{"id":776320,"date":"2021-09-23T06:00:00","date_gmt":"2021-09-23T13:00:00","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?p=776320"},"modified":"2023-03-14T21:09:43","modified_gmt":"2023-03-15T04:09:43","slug":"real-world-evidence-and-the-path-from-data-to-impact","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/research\/blog\/real-world-evidence-and-the-path-from-data-to-impact\/","title":{"rendered":"Real-world evidence and the path from data to impact"},"content":{"rendered":"\n<div class=\"annotations \" data-bi-aN=\"margin-callout\">\n\t<ul class=\"annotations__list card depth-16 bg-body p-4 annotations__list--left\">\n\t\t<li class=\"annotations__list-item\">\n\t\t\t\t\t\t<span class=\"annotations__type d-block text-uppercase font-weight-semibold text-neutral-300 small\">Group<\/span>\n\t\t\t<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/group\/societal-resilience\/\" target=\"_self\" class=\"annotations__link font-weight-semibold text-decoration-none\" data-bi-type=\"annotated-link\" aria-label=\"Societal Resilience\" data-bi-aN=\"margin-callout\" data-bi-cN=\"Societal Resilience\">\n\t\t\t\tSocietal Resilience&nbsp;<span class=\"glyph-append glyph-append-chevron-right glyph-append-xsmall\"><\/span>\n\t\t\t<\/a>\n\t\t\t\t\t<\/li>\n\t<\/ul>\n<\/div>\n\n\n\n<p>From the&nbsp;intense&nbsp;shock of the&nbsp;COVID-19 pandemic&nbsp;to the effects of climate change, our global society has never faced greater risk.&nbsp;The&nbsp;<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/group\/societal-resilience\/\" target=\"_blank\" rel=\"noreferrer noopener\">Societal Resilience<\/a> team at Microsoft Research was&nbsp;established&nbsp;in recognition of this risk and tasked with&nbsp;developing&nbsp;open technologies that enable&nbsp;a&nbsp;<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/group\/societal-resilience\/#!guiding-principles\" target=\"_blank\" rel=\"noreferrer noopener\">scalable response<\/a> in times of crisis.&nbsp;And just as we think about scalability in a holistic way\u2014scaling across different forms of&nbsp;common&nbsp;problems, for different partners, in different domains\u2014we also&nbsp;take a multi-horizon view of what it means to respond to crisis.<\/p>\n\n\n\n<p>When&nbsp;an acute crisis strikes,&nbsp;it creates&nbsp;an urgency&nbsp;to help real people, right now. However, not all crises are acute, and not all&nbsp;forms of&nbsp;response&nbsp;deliver&nbsp;direct assistance.&nbsp;While we need to attend to foreground&nbsp;crises&nbsp;like floods, fire, and famine, we also need to pay attention to the background crises that precipitate them\u2014for many, the background crisis is&nbsp;already&nbsp;the foreground of their lives. To&nbsp;give&nbsp;an&nbsp;example, with&nbsp;climate&nbsp;change, the potential long-term casualty is the human race.&nbsp;But&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.un.org\/sustainabledevelopment\/blog\/2019\/09\/climate-change-and-migration-in-vulnerable-countries\/\" target=\"_blank\" rel=\"noreferrer noopener\">climate migration<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;is happening all over the world&nbsp;already, and it disproportionately affects&nbsp;some of&nbsp;the poorest&nbsp;and most vulnerable countries.&nbsp;<\/p>\n\n\n\n<p>Crises can also feed into and amplify one another. For example,&nbsp;the United Nations\u2019 International Organization for&nbsp;Migration&nbsp;(IOM)&nbsp;reports that&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/publications.iom.int\/system\/files\/pdf\/migrant_vulnerability_to_human_trafficking_and_exploitation.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">migration in general<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>,&nbsp;and&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/publications.iom.int\/system\/files\/pdf\/addressing_human_trafficking_dec2015.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">crisis events<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;in particular,&nbsp;are key drivers of&nbsp;human trafficking&nbsp;and exploitation.&nbsp;Migration push factors can become exacerbated&nbsp;during times of crisis,&nbsp;and people may&nbsp;face&nbsp;extreme vulnerability&nbsp;when&nbsp;forced to migrate amid a <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/publications.iom.int\/system\/files\/pdf\/migrant_vulnerability_to_human_trafficking_and_exploitation.