<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Fluree PBC - Medium]]></title>
        <description><![CDATA[Fluree, PBC is a Public Benefit Corporation that offers a decentralized platform for enterprises to create better, more efficient software applications and a blockchain database that securely powers them to scale. - Medium]]></description>
        <link>https://medium.com/fluree?source=rss----5f522fe3f91d---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>Fluree PBC - Medium</title>
            <link>https://medium.com/fluree?source=rss----5f522fe3f91d---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 16:58:00 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/fluree" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How Linked Data Can Help Improve LLM Accuracy and Effectiveness]]></title>
            <link>https://medium.com/fluree/how-linked-data-can-help-improve-llm-accuracy-and-effectiveness-e26f00208815?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/e26f00208815</guid>
            <category><![CDATA[knowledge-graph]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[linked-data]]></category>
            <category><![CDATA[semantics]]></category>
            <category><![CDATA[genai]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Wed, 27 Aug 2025 19:55:29 GMT</pubDate>
            <atom:updated>2025-08-27T19:55:29.071Z</atom:updated>
            <content:encoded><![CDATA[<p>Providing your LLM with linked data is like giving it not just a direction, but a detailed map and compass to follow precise, step-by-step instructions.</p><p>Large Language Models continue to evolve, and organizations are trying to move from the low-hanging fruit use cases (blog post generation, customer service bots) to more complex, specialized use cases (knowledge co-pilots, decisioning agents).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ASlCSzCsE5OZyWyG.png" /></figure><p>To support these more sophisticated GenAI use cases, we need access to better enterprise data. There are a few patterns that have emerged to support specialized information retrieval, but all of them must tackle a deep-rooted problem that has existed far before the age of Chat-GPT: data silos.</p><p>Linked Data has been around for decades — largely to help the internet of URLs connect web pages together under global concepts for better search engine experiences (see schema.org).</p><p>By connecting disparate data sources through standardized frameworks, Linked Data allows information to be more easily accessed, shared, and understood. Now, as organizations push the boundaries of what LLMs can do, Linked Data is taking center stage again, offering a way to unify fragmented enterprise knowledge and provide the high-quality, interconnected data GenAI needs. In this article, we’ll explore how Linked Data can help bridge the gap between siloed data and the advanced, decision-making capabilities organizations are striving to achieve with their AI initiatives.</p><h3>Machine Readability: The Foundation of LLM Accuracy</h3><p>At the heart of LLMs’ ability to generate coherent and accurate information is machine readability — the ease with which AI models can process and interpret data. Linked Data standards are designed to ensure that data is structured in a way that machines can understand, making them a perfect fit for improving LLM performance.</p><p>Linked Data formats like RDF use triples — subject, predicate, object — to describe relationships between pieces of information. This structured approach enables LLMs to “read” the data more effectively, reducing ambiguity and improving the accuracy of the model’s output. When an LLM pulls information from a knowledge graph based on Linked Data, it can leverage these relationships to generate more precise answers, especially in complex or highly specialized queries.</p><p>For example, in a pharmaceutical context, Linked Data can help an LLM accurately connect information between drugs, clinical trials, and diseases, ensuring that the AI produces scientifically sound and context-aware outputs. Without the structured relationships provided by Linked Data standards, LLMs would struggle to make these crucial connections, leading to less accurate or incomplete responses.</p><h3>Enhancing Data Connectivity for LLMs with Linked Data</h3><p>Another major benefit of Linked Data standards is how they enable better data connectivity. In traditional data structures, information is often siloed, making it difficult for AI models to retrieve the most relevant data. However, Linked Data standards allow different datasets to be connected across platforms, creating a unified ecosystem of information.</p><p>When LLMs are trained on this interconnected data, their ability to retrieve the right information improves significantly. Linked Data standards like SPARQL enable precise querying, meaning LLMs can pull from specific data points within a knowledge graph that are directly relevant to the user’s question. This increased precision enhances both the retrieval and generation capabilities of the LLM, resulting in more accurate and effective outputs.</p><p>For instance, when querying financial data, an LLM can use Linked Data standards to link transactional data with market analysis, providing predictions based on a well-connected, comprehensive dataset. This capacity to connect and understand data relationships helps LLMs generate more robust insights, improving their overall effectiveness.</p><h3>Ontologies and Their Role in LLM Effectiveness</h3><p>When integrating multiple data sets together, we need to apply a global blueprint of terms and concepts germain to the business across the data environment. This allows organizations to apply standards to data and metadata, define global relationships, and ultimately reduce errors, inconsistencies, and duplication</p><p>Ontologies — formal representations of a set of concepts and their relationships within a domain — are another key component of Linked Data standards that boost LLM performance. By defining the relationships between concepts in a structured way, ontologies help LLMs better understand the context of the data they’re processing.</p><p>Providing your LLM with data modeled against a global ontology is like giving it not just a direction, but a detailed map and compass to follow precise, step-by-step instructions.</p><p>Read more about ontologies <a href="https://flur.ee/fluree-blog/semantic-interoperability-exchanging-data-with-meaning/">here</a>.</p><p>Because LLMs thrive on understanding context, the use of ontologies dramatically increases their ability to provide accurate responses. Linked Data standards like OWL, which support the creation of rich ontologies, are essential for making sure that LLMs aren’t just regurgitating information but rather generating insightful, context-aware responses. This structure allows LLMs to better interpret data and generate outputs that reflect the nuanced relationships within the field.</p><h3>Linked Data Standards for Enhanced LLM Retrieval-Augmented Generation (RAG)</h3><p>In Retrieval-Augmented Generation (RAG), Linked Data standards can play a transformative role in improving accuracy. RAG models combine the retrieval of relevant data with generative capabilities to provide users with well-informed and contextually relevant responses.</p><p>The precision of data retrieval in RAG depends heavily on the quality of the data connections and structure. Linked Data standards ensure that these connections are well-defined and machine-readable, allowing RAG models to pull more accurate information from various data sources. By improving the data that RAG models have access to, Linked Data enhances both the retrieval and generation aspects of the process, leading to higher-quality results.</p><p>For example, in regulatory compliance for pharmaceuticals, RAG models using Linked Data standards can retrieve data on clinical trials, regulatory guidelines, and drug safety records. With this precise, interconnected data, the LLM can then generate compliance reports that are not only accurate but also comprehensive, reducing the risk of error in critical decision-making processes.</p><p>Read more about RAG <a href="https://flur.ee/fluree-blog/resource-download/fluree-report-decentralized-graphrag-for-genai-accuracy/">here</a>.</p><p><strong>Thanks for Reading!</strong></p><p>Fluree offers a full suite of semantic data management products to build, maintain, leverage and share enterprise knowledge. Learn more about our comprehensive knowledge graph solutions or book a call to speak with an expert here. Or, dive into more information about our product suite below:</p><ul><li>Fluree Core — Semantic Graph Database with Built-In Security and Digital Trust</li><li>Fluree ITM — Intelligent Taxonomy and Ontology Manager</li><li>Fluree Sense– Semantic Pipeline to Transform Legacy Structured Data to Knowledge Graphs</li><li>Fluree CAM– Intelligent Content Auto-Tagging for Digital Media Assets</li></ul><h3>Master Enterprise AI with Semantic GraphRAG</h3><p>Unlock the full potential of your enterprise data with Fluree’s comprehensive guide to semantic GraphRAG. Learn how to build intelligent, context-aware AI systems that deliver accurate, explainable results.</p><p><a href="https://flur.ee/fluree-blog/resource-download/everything-you-need-to-know-about-semantic-graphrag/">Download Enterprise AI Guide</a></p><h3>Unlock the Power of Semantic AI</h3><p>Join our expert-led webinar to discover how semantic AI transforms enterprise data management. See real-world examples and learn implementation strategies that drive business value.</p><p><a href="https://app.livestorm.co/fluree/unlocking-the-power-of-semantic-ai-for-enterprise-data?utm_source=Livestorm+company+page">Watch Semantic AI Webinar</a></p><h3>Experience Fluree in Action</h3><p>Ready to see how Fluree’s semantic graph database can transform your data architecture? Get hands-on experience with our platform and discover the power of connected, intelligent data.</p><p><a href="https://flur.ee/get-started/">Try Fluree in Action</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e26f00208815" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/how-linked-data-can-help-improve-llm-accuracy-and-effectiveness-e26f00208815">How Linked Data Can Help Improve LLM Accuracy and Effectiveness</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Taxonomies Versus Ontologies: A Short Guide]]></title>
            <link>https://medium.com/fluree/taxonomies-versus-ontologies-a-short-guide-71ae41fe0d54?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/71ae41fe0d54</guid>
            <category><![CDATA[ontology]]></category>
            <category><![CDATA[controlled-vocabulary]]></category>
            <category><![CDATA[taxonomy]]></category>
            <category><![CDATA[semanticweb]]></category>
            <category><![CDATA[metadata]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Wed, 27 Aug 2025 19:34:50 GMT</pubDate>
            <atom:updated>2025-08-27T19:34:49.951Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5McK3FDCd0huN2XfXgdGXA.png" /></figure><h3>Discover the key differences between taxonomy and ontology and learn how each can optimize your business information management.</h3><p>In the world of information organization, taxonomies and ontologies play crucial roles but are often misunderstood or used interchangeably. Both serve to structure data and knowledge, but they do so in fundamentally different ways. Here’s a clear breakdown of their distinctions:</p><h3>What is a Taxonomy?</h3><p>A taxonomy is a hierarchical classification system used to categorize and organize information into groups and subgroups. Think of it as a structured way of grouping entities based on shared characteristics. It’s often represented as a tree-like structure where each node is a category or a subcategory.</p><p><strong>Key Characteristics of Taxonomies:</strong></p><ul><li><strong>Hierarchical Structure:</strong> Taxonomies are typically organized in a tree structure with parent-child relationships. For example, a biological taxonomy might classify organisms into Kingdom, Phylum, Class, Order, Family, Genus, and Species.</li><li><strong>Simple Relationships:</strong> Taxonomies focus on hierarchical relationships (e.g., broader and narrower terms) and do not usually capture complex interrelationships between categories.</li><li><strong>Fixed Vocabulary:</strong> Taxonomies often have a predefined set of terms or categories that are less flexible in accommodating new or unforeseen concepts.</li></ul><p><strong>Example of Taxonomy:</strong></p><p>In a library catalog, books might be categorized as follows:</p><h3><strong>Fiction</strong></h3><blockquote>Mystery</blockquote><blockquote>Romance</blockquote><blockquote>Science Fiction</blockquote><h3>Non-Fiction</h3><blockquote>Biography</blockquote><blockquote>Self-Help</blockquote><blockquote>History</blockquote><h3>What is an Ontology?