{"id":2203,"date":"2020-01-20T02:56:40","date_gmt":"2020-01-19T23:56:40","guid":{"rendered":"http:\/\/kusuaks7\/?p=1808"},"modified":"2024-01-25T15:47:47","modified_gmt":"2024-01-25T15:47:47","slug":"beyond-machine-learning-capturing-cause-and-effect-relationships","status":"publish","type":"post","link":"https:\/\/www.experfy.com\/blog\/ai-ml\/beyond-machine-learning-capturing-cause-and-effect-relationships\/","title":{"rendered":"Beyond Machine Learning: Capturing Cause-and-Effect Relationships"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"2203\" class=\"elementor elementor-2203\" data-elementor-post-type=\"post\">\n\t\t\t\t\t\t<section class=\"has_eae_slider elementor-section elementor-top-section elementor-element elementor-element-757459a4 elementor-section-boxed elementor-section-height-default elementor-section-height-default\" data-id=\"757459a4\" data-element_type=\"section\" data-e-type=\"section\">\n\t\t\t\t\t\t<div class=\"elementor-container elementor-column-gap-default\">\n\t\t\t\t\t<div class=\"has_eae_slider elementor-column elementor-col-100 elementor-top-column elementor-element elementor-element-2859ff4f\" data-id=\"2859ff4f\" data-element_type=\"column\" data-e-type=\"column\">\n\t\t\t<div class=\"elementor-widget-wrap elementor-element-populated\">\n\t\t\t\t\t\t<div class=\"elementor-element elementor-element-36c20c6d elementor-widget elementor-widget-text-editor\" data-id=\"36c20c6d\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tArtificial intelligence is rapidly becoming one of the most important technologies of our era.\u00a0\u00a0Every day we can read about the latest AI advances from startups and large companies.\u00a0\u00a0AI technologies\u00a0 are approaching or surpassing human levels of performance in vision, speech recognition, language translation, and other human domains.\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Machine_learning\" rel=\"noopener\">Machine learning<\/a>\u00a0advances, like\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Deep_learning\" rel=\"noopener\">deep learning<\/a>, have played a central role in AI\u2019s recent achievements, giving computers the ability to\u00a0<a href=\"https:\/\/blog.irvingwb.com\/blog\/2017\/02\/machine-learning-and-knowledge-discovery.html\" rel=\"noopener\">be trained<\/a>\u00a0by ingesting and analyzing large amounts of data instead of being explicitly programmed.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-ae075c5 elementor-widget elementor-widget-text-editor\" data-id=\"ae075c5\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tDeep learning is a powerful statistical technique for classifying patterns using large training data sets and multi-layer\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Artificial_neural_network\" rel=\"noopener\">AI neural networks<\/a>.\u00a0 It\u2019s essentially a method for machines to learn from all kinds of data, whether structured or unstructured, that\u2019s loosely modeled on the way a\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Brain\" rel=\"noopener\">biological brain<\/a>\u00a0learns new capabilities.\u00a0\u00a0Each artificial neural unit is connected to many other such units, and the links can be statistically strengthened or decreased based on the data used to train the system.\u00a0\u00a0Each successive layer in a multi-layer network uses the output from the previous layer as input.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-42ddab3 elementor-widget elementor-widget-text-editor\" data-id=\"42ddab3\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tMachine learning can be applied to just about any domain of knowledge given our ability to gather valuable data in almost any area of interest.\u00a0\u00a0But, machine learning methods are significantly narrower and more specialized than humans.\u00a0\u00a0There are many tasks for which they\u2019re not effective given the current state-of-the-art.\u00a0\u00a0In an\u00a0<a href=\"http:\/\/science.sciencemag.org\/content\/358\/6370\/1530\" class=\"broken_link\" rel=\"noopener\">article<\/a>\u00a0recently published in\u00a0<a href=\"http:\/\/science.sciencemag.org\/\" class=\"broken_link\" rel=\"noopener\"><em>Science<\/em><\/a>, professors\u00a0Erik Brynjolfsson\u00a0and\u00a0<a href=\"https:\/\/www.cs.cmu.edu\/~tom\/\" rel=\"noopener\">Tom Mitchell<\/a>\u00a0identified the key criteria that help distinguish tasks that are particularly suitable for machine learning from those that are not.\u00a0\u00a0<a href=\"https:\/\/blog.irvingwb.com\/blog\/2018\/07\/what-can-machine-learning-do.html\" rel=\"noopener\">These include<\/a>:\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d17857d elementor-widget elementor-widget-text-editor\" data-id=\"d17857d\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t<ul>\n \t<li><strong>Tasks that map well-defined inputs to well-defined outputs<\/strong>, &#8211; e.g., labeling images of specific animals, the probability of cancer in medical record, the likelihood of defaulting on a loan application;<\/li>\n \t<li><strong>Large data sets exist or can be created containing such input-output pairs<\/strong>, &#8211; the bigger the training data sets the more accurate the learning;<\/li>\n \t<li><strong>The capability being learned should be relatively static,\u00a0<\/strong>&#8211; If the function changes rapidly, retraining is typically required, including the acquisition of new training data; and<\/li>\n \t<li><strong>No need for detailed explanation of how the decision was made, &#8211;\u00a0<\/strong>the methods behind a machine\u00a0\u00a0learning recommendation, &#8211; subtle adjustments to the numerical weights that interconnect its huge number of artificial neurons, &#8211; are difficult to explain because they\u2019re so different from those used by humans.