https://www.yiliao.com/index.php?action=history&feed=atom&title=OpenEvidence
OpenEvidence - 版本历史
2026-04-09T20:49:47Z
本wiki的该页面的版本历史
MediaWiki 1.35.1
https://www.yiliao.com/index.php?title=OpenEvidence&diff=316211&oldid=prev
77921020：建立内容为“<div style="padding: 0 4%; line-height: 1.8; color: #1e293b; font-family: 'Helvetica Neue', Helvetica, 'PingFang SC', Arial, sans-serif; background-color: #ffffff…”的新页面
2026-02-02T23:30:07Z
<p>建立内容为“<div style="padding: 0 4%; line-height: 1.8; color: #1e293b; font-family: 'Helvetica Neue', Helvetica, 'PingFang SC', Arial, sans-serif; background-color: #ffffff…”的新页面</p>
<p><b>新页面</b></p><div><div style="padding: 0 4%; line-height: 1.8; color: #1e293b; font-family: 'Helvetica Neue', Helvetica, 'PingFang SC', Arial, sans-serif; background-color: #ffffff; max-width: 1200px; margin: auto;"><br />
<br />
<div style="margin-bottom: 30px; border-bottom: 1.2px solid #e2e8f0; padding-bottom: 25px;"><br />
<p style="font-size: 1.1em; margin: 10px 0; color: #334155; text-align: justify;"><br />
<strong>Open Evidence</strong> 是一个专为医疗卫生专业人士设计的、基于人工智能的医学信息检索与决策支持平台。与通用的 AI 模型（如 ChatGPT）不同，Open Evidence 采用了先进的<strong>检索增强生成 (Retrieval-Augmented Generation, RAG)</strong> 技术，确保其生成的每一个回答都严格基于高质量的、同行评审的医学文献（如 PubMed, NCCN 指南, 权威医学期刊）。Open Evidence 的核心价值在于<strong>“无幻觉” (Hallucination-free)</strong> 和<strong>“可溯源” (Verifiable)</strong>——它会在回答的每一句话后标注具体的引文出处，点击即可跳转至原始论文，从而让医生和研究者能够快速、放心地获取准确的循证医学证据。<br />
</p><br />
</div><br />
<br />
<div class="medical-infobox mw-collapsible mw-collapsed" style="width: 100%; max-width: 320px; margin: 0 auto 35px auto; border: 1.2px solid #bae6fd; border-radius: 12px; background-color: #ffffff; box-shadow: 0 8px 20px rgba(0,0,0,0.05); overflow: hidden;"><br />
<br />
<div style="padding: 15px; color: #1e40af; background: linear-gradient(135deg, #e0f2fe 0%, #bae6fd 100%); text-align: center; cursor: pointer;"><br />
<div style="font-size: 1.2em; font-weight: bold; letter-spacing: 1.2px;">Open Evidence</div><br />
<div style="font-size: 0.7em; opacity: 0.85; margin-top: 4px; white-space: nowrap;">AI-Powered Medical Search (点击展开)</div><br />
</div><br />
<br />
<div class="mw-collapsible-content"><br />
<div style="padding: 25px; text-align: center; background-color: #f8fafc;"><br />
<div style="width: 100px; height: 100px; background-color: #e2e8f0; border-radius: 50%; margin: 0 auto; display: flex; align-items: center; justify-content: center; color: #94a3b8; font-size: 0.8em;"><br />
<br />
</div><br />
<div style="font-size: 0.8em; color: #64748b; margin-top: 12px; font-weight: 600;">核心：句句有出处</div><br />
</div><br />
<br />
<table style="width: 100%; border-spacing: 0; border-collapse: collapse; font-size: 0.85em;"><br />
<tr><br />
<th colspan="2" style="padding: 8px 12px; background-color: #e0f2fe; color: #1e40af; text-align: left; font-size: 0.9em; border-top: 1px solid #bae6fd;">平台参数</th><br />
</tr><br />
<tr><br />
<th style="text-align: left; padding: 6px 12px; background-color: #f8fafc; color: #475569; border-bottom: 1px solid #e2e8f0; width: 40%;">技术架构</th><br />
<td style="padding: 6px 12px; border-bottom: 1px solid #e2e8f0; color: #0f172a;">LLM + RAG (检索增强)</td><br />
</tr><br />
<tr><br />
<th style="text-align: left; padding: 6px 12px; background-color: #f8fafc; color: #475569; border-bottom: 1px solid #e2e8f0;">数据来源</th><br />
<td style="padding: 6px 12px; border-bottom: 1px solid #e2e8f0; color: #0f172a;">PubMed, ClinicalTrials, Guidelines</td><br />
</tr><br />
<tr><br />
