Expand Queries Using Semsets

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN">    <html lang="en">  <head> <title>Expand Queries Using Semsets</title> <link rel="alternate" type="application/rss+xml" title="" href="/resources/feeds/news.xml" />  <meta charset="utf-8"> <meta name="viewport" content="width=device-width, initial-scale=1.0"> <meta name="description" content=""> <meta name="author" content="">  <link rel="shortcut icon" href="/favicon.ico">  <link rel="stylesheet" href="/assets/plugins/bootstrap/css/bootstrap.min.css"> <link rel="stylesheet" href="/assets/css/style.css">  <link rel="stylesheet" href="/assets/plugins/line-icons/line-icons.css"> <link rel="stylesheet" href="/assets/plugins/font-awesome/css/font-awesome.min.css">  <link rel="stylesheet" href="/assets/css/theme-colors/blue.css"> <link rel="stylesheet" href="/assets/css/footers/footer-v1.css"> <link rel="stylesheet" href="/assets/css/headers/header-v1.css">  <link rel="stylesheet" href="/css/custom.css"> <meta property="og:site_name" content="KBpedia"/> <meta property="og:title" content="Machine Learning Use Cases: Expand Queries Using Semsets"> <meta property="og:type" content="article"/> <meta property="og:description" content="We explain the importance of semsets to capture how we refer to things "> <meta property="og:image" content="/imgs/kbpedia-logo-350.png"> <meta property="og:image:width" content="350"> <meta property="og:image:height" content="390"> </head> <body> <div class="wrapper"> <div class="header-v1">  <div class="topbar"> <div class="container"> <div class="col-md-1"> </div> <a class="navbar-brand" href="/"> <img id="logo-header" src="/imgs/kbpedia-logo-420-horz.png" height="75" alt="KBpedia Knowledge Structure" name="logo-header"> </a> </div> </div>  <div class="navbar navbar-default mega-menu" role="navigation"> <div class="container">  <div class="navbar-header"> <button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-responsive-collapse"> <span class="sr-only">Toggle navigation</span> </button> </div> <div style="clear:both; height: 1px;">   </div>  <div class="col-md-1">   </div> <div class="collapse navbar-collapse navbar-responsive-collapse col-md-10"> <ul class="nav navbar-nav pull-left">  <li> <a href="/">Home</a> </li>  <li> <a href="/knowledge-graph/">knowledge Graph</a> </li> <li> <a href="http://sparql.kbpedia.org/">SPARQL</a> </li>  <li class="dropdown"> <a href="/background/">Background</a> <ul class="dropdown-menu"> <li> <a href="/background/overview/">Overview</a> </li> <li> <a href="/background/features-and-benefits/">Features & Benefits</a> </li> <li> <a href="/background/data-and-knowledge-structures/">Data and Knowledge Structures</a> </li> <li> <a href="/background/technology/">Technology</a> </li> <li> <a href="/background/machine-learning/">Machine Learning</a> </li> <li> <a href="/background/uses/">Uses</a> </li> </ul> </li> <li class="dropdown"> <a href="/use-cases/">Use Cases</a> <ul class="dropdown-menu"> <li class="dropdown-submenu"> <a href="/use-cases/knowledge-graph/">Knowledge Graph (KG)</a> <ul class="dropdown-menu"> <li><a href="/use-cases/browse-the-knowledge-graph/">Browse the Knowledge Graph</a></li> <li><a href="/use-cases/search-the-knowledge-graph/">Search the Knowledge Graph</a></li> <li><a href="/use-cases/expand-queries-using-semsets/">Expand Queries Using Semsets</a></li> <li><a href="/use-cases/use-and-control-of-inferencing/">Uses and Control of Inferencing</a></li> </ul> </li> <li class="dropdown-submenu"> <a href="/use-cases/machine-learning-use-case/">Machine Learning (KBAI)</a> <ul class="dropdown-menu"> <li><a href="/use-cases/text-classification-using-esa-and-svm/">Create Supervised Learning Training Sets</a></li> <li><a href="/use-cases/document-specific-word2vec-training-corpuses/">Create Word Embedding Corpuses</a></li> <li><a href="/use-cases/extending-kbpedia-with-kbpedia-categories/">Create Graph Embedding Corpuses</a></li> <li><a href="/use-cases/text-classification-using-esa-and-svm/">Classify Text</a></li> <li><a href="/use-cases/dynamic-machine-learning/">Create 'Gold Standards' for Tuning Learners</a></li> <li><a href="/use-cases/disambiguating-kbpedia-knowledge-graph-concepts/">Disambiguate