</h3><p>An ontology, on the other hand, is a more complex and flexible framework used to model the relationships between entities and their properties. It provides a rich, formal representation of knowledge within a domain, capturing not only the hierarchy but also the various relationships between concepts.</p><p><strong>Key Characteristics of Ontologies:</strong></p><ul><li>Rich Relationships: Ontologies describe multiple types of relationships between concepts (e.g., parent-child, part-whole, causal). They also include properties and constraints that define how entities interact.</li><li>Formal Representation: Ontologies use formal languages, such as OWL (Web Ontology Language), to provide precise definitions and infer logical relationships.</li><li>Dynamic Vocabulary: Unlike taxonomies, ontologies can be more adaptable, allowing for the inclusion of new concepts and relationships as knowledge evolves.</li></ul><p><strong>Example of Ontology:</strong></p><p>In a medical ontology, the concept of “Diabetes” might be connected to various other concepts through multiple relationships:</p><h3><strong>Disease:</strong> Diabetes</h3><blockquote><strong>Has Symptom:</strong> Increased Thirst</blockquote><blockquote><strong>Has Treatment:</strong> Insulin</blockquote><blockquote><strong>Causes:</strong> High Blood Sugar</blockquote><blockquote><strong>Subtype:</strong> Type 1 Diabetes, Type 2 Diabetes</blockquote><h3>Key Differences</h3><p><strong>Complexity:</strong> Taxonomies offer a simple hierarchical structure, while ontologies provide a detailed and nuanced representation of entities and their interrelations.</p><p><strong>Flexibility:</strong> Ontologies are generally more flexible, accommodating changes and additions to the domain more easily than taxonomies.</p><p><strong>Purpose:</strong> Taxonomies are primarily used for classification and organization, whereas ontologies are used for more sophisticated knowledge modeling and reasoning.</p><h3>When to Use Each</h3><p><strong>Taxonomies</strong> are ideal for straightforward categorization tasks where the primary goal is to group similar items and navigate through a hierarchical structure. For example, categorizing books in a library or organizing products in an online store.</p><p><strong>Ontologies</strong> are suited for scenarios where understanding and reasoning about complex relationships and interactions is critical. They are often used in fields like artificial intelligence, semantic web, and data integration where detailed knowledge representation is essential.</p><p>In summary, while both taxonomies and ontologies help structure and organize information, they do so in different ways and for different purposes. Understanding these differences can help in choosing the right tool for the task at hand, ensuring that information is managed effectively and efficiently.</p><p>Want to extract, build and manage your enterprise ontology or taxonomy? Check out <a href="https://flur.ee/taxonomy-manager/">Fluree ITM</a>, our intelligent taxonomy manager.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=71ae41fe0d54" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/taxonomies-versus-ontologies-a-short-guide-71ae41fe0d54">Taxonomies Versus Ontologies: A Short Guide</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building Corporate Memory for Enterprise LLMs]]></title>
            <link>https://medium.com/fluree/building-corporate-memory-for-enterprise-llms-bbd4a3af47b8?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/bbd4a3af47b8</guid>
            <category><![CDATA[generative-ai-tools]]></category>
            <category><![CDATA[enterprise-technology]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[knowledge-graph]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Wed, 27 Aug 2025 19:31:00 GMT</pubDate>
            <atom:updated>2025-08-27T19:31:33.357Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ad35iO_EjAzm4x0p125ddA.png" /></figure><p>Giving LLMs and GenAI Applications access to context and knowledge accumulated by your organization</p><p>Organizations today are likely hitting their <a href="https://www.techrepublic.com/article/generative-ai-dominates-2023-gartner-hype-cycle/">peak of inflated expectations</a> after a few rounds of LLM Proof of Concepts. While new and powerful models hit the market to provide greater intelligence at a lower cost, organizations are still hitting the same barriers when trying to scale enterprise LLMs to production:</p><ol><li>Accuracy</li><li>Relevance</li><li>Context</li></ol><p>These barriers — among others — can be generally attributed to how enterprise data and content are organized, managed, and accessed. Mckinsey <a href="https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/a-data-leaders-technical-guide-to-scaling-gen-ai">recently stated</a> that “Managing data remains a top barrier to value creation from GenAI.” Let’s get more specific: GenAI that isn’t grounded in contextual, trusted, and real-time data is never leaving the proof-of-concept bucket.</p><p>In order for organizations to truly deploy LLM applications that are context-aware and hallucination-proof, they must build and maintain their corporate memory.</p><h3>What is Corporate Memory?</h3><p>Corporate memory refers to the collective experience, information, and knowledge accumulated by an organization over time.</p><p>Corporate memory is more than just your operational or analytical data; it’s the essence of your organization’s history and positioning, key processes, business strategies, metrics, and lessons learned that have shaped your journey. In other words, it’s enterprise context.</p><p>Corporate Memory lives in unstructured content like PDFs, presentations, FAQs and email documents, as well as in structured data stored in databases, SaaS applications, and data lakes. More importantly, it resides in the minds of your subject matter experts (SMEs). Wherever your corporate memory lives, the ability to bring it together into a singular “corporate brain” is the key to building context-aware information retrieval for Generative AI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*SjEaIKR6O54qEcfO.jpg" /></figure><p>But oftentimes, organizations have a lack of good corporate memory — they have what we might call corporate amnesia.</p><h3>What is Corporate Amnesia?</h3><p>Corporate Amnesia occurs when an organization has no easy, repeatable access to the knowledge it generates or captures. The organization in question might have some semblance of corporate memory, but it is spread out in artifacts that are held in isolated information systems.</p><p>Symptoms of corporate amnesia might look like the following:</p><ul><li><strong>Data is difficult to find, access, or interpret</strong>: Information is scattered across systems, making it hard for employees to locate and use.</li><li><strong>Information is duplicated, outdated, or incorrect</strong>: Multiple versions of the same data lead to confusion and errors in decision-making.</li><li><strong>High costs for data preparation and integration</strong>: Manual data wrangling consumes resources that could be better spent on innovation.</li><li><strong>Delayed results from business inquiries</strong>: Business questions require time-consuming manual intervention to get answers.</li><li><strong>Data security and governance challenges</strong>: Fragmented policies prevent access to critical business information.</li><li><strong>Inconsistent knowledge sharing</strong>: Valuable insights from one department are not shared or utilized by others, leading to missed opportunities.</li><li><strong>Lack of a common data language</strong>: Different teams use varying terminologies and formats, creating barriers to effective collaboration.</li><li><strong>Over-reliance on key individuals</strong>: When knowledge is concentrated in a few employees, the organization is vulnerable if those individuals are not readily-available.</li></ul><p>Corporate Amnesia might look like a banker trying to piece together a full customer profile from duplicate CRM records, an insurance salesman sifting through thousands of PDFs to find the right policy group for a client, a customer service representative frantically searching through internal FAQ documents to find the answer to a rare question, or the risk department at a financial institution struggling to identify patterns in historical data because key insights are buried in outdated spreadsheets and isolated databases.</p><p>If humans are <a href="https://www.articlecube.com/research-shows-searching-information-work-wastes-time-and-money">struggling with getting access</a> to relevant, real-time knowledge from across their organization, LLMs can’t be doing too much better.</p><h3>LLMs Don’t Have Access to Your Corporate Memory — Yet.</h3><p>A lack of available, accessible, and well-documented institutional knowledge might be the result of the “big data” boom in the early 2000s: Lots of data, stored in different formats, living across isolated, separate systems. These systems and applications were not built for re-usability beyond basic query/retrieval.</p><p>Today, organizations are responding to these issues with a few bandaids: ETL, data lakes, and manual data cleansing in order to get information from systems into a <a href="https://towardsdatascience.com/retrieval-augmented-generation-rag-from-theory-to-langchain-implementation-4e9bd5f6a4f2">RAG pipeline</a>. But these are neither sustainable nor scalable methods if LLMs need continuous, grounded access to information in real-time.</p><h3>Why Semantics is the Key First Step</h3><p>A fundamental challenge in bringing any kind of disparate information together will be in aligning concepts. What a sales report located in a CRM might call “customer,” an invoice located in a finance ERP system might call “client.” If we were to simply dump information together in a massive data lake, we wouldn’t really have integrated data — we’d just have a mess to clean up before it can be of any use.</p><p>Last year, we wrote a fairly extensive piece on <a href="https://flur.ee/fluree-blog/semantic-interoperability-exchanging-data-with-meaning/">Semantic Interoperability</a> — the ability for disparate systems to “speak the same language.” Implementing a semantic layer can serve as a universal framework to define the concepts and relationships across your data assets. This is a key step in moving towards building corporate memory for LLMs — understanding and representing the semantic meaning behind data allows information to be linked together and retrieved in a context-aware manner.</p><p>Semantics gives us a few key advantages in bringing information together:</p><ul><li><strong>Established relationships across data assets — </strong>Semantics enables the explicit definition and management of relationships between data entities through ontologies. By inferring and maintaining these relationships, it allows for more accurate and meaningful connections across disparate data assets. This not only enhances data integration but also ensures that queries and information retrieval are contextually accurate, leveraging the full breadth of interconnected knowledge.</li><li><strong>Increased Accuracy and Precision</strong> — In working with customers as well as our own internal experiments, we’ve found that RAG against a semantic knowledge graph <a href="https://flur.ee/fluree-blog/resource-download/fluree-report-decentralized-graphrag-for-genai-accuracy/">increased baseline response accuracy by up to 4x</a> over a traditional relational database. GraphRAG has gained momentum as the best step towards hallucination-free information retrieval.</li><li><strong>Contextual Understanding Beyond Exact Terms </strong>Prompts — or natural language queries written by the user — don’t have to use the perfect, explicit terms located in the various schemas of the datasets containing key information. Regardless of phrasing used, we can tell the machine what we are looking for with semantic meaning as added context to both the prompt and the available data. This is far superior to <a href="https://technologyadvice.com/blog/information-technology/vector-similarity-search/">vector similarity searches</a> because instead of a probabilistic outcome, we achieve precise, contextually relevant results that align with the true intent of the query.</li></ul><p>A <a href="https://app.livestorm.co/fluree/a-practical-approach-to-implementing-semantic-data-layers">semantic layer in your data architecture</a> serves as a bridge between disparate data sources and LLMs, ensuring that your models have access to the right information at the right time. By applying semantic technologies such as knowledge graphs, ontologies, and metadata management, organizations can unify their data into a coherent, machine-readable format. This not only enhances the accuracy and relevance of the information retrieved but also reduces the chances of AI hallucinations, where the model generates responses based on incorrect or incomplete data.