<th style="text-align: left; padding: 6px 12px; background-color: #f8fafc; color: #475569; border-bottom: 1px solid #e2e8f0;">目标用户</th><br />
<td style="padding: 6px 12px; border-bottom: 1px solid #e2e8f0; color: #1e40af;">医生, 药剂师, 研究员</td><br />
</tr><br />
<tr><br />
<th style="text-align: left; padding: 6px 12px; background-color: #f8fafc; color: #475569; border-bottom: 1px solid #e2e8f0;">核心优势</th><br />
<td style="padding: 6px 12px; border-bottom: 1px solid #e2e8f0; color: #16a34a;"><strong>透明性 (Transparency)</strong></td><br />
</tr><br />
<tr><br />
<th style="text-align: left; padding: 6px 12px; background-color: #f8fafc; color: #475569;">竞品</th><br />
<td style="padding: 6px 12px; color: #0f172a;">UpToDate, DynaMed, ChatGPT</td><br />
</tr><br />
</table><br />
</div><br />
</div><br />
<br />
<h2 style="background: #f1f5f9; color: #0f172a; padding: 10px 18px; border-radius: 0 6px 6px 0; font-size: 1.25em; margin-top: 40px; border-left: 6px solid #0f172a; font-weight: bold;">技术原理：打破“黑箱”</h2><br />
<br />
<p style="margin: 15px 0; text-align: justify;"><br />
Open Evidence 解决了通用大模型在医学领域最大的痛点——<strong>一本正经地胡说八道 (Hallucination)</strong>。其工作流程如下：<br />
<br />
</p><br />
<br />
<ul style="padding-left: 25px; color: #334155;"><br />
<li style="margin-bottom: 12px;"><strong>实时检索 (Retrieval)：</strong> 当用户提问（例如“奥希替尼在一线非小细胞肺癌的最新生存数据”）时，系统首先在千万级的医学文献库中进行高精度搜索，而非直接让 AI 凭记忆生成。</li><br />
<li style="margin-bottom: 12px;"><strong>证据筛选 (Grounding)：</strong> 系统筛选出最相关、最权威的片段（如来自《NEJM》或《Lancet》的摘要）。</li><br />
<li style="margin-bottom: 12px;"><strong>溯源生成 (Citing)：</strong> 大模型基于这些筛选出的片段生成回答，并强制要求在句尾加上数字角标，链接回原始文献。如果找不到证据，系统会倾向于回答“未知”而非编造。</li><br />
</ul><br />
<br />
<h2 style="background: #f1f5f9; color: #0f172a; padding: 10px 18px; border-radius: 0 6px 6px 0; font-size: 1.25em; margin-top: 40px; border-left: 6px solid #0f172a; font-weight: bold;">工具对比：新旧范式之争</h2><br />
<br />
<div style="background-color: #fff7ed; border-left: 5px solid #ea580c; padding: 15px 20px; margin: 20px 0; border-radius: 4px;"><br />
<p style="margin: 0; text-align: justify; font-size: 0.95em; color: #334155;"><br />
<strong>定位差异：</strong> <strong>UpToDate</strong> 是由专家人工撰写的静态“电子书”，权威但更新慢；<strong>ChatGPT</strong> 是通才，但这容易编造数据；<strong>Open Evidence</strong> 是“阅读速度极快的 AI 助手”，兼顾了时效性与真实性。<br />
</p><br />
</div><br />
<br />
<div style="overflow-x: auto; margin: 20px auto;"><br />
<table style="width: 100%; border-collapse: collapse; border: 1.2px solid #cbd5e1; font-size: 0.9em; text-align: left;"><br />
<tr style="background-color: #f1f5f9; border-bottom: 2px solid #0f172a;"><br />
<th style="padding: 12px; border: 1px solid #cbd5e1; color: #0f172a; width: 25%;">特性</th><br />
<th style="padding: 12px; border: 1px solid #cbd5e1; color: #1e40af; width: 35%;">Open Evidence</th><br />
<th style="padding: 12px; border: 1px solid #cbd5e1; color: #475569;">UpToDate / DynaMed</th><br />
</tr><br />
<tr><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; font-weight: 600;">内容生成</td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; color: #16a34a;"><strong>AI 实时聚合</strong></td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1;">专家人工撰写</td><br />
</tr><br />
<tr><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; font-weight: 600;">时效性</td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; color: #16a34a;"><strong>极高 (覆盖昨日新刊)</strong></td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1;">中等 (需等待专家更新)</td><br />
</tr><br />