<\/li>\n<\/ul>\n<a id=\"more\"><\/a>And, in particular, as we\u2019re frequently reminded,\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Correlation_does_not_imply_causation\" rel=\"noopener\"><em>correlation does not imply causation<\/em><\/a>.\u00a0\u00a0Machine learning is a\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Statistical_model\" rel=\"noopener\">statistical modelling<\/a>\u00a0technique, like\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Data_mining\" rel=\"noopener\">data mining<\/a>\u00a0and\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Business_analytics\" rel=\"noopener\">business analytics<\/a>, which finds and correlates patterns between inputs and outputs without necessarily capturing their cause-and-effect relationships.\u00a0 Determining causal relationships requires tried-and true\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_method\" rel=\"noopener\">scientific methods<\/a>,\u00a0that is, empirical and measurable evidence subject to testable explanations and predictions.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d051b5a elementor-widget elementor-widget-text-editor\" data-id=\"d051b5a\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tPhysics, biology and other natural sciences have long relied on\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Scientific_modelling\" rel=\"noopener\">scientific models and principles<\/a>\u00a0to understand and explain the cause-and-effect relationships that enable them to detect faint signals within large and\/or noisy data sets, &#8211; i.e., the proverbial\u00a0<em>needle in a haystack<\/em>.\u00a0\u00a0 For example, tracking\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Potentially_hazardous_object\" rel=\"noopener\">potentially hazardous<\/a>, fast moving,\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Near-Earth_object\" rel=\"noopener\">near-Earth objects<\/a>\u00a0is based on the detection of very small changes in the sky.\u00a0 Similarly, searching for Earth-size\u00a0extrasolar\u00a0planets is based on detecting the faint changes in a star\u2019s light caused by a potential planet quickly passing by.\u00a0\u00a0No matter how much data we might have access to, it would be near impossible to detect the weak, noisy signals associated with either task without the models developed over the past few hundred years indicating a potential entity of interest that should be further investigated.\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-048b6e0 elementor-widget elementor-widget-text-editor\" data-id=\"048b6e0\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tSuch scientific models have enabled the discovery of very short lived\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Elementary_particle\" rel=\"noopener\">elementary particles<\/a>, &#8211; like the\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Higgs_boson\" rel=\"noopener\">Higgs boson<\/a>, &#8211; amidst the huge amounts of data generated by high energy\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Particle_accelerator\" rel=\"noopener\">particle accelerators<\/a>.\u00a0\u00a0\u201cAccording to the casino rules of modern quantum physics, anything that can happen will happen eventually,\u201d explained a recent\u00a0<a href=\"https:\/\/www.nytimes.com\/interactive\/2018\/12\/21\/science\/cern-large-hadron-collider-ar-ul.html\" rel=\"noopener\">article<\/a>\u00a0on\u00a0CERN\u2019s\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Large_Hadron_Collider\" rel=\"noopener\">Large Hadron Collider<\/a>.\u00a0\u00a0 \u201cBefore a single proton is fired through the collider, computers have calculated all the possible outcomes of a collision according to known physics.\u00a0 Any unexpected bump in the real data at some energy could be a signal of unknown physics, a new particle.\u00a0\u00a0That was how the Higgs was discovered, emerging from the statistical noise in the autumn of 2011.\u201d\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-84983ca elementor-widget elementor-widget-text-editor\" data-id=\"84983ca\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tA few months ago,\u00a0<a href=\"https:\/\/www.gartner.com\/en\" class=\"broken_link\" rel=\"noopener\">Gartner<\/a>\u00a0published a report on\u00a0The Top 10 Strategic Technology Trends for 2019, that is, trends with the potential to impact and transform industries over the next 5 years.\u00a0\u00a0The report included three AI trends, noting that AI is opening up a new frontier for digital business, \u201cbecause virtually every application, service and Internet of Things (IoT) object incorporates an intelligent aspect to automate or augment application processes or human activities.