</p><h3>Injecting Corporate Memory into LLMs</h3><p>We’ve found that the best possible data management platform for creating and managing “corporate memory” involves the following components:</p><ul><li><strong>Data</strong> — We must surface information held in structured and unstructured systems</li><li><strong>Ontologies</strong> — We must classify and interconnect information against a universal set of concepts</li><li><strong>Security</strong> — We must use the ontology to define key policy frameworks. We’ll discuss security in our last section</li></ul><p>Combining the above elements into a semantic knowledge graph layer allows us to dynamically tap into the collective knowledge of the organization.</p><p>This approach ensures that all relevant data, regardless of its original format or location, is accessible and usable by LLMs. Because we’ve made use of a universal ontology, we can classify every piece of data/content at the source of ingestion into the knowledge graph.</p><p>The semantic knowledge graph acts as a living repository of corporate memory, continuously evolving as new data is ingested and existing information is refined. By integrating this semantic layer into your data architecture, you empower your AI models to make informed decisions, improve operational efficiency, and maintain consistency across all levels of your organization. This approach not only maximizes the utility of your data but also ensures that your AI-driven applications remain reliable and contextually aware.</p><p>To truly integrate corporate memory into LLMs, organizations must go beyond simply moving their stored data into a single lake or warehouse. They need to create a dynamic, living corporate memory that is continuously updated and enriched with new knowledge.</p><h3>The Future of Enterprise LLMs with Corporate Memory</h3><p>The majority of LLM applications that are deployed today are useful tools that help automate a simple process — such as generating blog posts for marketing, managing call center chat interactions, summarizing documents, drafting emails, and even providing basic customer support.</p><p>But we can start to see the true potential of Generative AI just around the corner: complex knowledge retrieval that can enhance critical components of our business processes; use cases that involve both a deep understanding of the business and true real-time access to available knowledge held across multiple systems.</p><p>The ability to manage and utilize corporate memory effectively will become a critical differentiator for organizations to achieve these advanced GenAI applications.</p><p><strong>Thanks for Reading!</strong></p><p>Fluree offers a full suite of semantic data management products to build, maintain, leverage and share enterprise knowledge. Learn more about our comprehensive <a href="https://flur.ee/initiatives-landing/enterprise-knowledge-graph/">knowledge graph solutions</a> or book a call to speak with an expert <a href="https://flur.ee/initiatives-landing/enterprise-knowledge-graph/">here</a>. Or, dive into more information about our product suite below:</p><ul><li><a href="http://flur.ee/core"><strong>Fluree Core</strong></a> — Semantic Graph Database with Built-In Security and Digital Trust</li><li><a href="https://flur.ee/taxonomy-manager/"><strong>Fluree ITM</strong></a> — Intelligent Taxonomy and Ontology Manager</li><li><a href="https://flur.ee/fluree-products-ai-ml-data-pipeline/"><strong>Fluree Sense</strong></a>– Semantic Pipeline to Transform Legacy Structured Data to Knowledge Graphs</li><li><a href="https://flur.ee/sense-unstructured/"><strong>Fluree CAM</strong></a>– Intelligent Content Auto-Tagging for Digital Media Assets</li></ul><h3>Master Enterprise AI with Semantic GraphRAG</h3><p>Unlock the full potential of your enterprise data with Fluree’s comprehensive guide to semantic GraphRAG. Learn how to build intelligent, context-aware AI systems that deliver accurate, explainable results.</p><p><a href="https://flur.ee/fluree-blog/resource-download/everything-you-need-to-know-about-semantic-graphrag/">Download Enterprise AI Guide</a></p><h3>Unlock the Power of Semantic AI</h3><p>Join our expert-led webinar to discover how semantic AI transforms enterprise data management. See real-world examples and learn implementation strategies that drive business value.</p><p><a href="https://app.livestorm.co/fluree/unlocking-the-power-of-semantic-ai-for-enterprise-data?utm_source=Livestorm+company+page">Watch Semantic AI Webinar</a></p><h3>Experience Fluree in Action</h3><p>Ready to see how Fluree’s semantic graph database can transform your data architecture? Get hands-on experience with our platform and discover the power of connected, intelligent data.</p><p><a href="https://flur.ee/get-started/">Try Fluree in Action</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bbd4a3af47b8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/building-corporate-memory-for-enterprise-llms-bbd4a3af47b8">Building Corporate Memory for Enterprise LLMs</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How Decentralized GraphRAG Improves GenAI Accuracy]]></title>
            <link>https://medium.com/fluree/how-decentralized-graphrag-improves-genai-accuracy-0cb3fd861712?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/0cb3fd861712</guid>
            <category><![CDATA[knowledge-graph]]></category>
            <category><![CDATA[retrieval-augmented]]></category>
            <category><![CDATA[graphrag]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[genai]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Wed, 31 Jul 2024 15:52:10 GMT</pubDate>
            <atom:updated>2024-07-31T15:52:10.108Z</atom:updated>
            <content:encoded><![CDATA[<h3><em>Not all RAG methods are created equal. A decentralized knowledge graphRAG approach enables you to securely tap into a broad range of information sources in real-time, increasing the amount of trusted context for higher accuracy rates in GenAI applications.</em></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/0*iz5d5oKOPTeR8jBE.png" /></figure><h3>Introduction</h3><p>Businesses are continuously searching for the most effective approach to implement GenAI models without breaking the bank, failing their compliance obligations, or introducing error-prone hallucinations into their workflows. Organizations are realizing that simply fine-tuning models are not a sustainable method for delivering accurate answers from their internal bases of knowledge.</p><p><a href="https://flur.ee/fluree-blog/llms-are-becoming-less-accurate-heres-where-knowledge-graphs-can-help/">In a previous article</a>, we wrote about the synergistic relationship between knowledge graphs and GenAI. Today’s article will dive into a recent <a href="http://flur.ee/">Fluree report</a> that explores that relationship within the context of RAG.</p><h3>The Threat of Inaccuracy in AI</h3><p>One of the primary challenges with deploying AI in production environments is the risk of inaccuracy. Generative AI, such as large language models (LLMs), often “hallucinates” answers — producing responses that are plausible but incorrect. This is particularly problematic in enterprise settings where accuracy is paramount. Two significant barriers contribute to this issue:</p><ul><li><strong>Security and Privacy Concerns</strong>: Many valuable data assets remain untapped due to stringent security and privacy controls. These controls prevent the use of critical data in AI applications, leading to reliance on less strategic, generalized knowledge.</li><li><strong>Lack of Grounded Truth</strong>: AI models often generate answers based on statistical probabilities rather than verified truths, making it difficult to distinguish between accurate and fabricated responses.</li></ul><h3>Retrieval Augmented Generation (RAG): Enhancing AI Accuracy</h3><p>To address the inaccuracy, organizations are turning to Retrieval Augmented Generation (RAG), a form of information retrieval that integrates live sources of authoritative truth into AI models, reducing the risk of hallucination.</p><p>RAG involves teaching LLMs to retrieve knowledge from external sources during response generation, thus grounding the AI’s output in verified information.</p><p>RAG can utilize various data sources, including:</p><ul><li><strong>Relational Databases, Data Warehouses, and Data Lakes</strong>: For structured data, providing a familiar but limited data interaction model.</li><li><strong>Knowledge Graphs</strong>: Representing data in a network of nodes and relationships, offering a more intuitive structure for LLMs to interpret.</li></ul><h3>What is GraphRAG?</h3><p>Not all RAG methods are created equal. <a href="https://flur.ee/initiatives-landing/enterprise-knowledge-graph/">Knowledge Graphs</a> are emerging as a key enabler of RAG-based GenAI. In fact, Gartner recently indicated that Knowledge Graphs are now a “Critical Enabler” with immediate impact on GenAI. “GraphRAG” refers to an approach to retrieval augmented generation in which information retrieval is based on a structured, hierarchal knowledge graph. Importantly, knowledge graphs help you connect the dots between entities using ontologies to define semantic concepts and relationships. Learn more knowledge graphs <a href="https://flur.ee/fluree-blog/llms-are-becoming-less-accurate-heres-where-knowledge-graphs-can-help/">here</a>.</p><p>In our research conducted in April, 2024, knowledge graphs significantly outperformed traditional relational databases in RAG tasks. By expressing data and metadata semantically, LLMs achieve higher accuracy with less training effort. Here’s a breakdown of their performance:</p><ul><li><strong>Centralized Relational Data</strong>: Initial zero-shot accuracy is low (~20%), improving to 80% with extensive data integration and model fine-tuning.</li><li><strong>Centralized Knowledge Graphs</strong>: Zero-shot accuracy improves to 60–65%, reaching up to 95% with fine-tuning and enriched data.</li><li><strong>Decentralized Knowledge Graphs</strong>: These achieve the highest accuracy, consistently hitting 90–99% by connecting data sources across various domains without physically centralizing them.</li></ul><h3>Wait, What is a Decentralized Knowledge Graph?</h3><p>A decentralized Knowledge Graph is a network of independently managed Knowledge Graphs that can be connected at query time, based on rights and permissions.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*1V8QnXGX48BhNtX9.png" /></figure><p>While centralized RAG (relational or semantic) can solve most accuracy concerns, decentralized knowledge graphs solve the other major barrier: secure, real-time access to a broad range of enterprise knowledge.</p><p>In centralized scenarios, organizations have a hard time getting their employees or GenAI applications access to potential information related to a query, particularly because:</p><ul><li>Security is too cumbersome and risky to manage for one-off business questions.</li><li>The amount of time, energy, and resources it takes to perform a custom ETL from across multiple sources, mask the data as needed, and then hand it over as a RAG Source is not sustainable.</li></ul><p>Because of this, it’s an “all or nothing” culture of data accessibility within the organization. Most of the time, LLMs are used in a primitive, non-proprietary fashion to automate a knowledge task. But imagine giving an LLM access to real-time knowledge from across the organization with granular security built-in?</p><p>In Decentralized Knowledge Graphs, integrated access and usage policies ensure that queries can span across any possible data source and safely and securely access, link, and return decentralized data to the user.</p><p>Each Knowledge Graph could have thousands or millions of nodes (classes and instances), and can physically reside right next to the business applications (on prem, or on the Cloud in a particular geography). The data is not physically stored centrally, but rather a part of a decentralized network connected through semantic web standards.</p><p>Based on specific questions being asked, if a node in one Graph has a relationship to nodes in any of the other Graphs, if they have the permissions to access specific nodes in the other graphs, then they could be included in the answer to the question.</p><p>Simply by nature of creating an enterprise-wide knowledge graph for RAG source data, enabled by semantic standards, ontological context and secure data policies, we concluded that decentralized knowledge graphs consistently outperformed traditional data systems in reducing hallucinations in responses.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*25I4tiAQMvN3LdUp" /></figure><p>Decentralized Knowledge Graphs are also the sustainable choice for ROI: as data volumes grow in size and complexity across distinct domains, Decentralized Graphs are well equipped to support emerging LLM business cases in stride.