KG Concepts</a></li> <li><a href="/use-cases/dynamic-machine-learning/">Dynamic Machine Learning Using the KG</a></li> </ul> </li> <li class="dropdown-submenu"> <a href="/use-cases/mapping-use-case/">Mapping</a> <ul class="dropdown-menu"> <li><a href="/use-cases/mapping-external-data-and-schema/">Map Concepts</a></li> <li><a href="/use-cases/extending-kbpedia-with-kbpedia-categories/">Extend KBpedia with Wikipedia</a></li> <li><a href="/use-cases/benefits-from-extending-kbpedia-with-private-datasets/">Extend KBpedia for Domains</a></li> <li><a href="/use-cases/mapping-external-data-and-schema/">General Use of the Mapper</a></li> </ul> </li> </ul> </li> <li class="dropdown"> <a href="/resources/">Resources</a> <ul class="dropdown-menu"> <li><a href="/resources/downloads/">Download KBpedia</a></li> <li><a href="/resources/about/">About KBpedia</a></li> <li><a href="/resources/faq/">KBpedia FAQ</a></li> <li><a href="/resources/news/">News About KBpedia</a></li> <li><a href="/resources/statistics/">KBpedia Statistics</a></li> <li><a href="/resources/documentation/">Additional Documentation</a></li> <li><a href="/resources/support/">Support for KBpedia</a>.</li> </ul> </li> </ul> </div> <div class="col-md-1">   </div> </div> </div> </div>  <div class="breadcrumbs"> <div class="container"> <div class="col-md-1">   </div> <div class="col-md-10"> <h1 class="pull-left"></h1> <ul class="pull-right breadcrumb"> <li>Use Cases</li> <li class="active">Expand Queries Using Semsets</li> </ul> </div> <div class="col-md-1">   </div> </div> </div>    <div class="container content"> <div class="row"> <div class="col-md-2">   </div> <div class="col-md-8"> <div class="use-cases-header"> <table border="0" cellpadding="4" cellspacing="2"> <tbody> <tr> <td colspan="2" align="center"> <h2> <b>USE CASE</b> </h2> </td> </tr> <tr> <td style="width: 140px;" valign="top"> <b>Title:</b> </td> <td style="padding-left: 25px;" valign="top"> <span style="font-weight: bold;">Expand Queries Using Semsets</span> </td> </tr> <tr> <td valign="top"> <b>Short Description:</b> </td> <td style="padding-left: 25px;" valign="top"> We explain the importance of <span style="font-style: italic;">semsets</span> to capture how we refer to things </td> </tr> <tr> <td valign="top"> <b>Problem:</b> </td> <td style="padding-left: 25px;" valign="top"> Concepts, entities and relations are not known by a single label, agreed by all. Thus, in order to properly represent the <span style="font-style: italic;">idea of</span> or <span style="font-style: italic;">meaning</span> of these things, multiple labels must be assembled to accommodate the various ways for how they are referenced. </td> </tr> <tr> <td valign="top"> <b>Approach:</b> </td> <td style="padding-left: 25px;" valign="top"> We use best practices to help populate alternative labels for concepts, entities and relations. <br> <ul> <li>The initial best practice is to be attentive to gathering alternate labels, including sources of abbreviations, acronyms, argot, diminutives, epithets, idioms, jargon, lingo, misspellings, nicknames, pseudonyms, redirects, and slang, as well as, of course, synonyms. </li> <li>Second, it is extremely important when mapping new knowledge sources into an existing target knowledge graph to also harvest and map semset candidates. Sources like Wikipedia are rich repositories of semsets; </li> <li>Third, like all textual additions to a knowledge graph, it is important to include the language label for the specific language being used. In the baseline case of KBpedia, the specific baseline language is English (en). By tagging text fields with their appropriate language tag, it is possible to readily swap out labels in one language for another. This design approach leads to a better ability to represent the conceptual and entity nature of the knowledge graph in multiple natural languages; and, </li> <li>Fourth, in actual enterprise implementations, it is important to design and include workflow steps that enable subject matter experts to add new <code>altLabel</code> entries as encountered. Keeping semsets robust and up-to-date is an essential means for knowledge graphs to fulfill their