From 8ab9cb41de7139d5c707def7e56ab02ae495f3d8 Mon Sep 17 00:00:00 2001 From: WaylandYang Date: Thu, 13 Aug 2026 02:13:50 +0800 Subject: [PATCH 1/4] docs: remove readme pre-release warning --- README.md | 3 --- README.zh-CN.md | 3 --- 2 files changed, 6 deletions(-) diff --git a/README.md b/README.md index cddb8c4..65f09d9 100644 --- a/README.md +++ b/README.md @@ -18,9 +18,6 @@ Build, review, version, publish, and serve TBox, SKOS terminology, and ABox data -> [!IMPORTANT] -> OntoPilot is under active development and has not reached 1.0. Back up PostgreSQL, the OntoPilot data volume, and your token-encryption key before upgrades. Review release notes and validate migrations on a copy of production data. -
Contents diff --git a/README.zh-CN.md b/README.zh-CN.md index d789f5c..ddaca9a 100644 --- a/README.zh-CN.md +++ b/README.zh-CN.md @@ -18,9 +18,6 @@ -> [!IMPORTANT] -> OntoPilot 正在快速开发，尚未达到 1.0。升级前请备份 PostgreSQL、OntoPilot 数据卷和 Token 加密密钥，并先使用生产数据副本验证迁移。 -
目录 From b925b745f504928ac13c6b22178fcc1113314b45 Mon Sep 17 00:00:00 2001 From: WaylandYang Date: Thu, 13 Aug 2026 02:15:24 +0800 Subject: [PATCH 2/4] docs: streamline benchmark footnote --- README.md | 8 +++----- README.zh-CN.md | 6 ++---- 2 files changed, 5 insertions(+), 9 deletions(-) diff --git a/README.md b/README.md index 65f09d9..fe3d959 100644 --- a/README.md +++ b/README.md @@ -57,15 +57,13 @@ It goes beyond asking an LLM to “generate an ontology.” OntoPilot puts domai | Protocol F1 | Wine
Food & Beverage | GeoNames
Geography | OWL-Time
Units & Measurements | | --- | ---: | ---: | ---: | -| OntoLearner reference · Qwen3-8B | 18.60%¹ | 19.70%¹ | 14.08%² | +| OntoLearner reference · Qwen3-8B | 18.60% | 19.70% | 14.08% | | **OntoPilot evaluation · Qwen3-8B** | **28.95%** | **27.03%** | **16.67%** | | **Improvement** | **+10.35 pp / +55.6%** | **+7.33 pp / +37.2%** | **+2.58 pp / +18.3%** | | Result | **New SOTA** | Same-model lead | Prompt gain | -¹ OntoLearner paper result. ² Controlled OntoLearner-prompt baseline because the paper does not -report OWL-Time individually. Wine and OWL-Time use OntoPilot's frozen prompt; GeoNames currently -uses the unchanged OntoLearner prompt in our adapter. See the -[benchmark methodology and full results](docs/benchmarks/ontolearner-multidomain.md). +See the [benchmark methodology and full results](docs/benchmarks/ontolearner-multidomain.md) +for evaluation scope, baselines, prompt profiles, and reproducibility details. ## Capabilities diff --git a/README.zh-CN.md b/README.zh-CN.md index ddaca9a..bdb6cea 100644 --- a/README.zh-CN.md +++ b/README.zh-CN.md @@ -57,14 +57,12 @@ OntoPilot 是面向企业与业务团队的本体生产工作台：把散落在 | 协议 F1 | Wine
食品与饮料 | GeoNames
地理 | OWL-Time
单位与度量 | | --- | ---: | ---: | ---: | -| OntoLearner 参照 · Qwen3-8B | 18.60%¹ | 19.70%¹ | 14.08%² | +| OntoLearner 参照 · Qwen3-8B | 18.60% | 19.70% | 14.08% | | **OntoPilot 评测 · Qwen3-8B** | **28.95%** | **27.03%** | **16.67%** | | **提升** | **+10.35 个百分点 / +55.6%** | **+7.33 个百分点 / +37.2%** | **+2.58 个百分点 / +18.3%** | | 结论 | **新 SOTA** | 同模型领先 | 提示词提升 | -¹ OntoLearner 论文成绩。² 论文未单列 OWL-Time，因此使用受控的 OntoLearner 提示词基线。 -Wine 和 OWL-Time 使用 OntoPilot 冻结提示词；GeoNames 目前仍使用未修改的 OntoLearner -提示词运行于我们的适配器。完整方法、六个数据集、消融和复现见 +评测范围、基线口径、提示词配置与复现细节见 [Benchmark 方法与完整报告](docs/benchmarks/ontolearner-multidomain.md)。 ## 核心能力 From 68df3fcf4eaf2ad4b2d766e5fc973173af4d56f9 Mon Sep 17 00:00:00 2001 From: WaylandYang Date: Thu, 13 Aug 2026 02:17:26 +0800 Subject: [PATCH 3/4] docs: highlight review-driven evolution --- README.md | 3 +++ README.zh-CN.md | 3 +++ 2 files changed, 6 insertions(+) diff --git a/README.md b/README.md index fe3d959..b5cd02e 100644 --- a/README.md +++ b/README.md @@ -4,6 +4,8 @@ **Human-governed ontology engineering from source documents.** +`Evolves with every review · Learns from every decision` + Build, review, version, publish, and serve TBox, SKOS terminology, and ABox data from one self-hosted workspace. [简体中文](README.zh-CN.md) · [Documentation](#documentation) · [Architecture](docs/architecture.md) · [Changelog](CHANGELOG.md) · [Roadmap](ROADMAP.md) · [Contributing](CONTRIBUTING.md) · [Code of Conduct](CODE_OF_CONDUCT.md) · [Security](SECURITY.md) @@ -48,6 +50,7 @@ It goes beyond asking an LLM to “generate an ontology.” OntoPilot puts domai - **From documents to computable domain knowledge.