</p><h3>Implementing Decentralized Knowledge Graphs with Fluree</h3><p>Fluree’s platform is designed to facilitate the creation and management of DKGs.</p><p>Importantly, Fluree solves two unique barriers to RAG-GenAI adoption across a broad range of enterprise knowledge:</p><ul><li><strong>Privacy and Safety: </strong>A data-centric approach to security, in which information is protected by policies directly at the data layer. Fluree’s embedded security policies programmatically enforce data access policies, making it the only knowledge graph that can manage governance, regulation, copyright, and privacy issues dynamically on real-time data.</li><li><strong>Trust</strong>: Fluree uses cryptography to secure all data in a tamper-proof audit trail. This means you can trace back every piece of data to its origin, review its history over time, and build better explainability into your systems.</li></ul><p>Fluree is the only platform that (1) automates the conversion of any/all data silos/lakes/systems into a consistent semantic knowledge graph, ready for RAG, (2) secures data with policy in order to protect information from unwanted exposure or use, and (3) adds digital cryptography into the data tier so that organizations can trust but verify the authenticity of the data sources used in production AI.</p><p><strong>Summary:</strong></p><ul><li>RAG (Retrieval Augmented Generation) is the best way to “ground” GenAI in your data. This involves introducing trusted sources of authoritative data for GenAI to pull from, thus reducing hallucinations and making responses more accurate and contextual.</li><li><a href="https://flur.ee/fluree-blog/llms-are-becoming-less-accurate-heres-where-knowledge-graphs-can-help/">Knowledge Graphs</a> are emerging as a key enabler of RAG-based GenAI. In fact, Gartner recently indicated that Knowledge Graphs are now a “Critical Enabler” with immediate impact on GenAI.</li><li>Decentralized Knowledge Graphs enable on-demand access to a data plane of decentralized graphs with security built in.</li><li>Fluree provides a trusted decentralized knowledge graph for enterprises to build more accurate and precise GenAI applications, grounded in enterprise data.</li><li>Importantly, Decentralized Knowledge Graphs solve two unique barriers to RAG-GenAI adoption across a broad range of enterprise knowledge:</li><li><strong>Privacy and Safety: </strong>A data-centric approach to security, in which information is protected by policies directly at the data layer. Fluree’s embedded security policies programmatically enforce data access policies, making it the only knowledge graph that can manage governance, regulation, copyright, and privacy issues dynamically on real-time data.</li><li><strong>Trust</strong>: Fluree uses cryptography to secure all data in a tamper-proof audit trail. This means you can trace back every piece of data to its origin, review its history over time, and build better explainability into your systems.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0cb3fd861712" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/how-decentralized-graphrag-improves-genai-accuracy-0cb3fd861712">How Decentralized GraphRAG Improves GenAI Accuracy</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Model, Map, Connect: How Fluree Builds Sustainable Knowledge Systems]]></title>
            <link>https://medium.com/fluree/model-map-connect-how-fluree-builds-sustainable-knowledge-systems-d5caa4805491?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/d5caa4805491</guid>
            <category><![CDATA[data-fabric]]></category>
            <category><![CDATA[knowledge-management]]></category>
            <category><![CDATA[data-mesh]]></category>
            <category><![CDATA[fluree]]></category>
            <category><![CDATA[knowledge-graph]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Tue, 21 May 2024 17:02:19 GMT</pubDate>
            <atom:updated>2024-05-21T17:02:19.102Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zkPeaFE6FVVrsGud.png" /></figure><p>Exploring Fluree’s Stack through the Model, Map, Connect Formula</p><p>At Fluree, we’ve embraced the data-centric principles of data management, which ultimately de-couple applications from data. Data-Centric architectures allow data to exist independently of a singular application or silo, creating a network of “autonomous data” that can be re-used and shared across a diverse set of use cases.</p><p>As we’ve discussed in our <a href="https://flur.ee/fluree-blog/introduction-to-data-centricity/">Data-Centric Architecture series</a>, there are a few capabilities that must be set into place to enable autonomous and reusable data, namely:</p><ul><li><a href="https://flur.ee/fluree-blog/semantic-interoperability-exchanging-data-with-meaning/"><strong>Semantic Interoperability</strong></a> — Universal standards to represent data in a form that can be found, accessed, and used across disparate contexts.</li><li><a href="https://flur.ee/fluree-blog/data-centric-security/"><strong>Data-Centric Security</strong></a> — Policies embedded at the data tier to programmatically enforce Identity and Access Management, no matter where, how, or who has query access.</li><li><a href="https://flur.ee/fluree-blog/data-centric-trust/"><strong>Digital Trust</strong></a>– Digital proof in the form of cryptographic signatures to verify data authenticity and origin.</li><li><a href="https://flur.ee/fluree-blog/time-traveling-with-fluree/"><strong>Data-Centric Time</strong></a> — A temporal dimension that time-stamps every piece of data to enable historical queries and audits.</li></ul><p>This kind of radical change cannot happen overnight. Oftentimes, applications and data systems sprawl in heterogeneous formats across diverse departments, and are layered with various and complex governance controls. It might seem impossible to even assemble the people, processes, and technology to embark on such a journey.</p><p>At Fluree, we’ve found that starting small within a single domain and iterating over time yields the best return on investing in a data-centric architecture. We like to place our products into a continuum that defines and assembles knowledge from existing data within that domain: Model, Map, Connect. Think of this framework less as a hard-coded three-step process, but more as a set of interconnected modules that continuously grow alongside each other and your data systems.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F933027745%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F933027745&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F1831281725-7a9cbb24005e9ba8f9e940575909b9df24ccc2d0e0fbb84218a94fe08e41a1ad-d_1280&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1280" height="720" frameborder="0" scrolling="no"><a href="https://medium.com/media/b234a4bb9520b2065a5b3446e031e965/href">https://medium.com/media/b234a4bb9520b2065a5b3446e031e965/href</a></iframe><h3>Model</h3><p><strong>Start with a domain and model it- this could be a business application schema, business or industry ontology, or a standardized schema from schema.org.</strong></p><p>It’s important to understand what your data is about and what concepts your business cares about.<strong> </strong>Organizations can find themselves anywhere on the spectrum of controlled vocabulary maturity; from managing taxonomies in spreadsheets to a fully-fleshed ontology management program. In any case, controlled vocabulary management is an important tool for solidifying a single source of truth for your business. In the <strong>model</strong> step, we develop the conceptual model to represent entities, attributes, and relationships within the knowledge graph.</p><p><a href="https://flur.ee/taxonomy-manager/">Fluree’s Intelligent Taxonomy Manager</a> is a simple, powerful tool to describe business concepts and their interrelationships. Built for collaboration, ITM enables teams to define terminologies and semantic relationships to create consistent vocabularies. You can model pretty much anything and create a unique ID for each concept or piece of knowledge. Fluree ITM also comes with built-in data improvement — powered by AI, the ITM system will automatically recommend new candidate terms to improve the context and depth of your data.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F925544015%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F925544015&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F1821868515-3570317450e25976427b466846c9b60d6aa7978ffbd1d1607677c6c44b91b175-d_1280&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1918" height="920" frameborder="0" scrolling="no"><a href="https://medium.com/media/3738f0fc92af4b9426e6681249a33e36/href">https://medium.com/media/3738f0fc92af4b9426e6681249a33e36/href</a></iframe><h3>Map</h3><p><strong>Fluree’s technology automates the classification and mapping of both structured and unstructured data into a specified semantic vocabulary, automating knowledge representation from across your data systems.</strong></p><p>Once you’ve modeled your domain, it’s time to discover and map existing data to your controlled vocabulary. This is the step that many organizations find the most painstaking and costly. Traditionally, getting access to relevant data and then harmonizing, classifying, and cleansing it takes an overwhelming manual effort from data engineering teams and Subject Matter Experts. The challenges and costs only increase if data quality is overlooked, leading to a portfolio of “<a href="https://flur.ee/fluree-blog/how-to-eliminate-dirty-data-for-good/">dirty data</a>.”</p><p>Once again, Fluree’s approach to semantic data management provides a foundation for (1) automating the manual data mapping processes and (2) creating golden records that will never become obsolete, dirty, or outdated again.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*m22sNKUlTqf5BKc5.png" /></figure><h3>Structured Data</h3><p>For Structured Data, <a href="https://flur.ee/sense/">Fluree Sense</a> uses AI and ML to automatically ingest, classify, and master records from an array of disparate data sources. With more than 30 out-of-the-box connectors to common data sources (SAP, Oracle, Salesforce, and others), Fluree Sense can plug into any structured data source and handle all aspects of data discovery, cleansing, mapping, and conversion to your target data model. Fluree Sense uses subject matter expertise to reinforce its ML capabilities, which will ultimately scale to billions of records and 99% accuracy for batch or continuous jobs.</p><h3>Unstructured Data</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*qmGo4Ug4bNtYE815.png" /></figure><p>For Unstructured Data, <a href="https://flur.ee/sense-unstructured/">Fluree CAM</a> (Content AutoTagging Manager) uses NLP and various out-of-the-box ML workflows to automatically classify and tag content. Using your model managed in Fluree ITM, organizations can set up CAM workflows to integrate into any kind of unstructured data source (PDFs, Videos, Audio, Sharepoint, Zendesk, and more) and auto-classify content on the fly.</p><p>Importantly, Fluree CAM and Fluree Sense both publish out knowledge graph ready data:</p><ul><li><a href="https://flur.ee/fluree-blog/making-data-fair/">“FAIR” data</a>: information that is Findable, Accessible, Interoperable, and Reusable by humans and machines.</li><li>Classified under your target controlled vocabulary</li><li>Relationships between data are semantically defined</li><li>In RDF format, serialized as JSON-LD</li></ul><h3>Connect</h3><p><strong>Analyze, interpret, and securely share your data with security policies directly built-in. Connect to apps and collaborate on datasets.</strong></p><p>Once you’ve architected a semantic knowledge graph from your various systems by defining your controlled vocabularies and mapping existing data, it’s time to provide access to your information for various data use cases. We need to access, share, and re-use this data for analytics, regulatory reporting, and new applications.</p><p>Fluree CAM, Sense, and ITM all integrate into <a href="http://flur.ee/">Fluree Core</a>, our knowledge graph database built for secure and trusted data collaboration. Fluree Core uses JSON-LD as the de-facto standard for transacting and querying your knowledge graph database, making it accessible for developers and analysts alike to build queries on top of consumption-ready data.</p><p>Fluree Core was built with the three key ingredients to enable superior knowledge graph data management:</p><ul><li><strong>Security Policies</strong> — Build, manage, and enforce data policy as data within your knowledge graph. Embedded read and write access control at the data tier opens up unlimited sharing capabilities.</li><li><strong>Digital Trust</strong> — Fluree embeds lineage and cryptographic proof of provenance as key metadata for each and every new transaction — enabling data consumers to trust and verify the information they are using came from a reliable source and has not been tampered with.