purpose as knowledge structures that represent "things not strings." </li> </ul> </td> </tr> <tr> <td valign="top"> <b>Key Findings:</b> </td> <td style="padding-left: 25px;" valign="top"> <ul> <li>Expanded semsets provide better referencing about things, breaking the tyranny of strings </li> <li>Leverage online and electronic resources for semset alternative labels </li> <li>Rich harvest pools exist for gathering alternative labels </li> <li>Proper treatment of labels helps enable multi-linguality. </li> </ul> </td> </tr> </tbody> </table> </div> </div> <div class="col-md-2">   </div> </div> <div class="row"> </div> <div class="row"> </div> <div class="row"> <div class="col-md-2">   </div> <div class="col-md-8"> <p> One of the strongest contributions that semantic technologies make to knowledge-based artificial intelligence (KBAI) is to focus on what things mean, as opposed to how they are labeled. This focus on underlying meaning is captured by the phrase, "things not strings." </p> <p> The <span style="font-style: italic;">idea</span> of something -- that is, its <span style="font-style: italic;">meaning</span> -- is conveyed by how we define that something, the context in which the various tokens (terms) for that something is used, and in the variety of terms or labels we apply to that thing. A single label alone is not enough. The idea of a parrot is conveyed by our understanding of what the name parrot means. Yet, in languages other than English, the same idea of parrot may be conveyed by the terms Papagei, <span id="result_box" class="short_text" lang="fr"><span class="">perroquet</span></span>, <span id="result_box" class="short_text" lang="es"><span class="">loro</span></span>, попугай, or オウム, depending on our native language. </p> <p> The idea of the '<span style="font-style: italic;">United States</span>', even just in English, may be conveyed with labels ranging from <span style="font-style: italic;">America</span>, to <span style="font-style: italic;">US</span>, <span style="font-style: italic;">USA</span>, <span style="font-style: italic;">Amerika</span>, <span style="font-style: italic;">Uncle Sam</span>, or even the <span style="font-style: italic;">Great Satan</span>. As another example, the simple token '<span style="font-style: italic;">bank</span>' can mean a financial institution, a side of a river, turning an airplane, tending a fire, or a pool shot, depending on context. What these examples illustrate is that a single term is more often not the only way to refer to something, and a given token may mean vastly different things depending on the context in which it is used. </p> <p> Knowledge graphs are also not comprised of labels, but of concepts, entities and the relationships between those things. Knowledge graphs constructed from single labels for individual nodes and single labels for different individual relations are, therefore, unable to capture these nuances of context and varieties of reference. In order for a knowledge graph to be useful to a range of actors it must reflect the languages and labels meaningful to those actors. In order for us to be able to distinguish the accurate references of individual terms we need the multiple senses of terms to each be associated with its related concepts, and then to use the graph relationships for those concepts to help disambiguate the intended meaning of the term based on its context of use. </p> <p> In the lexical database of <a href="https://en.wikipedia.org/wiki/WordNet">WordNet</a>, the variety of terms by which a given term might be known is collectively called a <a href="https://en.wikipedia.org/wiki/Synonym_ring"><span style="font-style: italic;">synset</span></a>. According to WordNet, a <span style="font-style: italic;">synset</span> (short for <span style="font-style: italic;">synonym set</span>) is "defined as a set of one or more synonyms that are interchangeable in some context without changing the truth value of the proposition in which they are embedded." In our view, the concept of a synset is helpful, but still does not go far enough. Any name or label that draws attention to a given thing can provide the same referential power as a synonym. We