** Convert scattered language into a connected TBox, SKOS terminology, and ABox while retaining the source behind every statement. - **Human–AI co-creation with governance built in.** Models propose; experts review, correct, and approve through focused queues instead of rebuilding machine output by hand. +- **Every review makes the agent better.** Human decisions and review rationale become reusable memory, helping the agent resolve recurring ambiguity and improve the accuracy of future extraction and validation. - **From a promising draft to a production asset.** Semantic Diff, immutable releases, rollback, REST APIs, and MCP carry approved knowledge into business systems and agent workflows. - **Traceable by design, not by afterthought.** Every decision can be traced to its document chunk, model, prompt snapshot, actor, and review history. diff --git a/README.zh-CN.md b/README.zh-CN.md index bdb6cea..824e3bf 100644 --- a/README.zh-CN.md +++ b/README.zh-CN.md @@ -4,6 +4,8 @@ **从源文档构建由人治理、可追溯、可发布的本体。** +`在每次审核中进化 · 从每个决策中学习` + 在一个自托管工作台中完成 TBox、SKOS 术语、ABox 的构建、审阅、版本化、发布与服务。 [English](README.md) · [文档](#文档与接口) · [架构](docs/architecture.md) · [更新日志](CHANGELOG.md) · [路线图](ROADMAP.md) · [参与贡献](CONTRIBUTING.md) · [行为准则](CODE_OF_CONDUCT.md) · [安全策略](SECURITY.md) @@ -48,6 +50,7 @@ OntoPilot 是面向企业与业务团队的本体生产工作台：把散落在 - **从业务文档到可计算的领域知识。** 将分散的自然语言转化为相互关联的 TBox、SKOS 术语与 ABox，同时保留每条语句的原始依据。 - **让人机协作真正可治理。** 模型规模化提出候选，专家在聚焦的审核队列中修正与裁决，不必从头返工，也不必盲信生成结果。 +- **每一次审核，都让智能体更懂你的业务。** 人工决策与审核意见会沉淀为可复用的学习记忆，帮助智能体消除反复出现的歧义，并持续提高后续抽取与校验准确率。 - **从“看起来可用”走到生产可用。** 通过语义 Diff、不可变发布、回滚、REST API 与 MCP，把审核后的知识稳定交付给业务系统和 Agent。 - **可追溯不是补丁，而是底座。** 每项决策都能回到文档 chunk、模型、提示词快照、操作者与完整审核历史。 From 64980de518dd7b1e4b59ed8882bf60e6477d9eab Mon Sep 17 00:00:00 2001 From: WaylandYang Date: Thu, 13 Aug 2026 02:22:36 +0800 Subject: [PATCH 4/4] docs: illustrate review learning loop --- README.md | 2 +- README.zh-CN.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index b5cd02e..8e993c2 100644 --- a/README.md +++ b/README.md @@ -50,7 +50,7 @@ It goes beyond asking an LLM to “generate an ontology.” OntoPilot puts domai - **From documents to computable domain knowledge.** Convert scattered language into a connected TBox, SKOS terminology, and ABox while retaining the source behind every statement. - **Human–AI co-creation with governance built in.** Models propose; experts review, correct, and approve through focused queues instead of rebuilding machine output by hand. -- **Every review makes the agent better.** Human decisions and review rationale become reusable memory, helping the agent resolve recurring ambiguity and improve the accuracy of future extraction and validation. +- **Every review makes the agent better.** Suppose one document says “Ocean Explorer One” and another says “OE-1.” Once an expert confirms they are the same vehicle—and records why—OntoPilot retains that decision as reusable resolution memory. The next occurrence can map to the right entity instead of creating a duplicate; new or conflicting variants still return to human review. - **From a promising draft to a production asset.** Semantic Diff, immutable releases, rollback, REST APIs, and MCP carry approved knowledge into business systems and agent workflows. - **Traceable by design, not by afterthought.** Every decision can be traced to its document chunk, model, prompt snapshot, actor, and review history. diff --git a/README.zh-CN.md b/README.zh-CN.md index 824e3bf..8b6d29e 100644 --- a/README.zh-CN.md +++ b/README.zh-CN.md @@ -50,7 +50,7 @@ OntoPilot 是面向企业与业务团队的本体生产工作台：把散落在 - **从业务文档到可计算的领域知识。** 将分散的自然语言转化为相互关联的 TBox、SKOS 术语与 ABox，同时保留每条语句的原始依据。 - **让人机协作真正可治理。** 模型规模化提出候选，专家在聚焦的审核队列中修正与裁决，不必从头返工，也不必盲信生成结果。 -- **每一次审核，都让智能体更懂你的业务。** 人工决策与审核意见会沉淀为可复用的学习记忆，帮助智能体消除反复出现的歧义，并持续提高后续抽取与校验准确率。 +- **每一次审核，都让智能体更懂你的业务。** 假设一份文档写“海洋探测器一号”，另一份写“海探1”。专家确认它们是同一设备并留下判断理由后，OntoPilot 会把这项决策沉淀为可复用的实体消歧记忆；下次再遇到“海探1”时，它会映射到正确实体，而不是重复创建。遇到新的叫法或冲突证据时，仍会回到人工审核。 - **从“看起来可用”走到生产可用。** 通过语义 Diff、不可变发布、回滚、REST API 与 MCP，把审核后的知识稳定交付给业务系统和 Agent。 - **可追溯不是补丁，而是底座。** 每项决策都能回到文档 chunk、模型、提示词快照、操作者与完整审核历史。