</li><li><strong>Linked Data Standards</strong> — As an RDF Semantic graph database, Fluree uses universal standards for linked data, allowing for schema at query time and unlocking powerful knowledge graph capabilities.</li></ul><h3>Expand</h3><p>Model, Map, and Connect are modules that will continuously evolve. Perhaps there are updates to your existing taxonomy or ontology, or an entirely new vocabulary that needs to be set into place. Perhaps you’re ready to expand or otherwise connect to new data sources to grow your knowledge graph capabilities. Perhaps there are new applications or business partners that need controlled secure access to your knowledge.</p><p>Because we’ve built a knowledge ecosystem on top of semantic standards, you can integrate any kind of new workflow to adapt to changing business needs. That’s the beauty of data-centric architectures — data exists independently of a singular application or data silo and can serve across multiple use cases. Data becomes less of a cost-center and more of an investment that grows with your business over time.</p><h3>Recap</h3><p>Here’s a TL:DR version of the model map connect framework:</p><ol><li>Identify a domain and model it with a semantic ontology.</li><li>Map existing (or new) data to your model.</li><li>Securely connect consumption patterns, applications, and business users to the data.</li><li>Expand and evolve your data models, data consumers, and data sources using automation.</li></ol><p>Want to see this in action? Review this 20-minute video taking you through each module in the Fluree stack:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F933324154%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F933324154&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F1831729248-ee22f62b4ba64f4cf664ad442da5c21adcb0e7464a0f2d59a7785f035d6f2897-d_1280&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1920" height="1080" frameborder="0" scrolling="no"><a href="https://medium.com/media/60dac12e81da059890194e41e9142735/href">https://medium.com/media/60dac12e81da059890194e41e9142735/href</a></iframe><p>Interested in getting started? Reach out to us <a href="http://flur.ee/contact">here</a>!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d5caa4805491" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/model-map-connect-how-fluree-builds-sustainable-knowledge-systems-d5caa4805491">Model, Map, Connect: How Fluree Builds Sustainable Knowledge Systems</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is SHACL? (With Examples)]]></title>
            <link>https://medium.com/fluree/what-is-shacl-with-examples-2697f659d465?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/2697f659d465</guid>
            <category><![CDATA[knowledge-management]]></category>
            <category><![CDATA[data-governance]]></category>
            <category><![CDATA[shacl]]></category>
            <category><![CDATA[knowledge-graph]]></category>
            <category><![CDATA[data-quality]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Tue, 07 May 2024 20:17:21 GMT</pubDate>
            <atom:updated>2024-05-07T20:17:21.325Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*78f65hISmtGDmlCnFx04FA.png" /></figure><p>SHACL is a critical tool for anyone involved in the management, curation, and utilization of RDF data. Its ability to enforce complex data quality rules in a flexible and scalable manner makes it indispensable for maintaining high-quality, interoperable datasets in the age of big data and the semantic web.</p><p>Shapes Constraint Language (SHACL) is a powerful language for validating RDF (Resource Description Framework) knowledge graphs against a set of conditions. These conditions are known as shapes and can be used by data governance professionals to streamline consistent data quality across the organization’s data ecosystem. SHACL was developed by the World Wide Web Consortium (W3C) into the open, industry standard for data quality assurance in data-centric, semantic technologies and linked data projects.</p><p>In Fluree’s graph database, users leverage policy-based SHACL to express constraints like required properties, data types, cardinality, and closed class shapes.</p><h3>What is the difference between OWL and SHACL?</h3><p>OWL and SHACL are typically compared as both ways in which to manage and maintain datasets, but their core competencies differ. OWL helps with inference; SHACL helps with validation. Let’s break it down further:</p><ul><li><strong>OWL</strong> is primarily designed for defining and reasoning about ontologies. It provides a rich set of constructs for describing the relationships between concepts in a domain, enabling sophisticated inferencing capabilities about the types of entities and their relationships. OWL is used to create complex domain models and to infer new knowledge from existing data.</li><li><strong>SHACL</strong>,<strong> </strong>on the other hand, is focused on data validation. It allows developers and data architects to define constraints on the structure and content of RDF graphs, ensuring that the data adheres to specified patterns, value ranges, or other criteria. While OWL focuses on enabling inference, SHACL is specifically tailored for validation, offering a more direct approach to enforcing data quality rules.</li></ul><h3>What does SHACL look like?</h3><p>SHACL uses a graph-based syntax, where shapes are defined as RDF graphs. Let’s take a look at a simple example: let’s say that we want to ensure that all values assigned to the property “schema:birthday” are enforced as xsd:dateTime (in plain english: birthdays are formatted as valid dates and times). In Fluree, we could insert our constraint in a ledger like this:</p><pre>{<br>  &quot;@context&quot;: &quot;https://ns.flur.ee&quot;,<br>  &quot;ledger&quot;: &quot;ledger/data-type&quot;,<br>  &quot;insert&quot;: {<br>    &quot;@id&quot;: &quot;ex:UserShape&quot;,<br>    &quot;@type&quot;: [&quot;sh:NodeShape&quot;],<br>    &quot;sh:targetClass&quot;: { &quot;@id&quot;: &quot;ex:Person&quot; },<br>    &quot;sh:property&quot;: [<br>      {<br>        &quot;sh:path&quot;: { &quot;@id&quot;: &quot;schema:birthDate&quot; },<br>        &quot;sh:datatype&quot;: { &quot;@id&quot;: &quot;xsd:dateTime&quot; }<br>      }<br>    ]<br>  }<br>}</pre><p>This example defines a shape for `ex:Person` instances, ensuring that the ‘birthDate’ property of such entities, if present, must be a valid date-time value.</p><h3>What are the benefits of using SHACL for data validation?</h3><p>Data governance is increasingly becoming critical, as organizations need to leverage clean, available, and trusted data on a day-to-day basis. By employing a global framework for data validation, organizations can save an immense amount of time and effort in standardizing data for analytics and re-use.</p><p>SHACL offers several benefits for data validation:</p><ul><li><strong>Flexibility and Expressiveness</strong>: SHACL can express a wide range of constraints, including property existence, value type, value range, cardinality, and more complex conditions.</li><li><strong>Scalability</strong>: SHACL validation can be efficiently implemented, making it suitable for large datasets.</li><li><strong>Standardization</strong>: As a W3C standard, SHACL ensures interoperability and consistency across different tools and platforms. This helps make not only the data interoperable, but also the rules that govern data consistency and validation interoperable.</li><li><strong>Declarative Approach</strong>: SHACL allows the expression of validation rules in a declarative manner, separating the rules from their execution, which can improve maintainability and understandability of data constraints.</li></ul><h3>What kinds of quality constraints can SHACL enforce?</h3><p>SHACL can enforce a variety of quality constraints, including:</p><ul><li><strong>Structural Constraints</strong>: Ensuring that data adheres to a specific schema or model, such as required properties, permissible property values, or specific class hierarchies.</li><li><strong>Value Constraints</strong>: Limiting the values that can be taken by properties, including data types, value ranges, and pattern matching.</li><li><strong>Cardinality Constraints</strong>: Defining the minimum and maximum occurrences of properties.</li><li><strong>Logical Constraints</strong>: Applying logical conditions to properties and values, such as equality or inequality, and combinations thereof through logical operators.</li></ul><h3>How does SHACL improve enterprise data at scale?</h3><p>For enterprises dealing with vast amounts of data, maintaining data quality is paramount. SHACL provides a robust framework for ensuring that data across the organization conforms to agreed-upon standards and models. At scale, SHACL helps in:</p><ul><li><strong>Automating Data Validation</strong>: Automated tools can leverage SHACL shapes to validate data as it is ingested, updated, or transformed, ensuring continuous data quality without manual intervention.</li><li><strong>Enforcing Data Governance</strong>: SHACL shapes can embody data governance policies, ensuring compliance with internal and external data standards and regulations.</li><li><strong>Improving Data Interoperability</strong>: By enforcing standardized data models and structures, SHACL facilitates data sharing and interoperability both within the enterprise and with external partners.</li><li><strong>Enhancing Data Quality</strong>: Consistent application of SHACL validation helps identify and rectify data quality issues early, reducing errors and improving the reliability of data-driven decisions.</li></ul><p>SHACL is a critical tool for anyone involved in the management, curation, and utilization of RDF data. Its ability to enforce complex data quality rules in a flexible and scalable manner makes it indispensable for maintaining high-quality, interoperable datasets in the age of big data and the semantic web.</p><h3>Try SHACL out with Fluree!</h3><p>Head on over to our <a href="https://next.developers.flur.ee/docs/reference/cookbook/#shacl">cookbook documentation</a> for some examples that you can test out with your <a href="http://data.flur.ee/signup">free Fluree cloud account</a>!</p><p>*special thanks to Buck Flannigan and Laurent Bégin for their contributions to this article</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2697f659d465" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/what-is-shacl-with-examples-2697f659d465">What is SHACL? (With Examples)</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[LLMs Are Becoming Less Accurate. Here’s Where Knowledge Graphs Can Help.]]></title>
            <link>https://medium.com/fluree/llms-are-becoming-less-accurate-heres-where-knowledge-graphs-can-help-a4e804ca2f9e?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/a4e804ca2f9e</guid>
            <category><![CDATA[ontology]]></category>
            <category><![CDATA[semanticweb]]></category>
            <category><![CDATA[knowledge-graph]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Thu, 25 Apr 2024 16:00:43 GMT</pubDate>