\u201d\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-b847f47 elementor-widget elementor-widget-text-editor\" data-id=\"b847f47\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tThe\u00a0Augmented Developer\u00a0is one of Gartner\u2019s top AI trends.\u00a0\u00a0\u201cThe market is rapidly shifting from one in which professional data scientists must partner with application developers to create most AI-enhanced solutions to one in which professional developers can operate alone using predefined models delivered as a service. This provides the developer with an ecosystem of AI algorithms and models, as well as development tools tailored to integrating AI capabilities and models into a solution.\u201d\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-c63167f elementor-widget elementor-widget-text-editor\" data-id=\"c63167f\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\u201cSome AI services are complete models that a developer can simply call as a function, pass the appropriate parameters and data, and obtain a result.\u00a0\u00a0Others may be pretrained to a high level but require some additional data to complete the training\u2026 The advantage of these partially trained models is that they require much smaller datasets for training.\u00a0\u00a0Not only does the evolution of these AI platforms and suites of AI services enable a wider range of developers to deliver AI-enhanced solutions, but it also delivers much higher developer productivity.\u201d\n\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-78ccf40 elementor-widget elementor-widget-text-editor\" data-id=\"78ccf40\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tLet me discuss two examples of AI solutions that\u2019ve been enhanced by the inclusion of pre-defined models.\u00a0\u00a0The first deals with\u00a0<a href=\"https:\/\/blog.irvingwb.com\/blog\/2018\/09\/social-physics-making-ai-predictions-easily-accessible.html\" rel=\"noopener\">predicting human behaviors<\/a>, based on\u00a0<a href=\"https:\/\/news.mit.edu\/2014\/social-physics-0304\" rel=\"noopener\">research<\/a>\u00a0in MIT\u2019s\u00a0<a href=\"https:\/\/www.media.mit.edu\/groups\/human-dynamics\/overview\/\" rel=\"noopener\">Human Dynamics<\/a>\u00a0group led by Media Lab professor\u00a0<a href=\"https:\/\/www.media.mit.edu\/people\/sandy\/overview\/\" rel=\"noopener\">Alex (Sandy) Pentland<\/a>, &#8211; which is explained in detail in his 2014 book\u00a0<a href=\"https:\/\/www.amazon.com\/Social-Physics-Spread-Lessons-Science\/dp\/B00HUUO7NE\" rel=\"noopener\">Social Physics: How Good Ideas Spread<\/a>.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d889392 elementor-widget elementor-widget-text-editor\" data-id=\"d889392\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tData derived from human behavior is dynamic and ever-changing.\u00a0 Such messy data is difficult to analyze to make predictions, such as\u00a0who are our top customers and how do we acquire more of them?; and where should we open our next store?\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-10ebbee elementor-widget elementor-widget-text-editor\" data-id=\"10ebbee\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tAfter years of research,\u00a0Pentland\u2019s group discovered that all event-data representing human activity contain a special set of social activity patterns regardless of what the data is about.\u00a0\u00a0These patterns are common across all human activities and demographics, and can be used to detect emerging behavioral trends before they can be observed by any other technique.\u00a0 Detecting such fast-changing trends requires the ability to frequently analyze data sets collected over short periods of time looking for deviations from the patterns predicted by human behavior models.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d211803 elementor-widget elementor-widget-text-editor\" data-id=\"d211803\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\t\nWhy are there such universal human activity patterns?\u00a0 The answer likely lies in\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Human_evolution\" rel=\"noopener\">human evolution<\/a>.\u00a0\u00a0We are a\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Sociality\" rel=\"noopener\">social species<\/a>, with the drive\u00a0to learn\u00a0from others in our social group.\u00a0 Such social learning has helped us survive by adapting to drastically different environments, and has thus been reinforced by\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Natural_selection\" rel=\"noopener\">natural selection<\/a>.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-81bf308 elementor-widget elementor-widget-text-editor\" data-id=\"81bf308\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tThese social behavior patterns have been tested across a variety of applications involving people, including strategy formulation in business, economic activity in cities, and, &#8211; working with an intelligence agency, &#8211; the detection of\u00a0<a href=\"https:\/\/blog.irvingwb.com\/blog\/2018\/10\/social-physics-and-cybercrime.html\" rel=\"noopener\">potential terrorist activity<\/a>\u00a0based on Twitter data.