与「Ontology」相关的搜索结果

EvoOntology：给数据智能体一个会成长的语义层 数据智能体（Data Agent）面对异构数据时，能“看见”的往往只有表名、列名和文件路径；指标如何定义、实体如何关联、业务上有何约束，这些语义都藏在数据之外。让智能体每次从零探索，代价高且易出语义错误；把人工维护的语义层全量注入提示词，又难以扩展、随数据漂移而失效。这就是论文中的 agent-data gap。 中国人民大学 ruc-datalab 提出的 EvoOntology（arXiv:2609.15779）给出了新解法：首个面向数据智能体的自进化本体层（Self-Evolving Ontology Layer）。核心洞察是把本体当作“可训练的智能体状态”：不更新模型权重，而让语义知识随真实使用持续积累、验证、版本化演进，每次变更可检查、可比较、可回滚。 本体以三层结构封装为 MCP 服务器，供智能体运行时主动查询： · Schema Layer 定义本体的表示边界（节点族、语义关系、引用模式）； · Content Layer 是类型化语义图，由 Terms / Mappings / Constraints / Evidence 四类节点构成，每个语义对象都有工作负载证据支撑，落地验证后才提交； · Tool Layer 仅暴露 browse_semantics 与 resolve_semantics 两个 MCP 工具加紧凑会话 manifest，智能体按需检索当前步骤所需语义，而非全量注入上下文。 本体遵循 Build → Use → Evolve → Evaluate → Publish 生命周期：构建器智能体从真实工作负载推导候选概念并对照原始数据验证；交互轨迹被持续记录；进化时诊断重复行为、归因到具体层级并生成局部补丁；候选版本只有在相同数据、相同智能体、相同解码设置与交互预算的配对评估中可复现地胜过父版本，才会发布。 在 BIRD、DDR-10K、InsightBench 三个基准、四个 LLM 骨干（GPT-5.5 / GPT-5.6-sol / Claude-Sonnet-5 / Claude-Opus-4.8）上，EvoOntology 一致优于无本体 ReAct 与静态初始本体。其中 DDR-Bench 上仅“自进化”环节就贡献了 +7.7（81.8 → 89.5），验证了持续演进相对静态语义层的增量价值。 论文： 代码：

Ethereum is both an abstract philosophical concept (the ontological Turing machine) and a practical research and development effort that aims to answer this question that pertains to the nature of Ethereum. How the nature of "Ethereum" as a protocol comes to be determined through the ACD governance mechanism, how cryptographic truth comes to be defined, and how convergent consensus is reached using majority representation of Ethereum's observers is what reduces the ontology of the world computer down to what we today call the Ethereum network. Here's how the world computer comes to be defined:

PITN has assembled a portfolio of exact-match multilingual digital assets spanning approximately 30 of the world's most widely spoken languages and writing systems. Structured using Unicode-based normalization, transliteration, and cross-lingual terminology mapping, the portfolio is designed to function as a semantic gateway for robotics and humanoid AI. By aligning concepts such as ROBOT, ROBOTS, HUMANOID, HUMANOIDS, and HUMANOID ROBOT across native scripts, these assets may support ontology-driven knowledge organization, AI discovery, semantic search, multilingual SEO, and globally consistent concept representation as search evolves from keywords to meaning. As humanoid technologies scale worldwide, standardized multilingual semantic infrastructure may become an increasingly important layer for global interoperability and discoverability. The more compelling question is not whether this semantic layer will matter, but at what point a major robotics company will have its "aha" moment and recognize that has already assembled a semantic gateway that could be leveraged today to accelerate global multilingual concept alignment and future humanoid ecosystems. #PITNAI# #Humanoid# #HumanoidRobots# #Robotics# #Ontology# #SemanticWeb# #KnowledgeGraph# #SemanticSearch# #AI# #NLP# #Multilingual# #Unicode# #SEO# #FutureOfAI# #AgenticAI# #DigitalInfrastructure#

Take a look at another page from Toy Worlds Atlas, a philosophic colouring-book! This puzzle explores Chalmers’s ontology of virtual reality. Using logic, the player must distinguish between colourants representing real and virtual objects. #ShowcaseThursday# #SoloDev# #IndieGame#

At DevCon 6, Palantir Software Engineer Ankit Shankar unveils the Agent Stack, a set of products to help you build production agents for the enterprise, all built on the Ontology. ▪️Orchestrator for durable execution ▪️Agent Engine and Agent SDK for building agents of any shape ▪️Agent Manager for actionable telemetry and observability ▪️AIP Evolve for continuous, end-to-end optimization

Palantir CEO Alex Karp on why the model is not the moat. The application layer is: "To make them valuable in an enterprise, like a battlefield, regulated, or manufacturing context, you have to have what's called an application layer." "We have this thing called ontology that now everyone's copying. It takes a large language model and makes it safe and useful and precise." "Safe because it doesn't touch your underlying data." "Safe because it prevents the large language model from caching your data and replicating your business." The tell for any enterprise AI pitch: ask what sits between the model and your data. If the answer is nothing, you are not buying leverage. You are renting a fast way to hand your business to whoever owns the weights. Alex Karp, CEO of Palantir, on CNBC.

Palantir CEO：OpenAI 和 Anthropic 在抢劫每一家财富 500 强公司 Palantir CEO Alex Karp 在 CNBC 采访中表示，企业级 AI 的真正价值不只来自模型本身，而是模型、应用层和算力三者结合。他认为，企业和政府客户真正关心的是能否控制自己的数据、模型权重、算力和业务 Alpha，而不是在按 token 付费的同时，把核心业务能力交给第三方前沿 AI 实验室。 Karp 称，许多企业私下已经对这种模式感到不满，担心自己支付了高昂费用，却没有获得真实价值，反而让自身数据和竞争优势被转移。他同时表示，AI 并非没有价值，算力、Ontology 与模型结合正在改变历史进程，但行业不需要对已有技术进行过度推销。 内容不构成任何投资建议，请严格遵循当地法律法规。 《白线 WhiteLine》由吴说团队出品，从 Crypto 走向更广阔的资本市场，关注 AI 时代下的趋势变化与交易机会。

Today I released the public preprint of SPOO: The Primal Geometry of Coherence. It argues that what we call “AI” is not intelligence in the ordinary sense, but synthetic coherence: structured output without governed evidentiary grounding. This is the framework I’ve spent 13 months building for the public, academia, and industry. This is not mere philosophy. Public preprint below with free PDF link to the SPOO ontology. This is just the very beginning of where this will go.

Many people think any given ML project is 99% training. In reality, it’s 50% evaluation, 40% data cleaning, 8% integration, and 2% training. The first two set the noise floor for learning. No ML magic matters; the model cannot lower the noise floor, as that’s the optimal bound of Shannon encoding of your data. Thus, not a single day goes by without me thinking about ontology. Even the old labels have to be constantly reviewed.

作为苦逼ToB 人，正在研究 Palantir的本体论（Ontology）。 还不太搞的懂，到底是知识图谱？领域建模？还是什么。 倒是解决的问题很直观，用模型解决生产中的问题（比如设备故障诊断等）。