            <atom:updated>2024-04-25T16:00:43.317Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Qx8yCC3dmWRoZahi.png" /></figure><h3><em>While LLMs are adept at pattern recognition and generating human-like text, they often struggle with maintaining accuracy over time and across diverse topics. Knowledge graphs can bridge this gap by providing a constantly updated and structured source of context and truth, ensuring that the LLM’s outputs remain reliable and up-to-date.</em></h3><p>Large Language Models (LLMs) like GPT-4 have been at the forefront of the recent AI craze, offering jaw-dropping capabilities in (seemingly) understanding and generating human-like text. These models, trained on vast datasets, have demonstrated proficiency in a range of tasks, from writing essays to coding (as well as demonstrating their abilities in <a href="https://www.nbcnews.com/tech/tech-news/chatgpt-passes-mba-exam-wharton-professor-rcna67036">passing a final MBA exam</a>, <a href="https://news.ycombinator.com/item?id=37448005">cloning text dialect to automate Whatsapp Responses</a>, and producing a hilarious <a href="https://people.com/julia-louis-dreyfus-uses-chat-gpt-to-write-hilarious-wsj-magazine-innovators-award-acceptance-speech-8386425">acceptance speech at the WSJ Innovators Awards</a>).</p><p>However, recent findings have brought to light performance inconsistencies in these LLMs.</p><p>Studies have indicated <a href="https://www.zdnet.com/article/what-is-a-ai-drift-and-why-is-it-making-chatgpt-dumber/">fluctuating levels of accuracy and reliability,</a> raising questions about their dependability for critical applications, as highlighted this <a href="https://www.cnbc.com/2024/01/17/generative-ai-dominates-davos-discussions-as-companies-focus-on-accuracy.html?&amp;qsearchterm=davos%20ai">year in Davos</a>.</p><p>For the above light-hearted use cases, this level of model drift may not be problematic. But in high-stake scenarios, where precise and factual information is critical (such as healthcare, supply chain management, financial accounting, or beyond), organizations cannot afford to gamble on accuracy.</p><h3>Evidence Shows LLMS Become Less Accurate Over Time</h3><p>A <a href="https://arxiv.org/abs/2307.09009">study</a> conducted by researchers at Stanford and Berkeley found that the ability of LLMs to handle tasks like solving mathematical problems and generating code varied significantly over time. The study showed how GPT-4’s accuracy in certain tasks dropped dramatically, showcasing a major concern in the context of LLM reliability.</p><p>This variability in accuracy — particularly a downward slope — poses a significant challenge, especially for applications that heavily rely on the consistent and accurate output of these models. The findings underscore the need for solutions to enhance the stability and reliability of LLMs in practical applications.</p><h3>Why? The Answer is “Fuzzy.”</h3><p>Open AI’s models are proprietary, so the exact reasons why accuracy is slipping are unconfirmed. However, we can broadly attribute the cause as a cyclical combination of “fuzzy matching” and model drift.</p><p>Model drift simply refers to a degradation of performance, due to:</p><ol><li>A changing data environment where the LLM’s original model training becomes less relevant over time.</li><li>Attempts to fine-tune one area of the model while inadvertently causing another area to perform worse</li></ol><p>Combined with LLMs’ fuzzy matching approach to answering questions, we can start to see severe implications.</p><p>The “fuzzy matching” approach to input/output of a model is grounded in probabilistic modeling, where the LLM generates content based on the likelihood of word sequences as learned from vast datasets. While this allows for fluid and human-like text generation, it inherently lacks a precise understanding of the factual correctness or specific relationships between entities. It simply knows what’s probably next — think “auto-fill” in Google Docs, or suggested text replies on your iPhone, but instead of being trained on just your words, trained on the internet.</p><p>The technical basis of this approach lies in machine learning algorithms, particularly <a href="https://news.mit.edu/2017/explained-neural-networks-deep-learning-0414">neural networks</a>, which process input data (text) and generate outputs based on learned patterns. These models don’t “understand” content in the human sense but rather rely on statistical correlations. This means they can effectively mimic language patterns but may falter when precise, context-specific accuracy is required.</p><p>For instance, when asked about specific factual information or nuanced topics, LLMs might produce answers that are plausible but not necessarily accurate. We know this as the <a href="https://www.cnbc.com/2023/12/22/why-ai-chatbots-hallucinate.html">“hallucination” crisis</a>.</p><p>In scenarios where factual accuracy and consistency are paramount, this probabilistic approach can be a significant limitation, potentially leading to unreliable or misleading information. Combined with the inevitability of model drift, fuzzy matching can pose a serious threat to a business looking to deploy production AI models with sustainable accuracy.</p><h3>Enter Knowledge Graphs</h3><p>What is the opposite of “fuzzy matching” when it comes to LLM accuracy? Answer: Explicit, Structured Data.</p><p>Unlike fuzzy matching’s probabilistic approach, structured data representation relies on explicitly defined data, definitions, and relationships</p><p>This is where <a href="https://flur.ee/initiatives-landing/enterprise-knowledge-graph/">knowledge graphs</a> — like Fluree — shine.</p><p>Where LLMs might offer ambiguity, Knowledge Graphs offer explicitness.</p><h3>A Quick Knowledge Graph Refresher</h3><p>A knowledge graph is a database that represents information as a graph network of entities and relationships between them. They are commonly used for advanced analytics, recommendation systems, semantic search engines, or data discovery, cataloging and integration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*rMKwtditMWBGrK1L" /></figure><p>In knowledge graphs, data is represented in the form of semantic objects (or entities), and the connections or relationships between these objects are explicitly mapped. Importantly, they leverage ontologies, which provide formal, structured representations of data and relationships within data.</p><p>Most important to the matter at hand, knowledge graphs are explicit. They don’t just contain raw data; they define and map the relationships between data points in a global, understandable, and machine-readable way. This explicitness ensures that every piece of information is contextualized, making the data not only more accessible but also more meaningful and reliable. This characteristic of knowledge graphs fundamentally enhances their utility in applications where precision and context are critical.</p><h3>Explicit Knowledge + LLMs</h3><p>Knowledge graphs can significantly enhance the accuracy and consistency of LLMs, offering a tangible solution to the challenges of fluctuating performance and reliability faced by these models. The explicit, structured, and semantically-defined data that knowledge graphs organize are vital for leveraging LLMs to their fullest potential:</p><ul><li><strong>Accuracy</strong>: By integrating knowledge graphs, LLMs can tap into a well-organized repository of explicitly-defined, contextual information, significantly enhancing their accuracy and consistency.</li><li><strong>Reliability: </strong>The structured nature of knowledge graphs ensures that LLMs have a reliable reference point, reducing the likelihood of generating incorrect or inconsistent responses.</li></ul><p>Even further, knowledge graphs help prioritize what’s relevant and important.</p><p><a href="https://towardsdatascience.com/is-small-data-the-next-big-thing-in-data-science-9acc7f24907f">Andrew Nyguen</a> suggests in his argument for “data-centric AI” that smaller sets of high quality data are far superior for model training and deployment. While today’s Large Language Models are trained on vast, general data, enterprise AI will instead succeed with this small data approach. Knowledge graphs can serve this “small data” approach through interconnecting across various domains and providing context-rich, targeted information that enhances model understanding and decision-making.</p><h3>Knowledge Graphs Don’t Just Save LLMs; They Supercharge Them.</h3><p>It’s not that LLMs are failing. They are performing what they were engineered to do; generating probabilistic outputs based on statistical correlations. LLMs just need contextual grounding. If LLMS were unleashed on well-defined, well-structured, and well-governed data, their accuracy and reliability would remain very high at scale and over time.</p><p>A paper published in <a href="https://arxiv.org/pdf/2306.08302.pdf">IEEE</a> represents the synergistic relationship between Knowledge Graphs and LLMs, pointing to how each technology compliments the other. Where LLMs lack contextual grounding through implicit knowledge, knowledge graphs provide structured truth. And where knowledge graphs might lack completeness and generalized knowledge, LLMs help fill that gap.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Qx8yCC3dmWRoZahi.png" /></figure><p>Businesses aiming to truly build effective AI should prioritize structured data management and governance as their strategy and should appoint knowledge graphs as the vehicle to deliver and maintain these capabilities.</p><h3>Thanks for reading!</h3><p>Fluree offers a full suite of semantic data management products to build, maintain, leverage and share enterprise knowledge. Learn more about our comprehensive <a href="https://flur.ee/initiatives-landing/enterprise-knowledge-graph/">knowledge graph solutions</a> and book a call to speak with an expert <a href="https://flur.ee/initiatives-landing/enterprise-knowledge-graph/">here</a>. Or, dive into more information about our product suite below:</p><ul><li><a href="http://flur.ee/core"><strong>Fluree Core</strong></a> — Semantic Graph Database with Built-In Security and Digital Trust</li><li><a href="https://flur.ee/taxonomy-manager/"><strong>Fluree ITM</strong></a> — Intelligent Taxonomy and Ontology Manager</li><li><a href="https://flur.ee/fluree-products-ai-ml-data-pipeline/"><strong>Fluree Sense</strong></a>– Semantic Pipeline to Transform Legacy Structured Data to Knowledge Graphs</li><li><a href="https://flur.ee/sense-unstructured/"><strong>Fluree CAM</strong></a>– Intelligent Content Auto-Tagging for Digital Media Assets</li><li>Image Inspiration Credit: <a href="https://ai.plainenglish.io/llms-knowledge-graphs-234c443b3669">LLMs &amp; Knowledge Graphs, Medium</a></li></ul><p><strong>Did you enjoy this story? Consider signing up for </strong><a href="http://eepurl.com/gcsCV5"><strong>Fluree’s Newsletter</strong></a><strong>.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a4e804ca2f9e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/llms-are-becoming-less-accurate-heres-where-knowledge-graphs-can-help-a4e804ca2f9e">LLMs Are Becoming Less Accurate. Here’s Where Knowledge Graphs Can Help.</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is JSON-LD?]]></title>
            <link>https://medium.com/fluree/what-is-json-ld-e6ddaf1611f7?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/e6ddaf1611f7</guid>
            <category><![CDATA[linked-data]]></category>
            <category><![CDATA[json-ld]]></category>
            <category><![CDATA[json]]></category>
            <category><![CDATA[graph-database]]></category>
            <dc:creator><![CDATA[Kevin Doubleday]]></dc:creator>
            <pubDate>Wed, 24 Jan 2024 18:54:43 GMT</pubDate>
            <atom:updated>2024-01-24T18:54:43.060Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/920/1*WEjKrxdS6NG1BjlUOp-ohA.png" /></figure><p>Future-Proof Your Data with JSON-LD</p><p><em>Fluree has officially released native </em><a href="http://flur.ee/json-ld"><em>JSON-LD support</em></a><em>. But what exactly is JSON-LD, and why does it matter?</em></p><p>At Fluree, we are big fans of the semantic web standards — especially RDF — because they allow for <a href="https://flur.ee/2021/03/10/semantic-interoperability-exchanging-data-with-meaning/">data to be organized and exchanged with universal meaning</a>. Semantic standards are necessary for a future internet in which machines and humans can seamlessly connect to, leverage, and collaborate around disparate sets of linked data. The implications — especially in the era of emerging AI technologies, enterprise knowledge graphs, and industry data sharing ecosystems — are profound.</p><p>But we are also realists — in the last ten years, most developers have chosen convenience over the grand vision of the semantic web, opting for JSON document data stores. JSON’s incredibly lightweight interchange format has seen mass adoption thanks in part to the influence of javascript on software over the last decade.</p><p><strong>But the simplicity of JSON backfires when it comes to any case of reusing, sharing, or linking data</strong>. Without adding the context of data types, classifications, and relationships, data can easily become isolated, leading to a fragmented and siloed understanding of its implications.</p><p>That’s why we’re passionate about the JSON-LD specification — which Fluree has fully implemented and now supports.</p><h3>What is JSON-LD?</h3><p>JSON-LD — or JavaScript Object Notation Linked Data — combines the simplicity and ease of JSON documents with the power of linked data. JSON-LD allows you to encode meaning within your JSON document through the use of a shared vocabulary — a total upgrade from your typical JSON payload.</p><p>JSON-LD is JSON for linked data. In fact, A JSON-LD document is always a valid JSON document. Broadly speaking, there are two core capabilities that the “LD” part provides:</p><ol><li>It allows you to embed or reference the definition of objects within your JSON payload. For example, the @context element embedded in a JSON-LD document maps object properties to concepts from an RDF ontology (a universally standardized schema).</li><li>It allows you to assign IRIs to properties in order to establish uniqueness.</li></ol><h3>🔍 What is an IRI?</h3><p>An IRI (Internationalized Resource Identifier) is a generalization of the more commonly known URI (Uniform Resource Identifier). While URIs are limited to a subset of the ASCII character set, IRIs provide a way to include characters from a wider range, including Unicode. This expansion allows for the use of scripts and characters from various languages around the world, making it more globally inclusive.