb-trans"><p><strong>— Graham French，Ada Health 首席技術官</strong></p>
</div><p>Following the EPO patent grant, Ada will extend its hybrid approach across its pharma, health systems, and payers partners, where safety and explainability are prerequisites.</p>
<div class="lb-trans"><p>在 EPO 授予專利後，Ada 將把其混合方法擴展到其製藥、健康系統和支付方合作伙伴中，安全性和可解釋性是前提條件。</p>
</div><p><strong>About the Patent</strong></p>
<div class="lb-trans"><p><strong>關於專利</strong></p>
</div><p>The patent (EP4679451B1) covering Ada Health's hybrid architecture was officially granted by the European Patent Office and went into effect on March 25, 2026, with a priority date of July 12, 2024.</p>
<div class="lb-trans"><p>涵蓋 Ada Health 混合架構的專利（EP4679451B1）已由歐洲專利局正式授予，並於 2026 年 3 月 25 日生效，優先日期為 2024 年 7 月 12 日。</p>
</div><p><strong>About Ada Health</strong></p>
<div class="lb-trans"><p><strong>關於 Ada Health</strong></p>
</div><p>Ada Health combines clinical expertise and advanced AI to help people and health organizations act on insights earlier. Built by doctors, scientists, and engineers, Ada applies probabilistic reasoning technology to a medical expert-curated knowledge base that transforms real-world data into actionable guidance, empowering individuals to understand their symptoms and helping partners across healthcare and life sciences deliver more timely, personalized care. As a Class IIa medical device under the EU's Medical Device Regulation (MDR), with over 50 million assessments and 50+ independent peer-reviewed publications, Ada's assessment is trusted by leading health systems, insurers, and pharmaceutical companies around the world. Learn more at about.ada.com.</p>
<div class="lb-trans"><p>Ada Health 結合臨牀專業知識和先進的人工智能，幫助個人和健康組織更早地採取行動。由醫生、科學家和工程師構建，Ada 應用概率推理技術於一個醫療專家策劃的知識庫，將現實世界數據轉化為可操作的指導，賦能個人理解其症狀，並幫助醫療和生命科學領域的合作伙伴提供更及時、個性化的護理。作為歐盟醫療器械法規（MDR）下的 IIa 類醫療器械，Ada 的評估已進行超過 5000 萬次，並有 50 多篇獨立的同行評審出版物，受到全球領先健康系統、保險公司和製藥公司的信任。瞭解更多信息，請訪問 about.ada.com。</p>
</div><p>Logo - https://mma.prnewswire.com/media/2953902/Ada_Health_Logo.jpg</p>
<div class="lb-trans"><p>Logo - https://mma.prnewswire.com/media/2953902/Ada_Health_Logo.jpg</p>
</div><p>SOURCE Ada Health</p>
<div class="lb-trans"><p>來源：Ada Health</p>
</div><p><img src="https://imageproxy.pbkrs.com/https://rt.prnewswire.com/rt.gif/query-RGF0ZUlkPTIwMjYwNDEzJk5ld3NJdGVtSWQ9TE4zMTYzOCZUcmFuc21pc3Npb25fSWQ9MjAyNjA0MTMwODI0UFJfTkVXU19VU1BSX19fX19MTjMxNjM4" alt="" original-src="https://rt.prnewswire.com/rt.gif?NewsItemId=LN31638&amp;Transmission_Id=202604130824PR_NEWS_USPR_____LN31638&amp;DateId=20260413"/></p>

Ada Health 為確保大型語言模型在醫療保健中安全的臨牀層申請了專利