can include in this category abbreviations, acronyms, argot, diminutives, epithets, idioms, jargon, lingo, misspellings, nicknames, pseudonyms, redirects, and slang, as well as, of course, synonyms. Collectively, we call any of these terms to refer to a given concept or entity a <span style="font-style: italic;">semset</span>. In all cases, these terms are mere pointers to the actual something at hand. </p> <p> In the KBpedia knowledge graph, these terms are defined either as <code>skos:prefLabel</code> (the preferred term), <code>skos:altLabel</code> (all other semset variants) or <code>skos:hiddenLabel</code> (misspellings). In this use case, we show an example of semsets in use, discuss what we have done specifically in KBpedia to accommodate semsets, and summarize with some best practices for semsets in knowledge graph construction. </p> <h3> A KBpedia Example </h3> <p> Let's see how this concept of semset works in KBpedia. Though in practice much of what is done with KBpedia is done via SPARQL queries or programmatically, here we will simply use the <a href="/demo">online KBpedia demo</a>. Our example Web page references an earlier announcement for the release of KBpedia v. 1.40. We begin by entering the URL for this announcement into the Analyze Web Page demo box on the main page: </p> <center> <a href="altLabel_0.png"> <img style="border: 0px solid ; width: 650px; margin-bottom: 20px;" alt="Running the KBpedia Demo" title="Running the KBpedia Demo" src="altLabel_0.png" align="middle"></a> </center> <p> After some quick messages on the screen telling us how the Web page is being processed, we receive the first results page for the analysis of our KBpedia version announcement. The tab we are looking at here highlights the matching concepts we have found, with the most prevalent shown in red-orange. Note that one of those main concepts is a 'knowledge graph': </p> <center> <a href="altLabel_1.png"> <img style="border: 0px solid ; width: 650px; margin-bottom: 20px;" alt="Concept Tagging in the Web Page" title="Concept Tagging in the Web Page" src="altLabel_1.png" align="middle"></a> </center> <p> If we mouseover this 'knowledge graph' tag we see a little popup window that shows us what KBpedia concepts the string token matches. In this case, there is only one concept match, that of OntologyInformationScience; the term 'knowledge graph' itself is not a listed match (other highlighted terms may present multiple possible matches): </p> <center> <a href="altLabel_2.png"> <img style="border: 0px solid ; width: 650px; margin-bottom: 20px;" alt="Tag Links to KBpedia" title="Tag Links to KBpedia" src="altLabel_2.png" align="middle"></a> </center> <p> When we click on the live link to <a href="/knowledge-graph/reference-concept/?uri=OntologyInformationScience"> <code>OntologyInformationScience</code></a> we are then taken to that concept's individual entry within the KBpedia knowledge structure: </p> <center> <a href="altLabel_3.png"> <img style="border: 0px solid ; width: 650px; margin-bottom: 20px;" alt="'Knowledge Graph' in the Semset" title="'Knowledge Graph' in the Semset" src="altLabel_3.png" align="middle"></a> </center> <p> Other use cases describe more fully <a href="/use-cases/browse-the-knowledge-graph/">how to browse</a> and navigate KBpedia or <a href="/use-cases/search-the-knowledge-graph/">how to search</a> it. To summarize here, what we are looking at above is the standard concept header that presents the preferred label for the concept, followed by its URI and then its alternative labels (semset). Note that 'knowledge graph' is one of the alternative terms for <a href="/knowledge-graph/reference-concept/?uri=OntologyInformationScience"> <code>OntologyInformationScience</code></a>. </p> <p> We can also confirm that only one concept is associated with the 'knowledge graph' term by searching for it. Note, as well, that we have the chance to also search individual fields such as the title (preferred label), alternative labels (semset), URI or definitions of the concept: </p> <center> <a href="altLabel_4.png"> <img style="border: 0px solid ; width: 650px; margin-bottom: 20px;" alt="Searching altLabels Alone" title="Searching