</p><p>In the context of linked data and the semantic web, IRIs are used to uniquely identify resources. This ensures that there’s no ambiguity when referring to particular entities, concepts, or relationships, regardless of where they originate or the language they’re represented in.</p><h3>Understanding @context, @id and @type</h3><p><strong>@context</strong> — The @context tag in a JSON-LD payload gives us just that — context — in the form of mapping objects in the document to a vocabulary. For example, one might map @context to schema.org, a well-known reference site for structured data schemas.</p><p><strong>@id</strong> — The @id tag gives us a way to establish uniqueness for data — which is fundamentally critical to prevent ambiguity and duplication in linked data systems. The @id tag is sort of like a digital version of an address where we can ‘park’ unique data at URIs and point to them later on. Uniqueness in data allows us to distinguish between things or recognize that one thing in a dataset is the exact same as one thing in another dataset.</p><p><strong>@type</strong> — Represents the data type or class of the object, providing a way to categorize the data in terms of what it represents in the real world.</p><h3>Let’s take a look at a JSON-LD document:</h3><pre>{<br>  &quot;@context&quot;: {<br>    &quot;schema&quot;: &quot;https://schema.org/&quot;,<br>    &quot;name&quot;: &quot;schema:name&quot;,<br>    &quot;author&quot;: &quot;schema:author&quot;,<br>    &quot;datePublished&quot;: &quot;schema:datePublished&quot;,<br>    &quot;articleBody&quot;: &quot;schema:articleBody&quot;<br>  },<br>  &quot;@type&quot;: &quot;schema:Article&quot;,<br>  &quot;@id&quot;: &quot;https://example.com/articles/json-ld-introduction&quot;,<br>  &quot;name&quot;: &quot;Introduction to JSON-LD&quot;,<br>  &quot;author&quot;: {<br>    &quot;@type&quot;: &quot;schema:Person&quot;,<br>    &quot;name&quot;: &quot;John Doe&quot;<br>  },<br>  &quot;datePublished&quot;: &quot;2023-10-27&quot;,<br>  &quot;articleBody&quot;: &quot;JSON-LD is a powerful way to represent linked data using JSON. It combines the simplicity of JSON with the capabilities of linked data, allowing for richer metadata, better SEO, and improved data interoperability...&quot;<br>}</pre><ul><li><strong>@context:</strong> This defines the context of the JSON-LD document. It maps terms to IRIs so that properties such as “name” or “author” are clearly associated with concepts defined at “schema.org”, a popular vocabulary for structured data on the internet.</li><li><strong>@type:</strong> Specifies that this JSON-LD document represents an article. The “schema:Article” IRI defines this as an article type as per the schema.org vocabulary.</li><li><strong>@id:</strong> Provides a unique identifier for this article. In this case, it’s a URL to the article on “example.com”.</li><li><strong>author:</strong> Contains information about the author of the article. Nested within it, we see another @type that specifies this entity as a “Person” as per the schema.org vocabulary. The author has a “name” property.</li></ul><h3>Why use JSON-LD?</h3><p>JSON-LD gives developers a familiar format in which to deliver the value of linked data and provides organizations with an easy path to integrate the benefits of linked data into a broader range of their data management operations.</p><blockquote><strong>What’s the deal with Linked Data?</strong></blockquote><blockquote>Linked data is the practice of connecting data across domains through the use of semantic standards. Linked data uses standardized schemas known as ontologies to transform information into organized, connected, and contextual knowledge. These properties allow data to become interlinked and machine-readable on the web and promote knowledge discovery and federated analytics. Linked data can be applied within the enterprise in the form of semantic knowledge graphs, on the public internet in the form of Open Linked Data, or in a hybrid model in the form of semantic data sharing ecosystems.</blockquote><p>JSON-LD allows organizations to gain all the benefits of linked data — without the hassle and overhead of transforming data or forcing developers to use a foreign data format. JSON-LD immediately increases the value of JSON documents by making the data stored within them interoperable with other linked data systems — such as internal knowledge graphs or on the open linked web — as well as by providing an accessible medium for developers and applications to interact with linked data.</p><p>For developers used to working with document data stores like MongoDB, JSON-LD is a bridge to the world of linked data. Linked data is the reason <a href="https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/37430.pdf">Google has been able to provide</a> such unique and sophisticated services to consumers — but that power is about to expand to far more than simple search engine optimization. Linked data allows organizations, people, and applications to access and share distributed sets of knowledge under global semantic standards- a powerful concept in the context of the modern data age.</p><p>Here’s a few examples of how JSON-LD makes an impact:</p><ul><li><strong>Building Data-Centric Applications</strong> — Software increasingly produces data that has more than one purpose — whether analytics, sharing, compliance reporting, or re-operationalization into a new application — data is versatile. JSON-LD future-proofs data for re-use within a data-centric architecture, allowing data to exist independently of a singular application and to power multiple data applications without duplication. <a href="https://flur.ee/2020/10/05/introduction-to-data-centricity/">Read more about data-centricity here</a>.</li><li><strong>Participation in the Semantic Web</strong> — JSON-LD gives document data a path towards participation in the semantic web. This core initial use case has seen wide adoption in the SEO (Search Engine Optimization) field, because JSON-LD helps add metadata and structure to webpages. Beyond SEO, JSON data will need to interoperate at webscale for a variety of use cases — and JSON-LD is the perfect tool for the job.</li><li><strong>Broadening the Impact and Capabilities of Enterprise Knowledge Graphs</strong>: Linking data allows the formation of ontology-driven knowledge graphs, a data management trend that supports advanced analytics and insights through the use of semantic data standards. JSON-LD allows knowledge graphs to incorporate semantically-tagged documents as part of their data domains — but it also provides a medium in which to operationalize applications atop knowledge graphs. Knowledge graphs should be dynamic and operational in that they power multiple data applications and are contributed to by multiple data source domains. This is supported by our <a href="https://flur.ee/2020/10/05/introduction-to-data-centricity/#:~:text=The%20Data%2DCentric%20Architecture%20treats,the%20entire%20data%20value%20chain.">data-centric philosophy</a> in which data should be able to connect to, influence and power multiple data applications. Because JSON-LD extends JSON with linked data capabilities, adding and deleting objects in knowledge graphs is as simple as performing transactions against a document database like MongoDB. This read/write capability then allows a knowledge graph to serve as an operational backend to real-time web-based applications that use a JSON interface</li><li><strong>Feeding Analytics Better Data — </strong>JSON-LD extends the value and versatility of data with context, thereby preparing it for downstream activity. For instance, JSON-LD prepares data (that would otherwise be too simple or obscure) for an instantaneous analytical experience within data science teams. Because JSON-LD data arrives with self-describing context and a standardized semantic schema, downstream systems can immediately understand, evaluate, and take action on insights. For these systems, JSON-LD makes data immediately<a href="https://flur.ee/2020/07/08/making-data-fair/">F.A.I.R. (Findable, Accessible, Interoperable and Reusable)</a>.</li><li><strong>Machine-readability </strong>— JSON-LD upgrades JSON documents to become machine-readable, meaning computers can automatically exchange, access and understand information without the need for a human intermediary. This capability will be crucial for data to participate in the machine-to-machine web and power the next wave of automation and AI.</li><li><strong>Data sharing ecosystems</strong> — Data sharing ecosystems are on the rise, introducing complexities that data systems have not had to account for in the past. For example, data from a single repository might be contributed to by multiple organizations. Or, records in a consortium might need to flow across systems and applications — in a hospital system, for example, patient information is securely collaborated on by different clinics and insurance providers. Linked data supports these complexities — with semantic standards, we can exchange information across systems, execute federate queries across multiple disparate data sets, and collaborate around data sets with standardized schemas.</li><li><strong>Verifiable Credentials &amp; Decentralized Semantics</strong> — Allows for authentication and assurance of the data’s source, ensuring its validity and credibility. By incorporating decentralized semantics, data gets rooted in a universally accepted framework, making it trustworthy and interoperable across a network. (check out our cool work with the <a href="https://www.youtube.com/watch?v=XZUZSlsAepg">Trusted Learner Network</a>!)</li></ul><h3>Conclusion</h3><p>At Fluree, our vision of the Web3.0 involves an internet of trusted and linked data. The problem is that our tools haven’t addressed the need for data to be linked — developers have found document databases quite preferable for simplicity’s sake. It’s time to prepare our data — at the source of its creation — with capabilities for interoperability and exchange across disparate data systems. JSON-LD is that needed bridge. Try JSON-LD out for yourself <a href="http://dev.flur.ee/">here</a>. There are emerging innovations for trusted and interoperable data exchange that JSON-LD will help fuel — including personal knowledge graphs and <a href="https://www.w3.org/TR/vc-data-model/">verifiable credentials</a>. We’ll cover examples of those specifications in future blog posts.</p><h3>Resources</h3><p><a href="https://json-ld.org/">• https://json-ld.org/</a></p><p><a href="https://flur.ee/json-ld/">• https://flur.ee/json-ld/</a></p><p><a href="https://www.slideshare.net/gkellogg1/json-for-linked-data">• https://www.slideshare.net/gkellogg1/json-for-linked-data</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e6ddaf1611f7" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/what-is-json-ld-e6ddaf1611f7">What is JSON-LD?</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Fluree 2023 Year In Review]]></title>
            <link>https://medium.com/fluree/fluree-2023-year-in-review-023bb59c2fc5?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/023bb59c2fc5</guid>
            <category><![CDATA[fluree]]></category>
            <category><![CDATA[year-in-review]]></category>
            <category><![CDATA[highlights]]></category>
            <dc:creator><![CDATA[Malinda Massey]]></dc:creator>
            <pubDate>Thu, 04 Jan 2024 17:28:27 GMT</pubDate>