<tr><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; font-weight: 600;">权威性</td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1;">依赖原始文献质量</td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; color: #1e40af;"><strong>金标准 (专家背书)</strong></td><br />
</tr><br />
<tr><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; font-weight: 600; background-color: #eff6ff;">引用颗粒度</td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; background-color: #eff6ff;"><strong>精确到句 (Inline citation)</strong></td><br />
<td style="padding: 10px; border: 1px solid #cbd5e1; background-color: #eff6ff;">段落/章节级引用</td><br />
</tr><br />
</table><br />
</div><br />
<br />
<div style="font-size: 0.92em; line-height: 1.6; color: #1e293b; margin-top: 50px; border-top: 2px solid #0f172a; padding: 15px 25px; background-color: #f8fafc; border-radius: 0 0 10px 10px;"><br />
<span style="color: #0f172a; font-weight: bold; font-size: 1.05em; display: inline-block; margin-bottom: 15px;">学术参考文献 [Academic Review]</span><br />
<br />
<p style="margin: 12px 0; border-bottom: 1px solid #e2e8f0; padding-bottom: 10px;"><br />
[1] <strong>Thirunavukarasu AJ, et al. (2023).</strong> <em>Large language models in medicine.</em> <strong>[[Nature Medicine]]</strong>. <br><br />
<span style="color: #475569;">[点评]：综述了 LLM 在医学中的应用，特别强调了 RAG 技术（即 Open Evidence 的核心）在提高准确性和可验证性方面的关键作用。</span><br />
</p><br />
<br />
<p style="margin: 12px 0; border-bottom: 1px solid #e2e8f0; padding-bottom: 10px;"><br />
[2] <strong>Singhal K, et al. (2023).</strong> <em>Large language models encode clinical knowledge.</em> <strong>[[Nature]]</strong>. <br><br />
<span style="color: #475569;">[点评]：Google DeepMind 团队的研究，讨论了 Med-PaLM 等模型，虽然 Open Evidence 使用不同架构，但该文奠定了 AI 医疗问答的评估标准。</span><br />
</p><br />
<br />
<p style="margin: 12px 0;"><br />
[3] <strong>Lewis P, et al. (2020).</strong> <em>Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks.</em> <strong>[[NeurIPS]]</strong>. <br><br />
<span style="color: #475569;">[点评]：RAG 技术的奠基之作，解释了如何通过结合检索器和生成器来解决大模型的知识幻觉问题。</span><br />
</p><br />
</div><br />
<br />
<div style="margin: 40px 0; border: 1px solid #e2e8f0; border-radius: 8px; overflow: hidden; font-family: 'Helvetica Neue', Arial, sans-serif; font-size: 0.9em;"><br />
<div style="background-color: #eff6ff; color: #1e40af; padding: 8px 15px; font-weight: bold; text-align: center; border-bottom: 1px solid #dbeafe;"><br />
医疗 AI 技术 · 知识图谱<br />
</div><br />
<table style="width: 100%; border-collapse: collapse; background-color: #ffffff;"><br />
<tr style="border-bottom: 1px solid #f1f5f9;"><br />
<td style="width: 85px; background-color: #f8fafc; color: #334155; font-weight: 600; padding: 10px 12px; text-align: right; vertical-align: middle;">核心技术</td><br />
<td style="padding: 10px 15px; color: #334155;">[[RAG]] (检索增强生成) • [[LLM]] (大语言模型) • [[Hallucination]] (幻觉)</td><br />
</tr><br />
<tr style="border-bottom: 1px solid #f1f5f9;"><br />
<td style="width: 85px; background-color: #f8fafc; color: #334155; font-weight: 600; padding: 10px 12px; text-align: right; vertical-align: middle;">数据基座</td><br />
<td style="padding: 10px 15px; color: #334155;">[[PubMed]] • [[Clinical Guidelines]] (临床指南) • [[FDA Labels]]</td><br />
</tr><br />
<tr><br />
<td style="width: 85px; background-color: #f8fafc; color: #334155; font-weight: 600; padding: 10px 12px; text-align: right; vertical-align: middle;">应用场景</td><br />
<td style="padding: 10px 15px; color: #334155;">[[Clinical Decision Support]] (临床决策) • [[Literature Review]] (文献综述)</td><br />
</tr><br />
</table><br />
</div><br />
<br />
</div></div>
77921020