altLabels Alone" src="altLabel_4.png" align="middle"></a> </center> <p> What this example shows is that a term, 'knowledge graph', in our original Web page, while not having a concept dedicated to that specific label, has a corresponding concept of <a href="/knowledge-graph/reference-concept/?uri=OntologyInformationScience"> <code>OntologyInformationScience</code></a> as provided through its semset of <code>altLabels</code>. Semsets provide us an additional term pool by which we can refer to a given concept or entity. </p> <h3> Implementing Semsets in KBpedia </h3> <p> For years now, given that our knowledge graphs are grounded in semantic technologies, we have emphasized the generous use of <code>altLabels</code> to provide more contextual bases for how we can identify and get to appropriate concepts. In our prior public release of KBpedia (namely, v 120), we had more than 100,000 <code>altLabels</code> for the approximately 39,000 reference concepts in the system (an average of 2.6 <code>altLabels</code> per concept). </p> <p> As part of the reciprocal mapping effort we undertook in moving from version 1.20 to version 1.40 of KBpedia, as we describe in its <a href="/use-cases/extending-kbpedia-with-kbpedia-categories/"> own use case</a>, we also made a concerted effort to mine the alternative labels within Wikipedia. (The reciprocal mapping effort, you may recall, involved adding missing nodes and structure in Wikipedia that did not have corresponding tie-in points in the existing KBpedia.) Through these efforts, we were able to add more than a quarter million new alternative terms to KBpedia. Now, the average number of <code>altLabels</code> per RC exceeds 6.6, representing a <span class="double_u">2.5X increase</span> over the prior version. This is the kind of effort that enables us to match to 'knowledge graph'. </p> <h3> Best Practices for Semsets in Knowledge Graphs </h3> <p> At least three best-practice implications arise from these semset efforts: </p> <ul> <li>First, it is extremely important when mapping new knowledge sources into an existing target knowledge graph to also harvest and map semset candidates. Sources like Wikipedia are rich repositories of semsets; </li> <li>Second, like all textual additions to a knowledge graph, it is important to include the language label for the specific language being used. In the baseline case of KBpedia, the specific baseline language is English (en). By tagging text fields with their appropriate language tag, it is possible to readily swap out labels in one language for another. This design approach leads to a better ability to represent the conceptual and entity nature of the knowledge graph in multiple natural languages; and, </li> <li>Third, in actual enterprise implementations, it is important to design and include workflow steps that enable subject matter experts to add new <code>altLabel</code> entries as encountered. Keeping semsets robust and up-to-date is an essential means for knowledge graphs to fulfill their purpose as knowledge structures that represent "things not strings." </li> </ul> <p> Semsets, along with other semantic technology techniques, are essential tools in constructing <span style="font-style: italic; font-weight: bold;">meaningful</span> knowledge graphs. </p> </div> <div class="col-md-2">   </div> </div> </div>  <div class="footer-v1"> <div class="footer"> <div class="container"> <div class="row">  <div class="col-md-3 md-margin-bottom-40"> <table> <tbody><tr> <td> <a href="/"> <img id="logo-footer" class="footer-logo" src="/imgs/logo-simple-purple.png" alt="" name="logo-footer"> </a> </td> </tr> <tr> <td> <center> <p> KBpedia </p> </center> </td> </tr> </tbody></table> <p style="font-size: 0.85em;"> KBpedia exploits large-scale knowledge bases and semantic technologies for machine learning, data interoperability and mapping, and fact extraction and tagging. </p> </div>   <div class="col-md-3 md-margin-bottom-40"> <div class="posts"> <div class="headline"> <h2> Latest News </h2> </div> <ul class="list-unstyled latest-list">  <li> <a href="https://kbpedia.org/resources/news/kbpedia-adds-ecommerce/">KBpedia Adds Major eCommerce Capabilities</a> <small>06/15/2020</small> </li>  <li> <a href="http://kbpedia.org/resources/news/kbpedia-continues-quality-improvements/">KBpedia Continues Quality Improvements</a> <small>12/04/2019</small> </li>  <li> <a href="http://kbpedia.org/resources/news/wikidata-coverage-nearly-complete/">Wikidata Coverage Nearly Complete (98%)</a> <small>04/08/2019</small> </li> </ul> </div> </div> <div class="col-md-3 md-margin-bottom-40"> <div class="headline"> <h2> Other Resources </h2> </div> <ul class="list-unstyled link-list"> <li> <a href="/resources/about/">About</a> </li> <li> <a href="/resources/faq/">FAQ</a> </li> <li> <a href="/resources/news/">News</a> </li> <li> <a href="/use-cases/">Use Cases</a> </li> <li> <a href="/resources/documentation/">Documentation</a> </li> <li> <a href="/resources/privacy/">Privacy</a> </li> <li> <a href="/resources/terms-of-use/">Terms of Use</a> </li> </ul> </div>  <div class="col-md-3 map-img md-margin-bottom-40"> <div class="headline"> <h2> Contact Us </h2> </div> <address class="md-margin-bottom-40"> c/o <a href="mailto:info@mkbergman.com?subject=KBpedia%20Inquiry">Michael K. Bergman</a> <br> 380 Knowling Drive <br> Coralville, IA 52241 <br> U.S.A. <br> Voice: +1 319 621 5225 </address> </div>  </div> </div> </div> <div class="copyright"> <div class="container"> <div class="row"> <div class="col-md-7"> <p class="copyright" style="font-size: 10px;"> 2016-2022 © <a href="http://kbpedia.org" style="font-size: 10px;">Michael K. Bergman.</a> All Rights Reserved. </p> </div>  <div class="col-md-5"> <ul class="footer-socials list-inline"> <li> <a href="/resources/feeds/news.xml" class="tooltips" data-toggle="tooltip" data-placement="top" title="" data-original-title="RSS feed"> <i class="fa fa-rss-square"></i> </a> <br></li> <li> <a href="http://github.com/Cognonto" class="tooltips" data-toggle="tooltip" data-placement="top" title="" data-original-title="Github"> <i class="fa fa-github"></i> </a> <br></li> <li> <a href="http://twitter.com/cognonto" class="tooltips" data-toggle="tooltip" data-placement="top" title="" data-original-title="Twitter"> <i class="fa fa-twitter"></i> </a> <br></li> </ul> </div>  </div> </div> </div> </div>   <script type="text/javascript" src="/assets/plugins/jquery/jquery.min.js"></script> <script type="text/javascript" src="/assets/plugins/jquery/jquery-migrate.min.js"></script> <script type="text/javascript" src="/assets/plugins/bootstrap/js/bootstrap.min.js"></script>  <script type="text/javascript" src="/assets/plugins/back-to-top.js"></script>  <script type="text/javascript" src="/assets/js/custom.js"></script>  <script type="text/javascript" src="/assets/js/app.js"></script>  <script type="text/javascript" src="/assets/plugins/smoothScroll.js"></script> <script type="text/javascript" src="/assets/plugins/owl-carousel/owl-carousel/owl.carousel.js"></script> <script type="text/javascript" src="/assets/plugins/layer-slider/layerslider/js/greensock.js"></script> <script type="text/javascript" src="/assets/plugins/layer-slider/layerslider/js/layerslider.transitions.js"></script> <script type="text/javascript" src="/assets/plugins/layer-slider/layerslider/js/layerslider.kreaturamedia.jquery.js"></script>  <script type="text/javascript" src="/assets/js/custom.js"></script>  <script type="text/javascript" src="/assets/js/plugins/layer-slider.js"></script> <script type="text/javascript" src="/assets/js/plugins/style-switcher.js"></script> <script type="text/javascript" src="/assets/js/plugins/owl-carousel.js"></script> <script type="text/javascript" src="/assets/js/plugins/owl-recent-works.js"></script> <script type="text/javascript"> jQuery(document).ready(function() { App.init(); LayerSlider.initLayerSlider(); StyleSwitcher.initStyleSwitcher(); OwlCarousel.initOwlCarousel(); OwlRecentWorks.initOwlRecentWorksV2(); }); </script>   <script> (function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){ (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o), m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m) })(window,document,'script','https://www.google-analytics.com/analytics.js','ga'); ga('create', 'UA-84405507-1', 'auto'); ga('send', 'pageview'); </script>

CINXE.COM

Expand Queries Using Semsets