\u00a0\u00a0As long as the data involves human activity, &#8211; regardless of the type of data, the demographic of the users or the size of the data sets, &#8211; similar behavioral dynamics apply.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-3834e2d elementor-widget elementor-widget-text-editor\" data-id=\"3834e2d\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tThe second example comes from\u00a0<a href=\"https:\/\/r4.ai\/\" rel=\"noopener\">r4 Technologies,<\/a>\u00a0an AI-based company whose advisory board I recently joined.\u00a0 Human-based organizations, like companies and industries, are generally based on common elements, processes, and relationship patterns, making it possible to develop fairly universal models of the way they function.\u00a0\u00a0Over the past decade, r4 has developed such generic, customizable models of business organizations and the industry sectors in which they operate, based on three key entities, &#8211; people, products and places, &#8211; and their various attributes and interrelationships.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-d650e0d elementor-widget elementor-widget-text-editor\" data-id=\"d650e0d\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tThe models are then customized for each specific company using its own internal data as well as a variety of external data sources, thus creating a unique\u00a0<em><a href=\"https:\/\/en.wikipedia.org\/wiki\/Digital_twin\" rel=\"noopener\">digital twin<\/a>\u00a0<\/em>simulation of the company and its\u00a0<a href=\"https:\/\/en.wikipedia.org\/wiki\/Market_environment\" rel=\"noopener\">market environment<\/a>, which\u00a0is continually updated as new data comes in.\u00a0 This enables the company to detect emerging business and market trends before they can be detected by statistical methods, helping the company make better, faster decisions.\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-6a477f0 elementor-widget elementor-widget-text-editor\" data-id=\"6a477f0\" data-element_type=\"widget\" data-e-type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t\t\tLet me conclude by summarizing the key benefits of AI solutions based on augmenting statistical methods with domain-based models:\n<ul>\n \t<li>they can be trained or customized with much smaller data sets;<\/li>\n \t<li>they can tolerate much more noise in the data;<\/li>\n \t<li>they can be continually updated with new data reflecting changing conditions;<\/li>\n \t<li>it\u2019s easier to explain how a decision or recommendation was arrived at; and,<\/li>\n \t<li>such augmented AI solutions help capture cause-and-effect relationships.<\/li>\n<\/ul>\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>Machine learning is a&nbsp;statistical modelling&nbsp;technique, like&nbsp;data mining&nbsp;and&nbsp;business analytics, which finds and correlates patterns between inputs and outputs without necessarily capturing their cause-and-effect relationships.&nbsp; Determining causal relationships requires tried-and true&nbsp;scientific methods,&nbsp;that is, empirical and measurable evidence subject to testable explanations and predictions. And, in particular, as we&rsquo;re frequently reminded,&nbsp;correlation does not imply causation. Here are the key benefits of AI solutions based on augmenting statistical methods with domain-based models.<\/p>\n","protected":false},"author":612,"featured_media":3393,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"content-type":"","footnotes":""},"categories":[183],"tags":[92],"ppma_author":[3320],"class_list":["post-2203","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-ml","tag-machine-learning"],"authors":[{"term_id":3320,"user_id":612,"is_guest":0,"slug":"irving-wladawsky-berger","display_name":"Irving Wladawsky-Berger","avatar_url":"https:\/\/secure.gravatar.com\/avatar\/?s=96&d=mm&r=g","user_url":"","last_name":"Wladawsky-Berger","first_name":"Irving","job_title":"","description":"Irving Wladawsky-Berger, a Guest Columnist at WSJ CIO Journal, is Research Affiliate at MIT Sloan School of Management, Adjunct Professor at Imperial College, London, and Chairman Advisory Board at r4 Technologies."}],"_links":{"self":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/2203","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/users\/612"}],"replies":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/comments?post=2203"}],"version-history":[{"count":6,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/2203\/revisions"}],"predecessor-version":[{"id":35663,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/posts\/2203\/revisions\/35663"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media\/3393"}],"wp:attachment":[{"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/media?parent=2203"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/categories?post=2203"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/tags?post=2203"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.experfy.com\/blog\/wp-json\/wp\/v2\/ppma_author?post=2203"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}