            <atom:updated>2024-01-04T17:28:27.683Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nTRfr19flHgKWTvCjxlrmg.png" /></figure><blockquote>2023 was a big year for Fluree — from new partnerships to new product launches, here are 8 Fluree highlights from the last year.</blockquote><p>The 2023 chapter has ended and we’ve turned the page into 2024. We turn this page with excitement and optimism for the year ahead. Before this year escapes us, as they all do, we’d like to reflect on our favorite Fluree highlights and noteworthy moments from the last 365 days.</p><h3>Fluree Launches Version 3</h3><p><strong>In 2023: Fluree launched the all-new V3, a JSON-LD Database with access policies, digital trust, and shared vocabularies built right in.</strong></p><p>Fluree Version 3 presents a groundbreaking fusion of JSON-LD support and linked graph data capabilities. The release introduces a variety of advanced features including direct query endpoint access, immutability assurance, integrated policy controls, shared vocabularies, and a user-friendly collaboration interface. These features collectively bolster Fluree’s commitment to secure, trusted, and seamlessly interconnected data sharing — offering a pathway for developers, engineers, analysts, and the like to harness the boundless potential of linked data.</p><p>Learn more about Fluree V3 in <a href="https://app.livestorm.co/fluree/getting-started-with-fluree">this webinar</a>, or read about it <a href="https://flur.ee/fluree-blog/launching-fluree-version-3-the-next-generation-json-ld-database/">here</a>. Jumpstart your experience with Fluree V3 <a href="https://data.flur.ee/login">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*u49R6NUfiPUrafDK.png" /></figure><h3>Fluree Series A Funding</h3><p><strong>In 2023: Fluree Closes $10M Series A Round Funding.</strong></p><p>One of our most significant 2023 milestones included securing $10 million in a Series A funding round led by Sinewave Ventures. This funding helps fuel our efforts to develop data-centric infrastructure and assist enterprises in modernizing their legacy systems into collaborative data platforms. We’re grateful for the support of our investors and excited about the prospects this funding brings as we drive innovation in decentralized data management.</p><p>For more details on this funding round, read <a href="https://www.businesswire.com/news/home/20230420005076/en/Fluree-Closes-10M-Series-A-Round-for-Web3-Data-Management-Modern-Data-Infrastructure">here</a>.</p><h3>Fluree x Vitality TechNet</h3><p><strong>In 2023: Fluree technology joined forces with Vitality TechNet to Improve Pharmaceutical Product, Cost, and Success Rate Profiles</strong></p><p>In August, we announced a partnership with Vitality TechNet, an Orlando firm that helps pharmaceutical organizations digitize operations and securely use data effectively through the adoption of <a href="https://flur.ee/idmp/">IDMP semantic standards</a>. This partnership works to safely accelerate drug discovery processes and enables ongoing collaboration across internal departments, external partners, and regulatory offices through semantics, knowledge graphs, and digital trust technologies. This collaboration also allows for more efficient drug development processes and reduces the time and costs associated with late-stage failures. Read the whitepaper co-written by Vitality TechNet and Fluree <a href="https://flur.ee/idmp/">here</a>.</p><p>More details on the scope of this partnership are available <a href="https://app.livestorm.co/fluree/idmp-bridging-pharmaceutical-data-silos-with-semantics-ai-and-digital-trust-feat-vitality-technet">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*71b0O6ZmqVZRyngK.png" /></figure><h3>Fluree x Fabric</h3><p><strong>In 2023: Fluree partnered with Fabric.</strong></p><p><a href="https://fabricpay.com/">Fabric</a> uses Fluree technology to enable consumers to control and monetize their data without being exploited by companies like Meta and Google. With Fluree’s trusted ledger database, Fabric has built a business that seamlessly helps to improve the relationship between regular people and the brands appealing to them.</p><p>Fabric and Fluree both believe that by empowering data with additional security, integrity, and trust, the value of that data can be unlocked for consumers and enterprises alike. Read more about Fabric’s use of Fluree <a href="https://flur.ee/fluree-blog/fabric-reinstates-data-ownership-to-consumers-with-fluree/">here</a>.</p><h3>Fluree Board Expansion</h3><p><strong>In 2023: Fluree welcomed four industry experts to the company advisory board.</strong></p><p>In May we welcomed a team of technologists and strategic business advisors with unparalleled real-world experience that will work to build inroads for Fluree’s product-market fit as we continue to build partnerships and deliver outstanding data management solutions to our customers.</p><p>Julia Bardmesser, Robert Grossman, Sudeep Nadkarni, and Mark Silverman joined our advisory board alongside inaugural board member, Peter Serenita. Their combined industry knowledge and expertise has been invaluable as we continue to enhance the capabilities of Fluree and develop innovative data management solutions.</p><p>Learn more about the background and experience of each individual member <a href="https://www.businesswire.com/news/home/20230509005011/en">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ESVtP-VRYDo0_mCL.png" /></figure><h3>Fluree’s Favorite: Blog</h3><p><strong>So You Think You’re Ready for an Enterprise Knowledge Graph?</strong></p><p>Our favorite blog post from this year takes a deep dive into the world of enterprise knowledge graphs and their transformative potential for data integration and management. Knowledge graph use cases are growing rapidly, as the need to connect and integrate disparate data sources grows every day for organizations looking to play effective “data offense” and “data defense” simultaneously. With a knowledge graph, you can gain a unified view of your organization’s data and uncover insights and patterns that may have been hidden in disparate data sources.</p><p>If you missed its original publication, there’s no time like the present to learn more about enterprise knowledge graphs. Read more <a href="https://flur.ee/fluree-blog/so-you-think-youre-ready-for-an-enterprise-knowledge-graph/">here</a>.</p><h3>Fluree’s Favorite: Podcast</h3><p><strong>Episode #225 from the Data Futurology Podcast: The Future of Data Collaboration</strong></p><p>Topping the charts as our favorite podcast from 2023 is this episode of the Data Futurology Podcast featuring Fluree President, Eliud Polanco. Eliud joins host Felipe Flores to discuss the innovative blockchain-driven approach that Fluree brings to data management, emphasizing the promise of heightened security and efficiency for organizations looking to better manage data operations.</p><p>Eliud details Fluree’s decentralized approach to data management that allows automation to query data without compromising privacy, noting potential use cases in GDPR compliance and financial services. This episode of the Data Futurology podcast dives into all facets of Fluree technology and explores what the future of innovation in data management might look like. The full episode is available <a href="https://www.datafuturology.com/podcast/2023/3/1/225-the-future-of-data-collaboration">here</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*UnPhN67RLLFOTKwp.png" /></figure><h3>Fluree’s Favorite: Webinar</h3><p><strong>Data-Centric Strategies to Break the Cycle of Legacy Addiction</strong></p><p>In October we had an awesome conversation with Dave McComb, President of Semantic Arts, and Brian Platz, Fluree CEO, on “Data-Centric Strategies to Break the Cycle of Legacy Addiction.” The duo discussed practical steps organizations can take to break the cycle of legacy addiction and move to a more data-centric architecture, describing the path to data-centricity through a continuum: Legacy Amplification, Legacy Avoidance, Legacy Erosion, and Legacy Replacement.</p><p>This webinar, as a sequel to our first joint webinar with <a href="https://vimeo.com/857921554?share=copy">Semantic Arts</a>, explores strategies and tactics that will help organizations to asses their own data centric maturity, build technological and cultural paths to data-centric architectures, and dive in to the MMCE framework (Model, Map, Connect, Expand). View the full conversation <a href="https://app.livestorm.co/fluree/data-centric-strategies-to-break-the-cycle-of-legacy-addiction">here</a>.</p><h3>Conclusion</h3><p>Here at Fluree we are thankful for the incredible support from our clients and community. We feel fortunate to have had so many opportunities for growth, innovation, and learning. We’re proud of our ability to continue delivering innovative solutions to users old and new within the Fluree ecosystem.</p><p>We look forward to carrying the momentum from 2023 into this new year. Thank you for being part of the Fluree journey — here’s to a bright and successful future ahead!</p><p>Stay updated with future Fluree highlights and news by joining our <a href="https://mailchi.mp/flur.ee/subscribe">newsletter</a>!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=023bb59c2fc5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/fluree-2023-year-in-review-023bb59c2fc5">Fluree 2023 Year In Review</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Launching Fluree Version 3: The Next Generation JSON-LD Database]]></title>
            <link>https://medium.com/fluree/launching-fluree-version-3-the-next-generation-json-ld-database-22707b8c8773?source=rss----5f522fe3f91d---4</link>
            <guid isPermaLink="false">https://medium.com/p/22707b8c8773</guid>
            <category><![CDATA[linked-data]]></category>
            <category><![CDATA[data-collaboration]]></category>
            <category><![CDATA[json-ld]]></category>
            <category><![CDATA[immutability]]></category>
            <category><![CDATA[product-launch]]></category>
            <dc:creator><![CDATA[Malinda Massey]]></dc:creator>
            <pubDate>Thu, 30 Nov 2023 17:49:45 GMT</pubDate>
            <atom:updated>2023-11-30T17:49:45.864Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ad8_UfwbGJ_mOT719ECL3g.png" /></figure><blockquote>Fluree Version 3 is live in public preview, featuring native JSON-LD support and a cloud platform for management, development, and collaboration.</blockquote><p>Today marks an exciting milestone for us at Fluree. We’re thrilled to introduce <a href="http://flur.ee/fluree-v3">Fluree Version 3</a>, in conjunction with our hosted database platform in <a href="http://data.flur.ee/">public preview</a>. The journey to this day has been filled with anticipation, and we extend our heartfelt gratitude to the nearly 1000 testers from our cloud or JSON-LD beta programs that made this release possible.</p><h4><strong>What Makes Fluree Different?</strong></h4><p>Fluree was built from the ground up to enable secure and trusted data collaboration across many sources and applications. We’ve compiled the necessary capabilities to make data sharing across applications possible: interoperability, security, and trust.</p><ul><li>Direct Query Endpoint Access: Users can extract what they need, assured that our granular policy contracts will regulate data access.</li><li>Guaranteed Immutability and Traceability: Enjoy a graph database where each data alteration is cryptographically signed, ensuring digital trust.</li><li>Integrated Policy Contracts: Design, adjust, and administer access controls, letting your data “defend itself” throughout functional teams.</li><li>Shared Vocabularies: Enable universal data vocabularies and unique IRIs for seamless data integration across humans and machines</li><li>UI for Collaboration: Create data sets, connect APIs, and add collaborators with permissions and shared vocabularies built directly in.</li></ul><h4><strong>Embracing JSON-LD</strong></h4><p>Fluree isn’t just another graph database; it’s built on the JSON-LD standard. If you know JSON, diving into Fluree is a breeze. JSON-LD combines the simplicity of JSON with the power of linked graph data. With <a href="http://flur.ee/fluree-v3">Fluree Version 3,</a> you can now transact, query, and interact with your data using JSON-LD.</p><p>Our 2023 vision for Fluree encompassed:</p><ul><li>Embracing open standards.</li><li>Building a hosted database experience that caters to both the generic developer, the hardcore semantic engineer, and the business analysts</li><li>Pioneering a platform that truly enhances secure, trusted, and interoperable data collaboration.</li></ul><p>While Fluree has gained traction within the semantic knowledge community, our larger aspiration has always been to acquaint developers, engineers, business analysts, and the like with the immense potential of linked data. For those already comfortable with JSON, JSON-LD offers a straightforward yet powerful avenue into the realm of W3C-compliant RDF triples. It elegantly bridges the gap between traditional databases and the semantic web.</p><h4><strong>Engage with Fluree’s Public Preview</strong></h4><p>Fluree’s public preview isn’t just a test drive — it’s an invitation to help us bring the power of trusted linked data to the world. As we refine our roadmap, your feedback will be invaluable in aligning our product with real-world developer requirements. Fluree’s new cloud platform features an intuitive UI, empowering users to transact, query, model, and share data more efficiently than ever before.</p><p><a href="http://data.flur.ee/">Join the Fluree Public Preview</a></p><h4><strong>Start Building Today</strong></h4><p>We’re available for everyone today, so create your <a href="http://data.flur.ee/">free account</a>, read our <a href="http://docs.flur.ee/">docs</a>, and join our <a href="https://discord.com/invite/pgjsvPa9Nm">discord community</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=22707b8c8773" width="1" height="1" alt=""><hr><p><a href="https://medium.com/fluree/launching-fluree-version-3-the-next-generation-json-ld-database-22707b8c8773">Launching Fluree Version 3: The Next Generation JSON-LD Database</a> was originally published in <a href="https://medium.com/fluree">Fluree PBC</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>