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">lack of safe and regular migration pathways<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>. Human exploitation and trafficking are a breach of the&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.un.org\/en\/about-us\/universal-declaration-of-human-rights\" target=\"_blank\" rel=\"noreferrer noopener\">most fundamental human rights<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;and show what can happen when societies fail to prevent the emergence of systemic vulnerability within their populations. By tackling&nbsp;existing&nbsp;sources of vulnerability and exploitation now, we&nbsp;can&nbsp;learn how to&nbsp;deliver more effective responses to the interconnected crises of the future.<br>&nbsp;<br>To build resilience in these areas, researchers at Microsoft and their collaborators have been working on a number of tools&nbsp;that help domain experts&nbsp;translate&nbsp;real-world&nbsp;data into evidence.&nbsp;All three tools and case studies presented in this post share a common idea: that&nbsp;a&nbsp;hidden structure&nbsp;exists&nbsp;within the many combinations of attributes that constitute real-world data, and that both domain knowledge and data tools are needed to make sense of this structure and inform real-world response.&nbsp;To learn more about these&nbsp;efforts, read the accompanying <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/aka.ms\/AAdcshy\" target=\"_blank\" rel=\"noreferrer noopener\">AI for Business and Technology blog post<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>.&nbsp;Note that several&nbsp;of the&nbsp;technologies&nbsp;in this post&nbsp;will be presented in greater detail at the&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/researchsummit.microsoft.com\/\" target=\"_blank\" rel=\"noreferrer noopener\">Microsoft Research Summit<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;on October 19\u201321, 2021.&nbsp;<\/p>\n\n\n\n<div style=\"padding-bottom:0; padding-top:0\" class=\"wp-block-msr-immersive-section alignfull row has-background has-light-gray-background-color wp-block-msr-immersive-section\">\n\t\n\t<div class=\"container\">\n\t\t<div class=\"wp-block-msr-immersive-section__inner wp-block-msr-immersive-section__inner--narrow\">\n\t\t\t<div class=\"wp-block-media-text has-vertical-margin-small  has-vertical-padding-none  is-stacked-on-mobile is-vertically-aligned-center\" style=\"grid-template-columns:58% auto\" data-bi-an=\"media-text\"><figure class=\"wp-block-media-text__media\"><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/researchsummit.microsoft.com\/\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"535\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-1024x535.png\" alt=\"Promo for the Microsoft Research Summit on October 19-21, 2021\" class=\"wp-image-776809 size-full\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-1024x535.png 1024w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-300x157.png 300w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-768x401.png 768w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-1536x803.png 1536w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-2048x1070.png 2048w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1200x627_Research_Summit_ad1_with_logo_collage-240x125.png 240w\" sizes=\"(max-width: 1024px) 100vw, 1024px\" \/><\/a><\/figure><div class=\"wp-block-media-text__content\" data-bi-an=\"media-text\">\n<h3 id=\"microsoft-research-summit\">Microsoft Research Summit<\/h3>\n\n\n\n<p><strong>October 19\u201321, 2021<\/strong><\/p>\n\n\n\n<p>At this inaugural event, researchers and engineers across Microsoft, and our colleagues in academia, industry, and government will come together to discuss cutting-edge work that is pushing the limits of science and technology.&nbsp;<\/p>\n\n\n\n<div class=\"wp-block-buttons is-layout-flex wp-block-buttons-is-layout-flex\">\n<div class=\"wp-block-button is-style-fill-chevron\"><a data-bi-type=\"button\" class=\"wp-block-button__link wp-element-button\" href=\"https:\/\/researchsummit.microsoft.com\/home_public?OCID=msr-researchsummit_Blog_CustPromo\" target=\"_blank\" rel=\"noreferrer noopener\">Register<\/a><\/div>\n<\/div>\n<\/div><\/div>\t\t<\/div>\n\t<\/div>\n\n\t<\/div>\n\n\n\n<h2 id=\"supporting-evidence-based-policy\">Supporting&nbsp;evidence-based policy<\/h2>\n\n\n\n<p>For crisis response at the level above&nbsp;individual assistance, we need to&nbsp;think in terms of policy\u2014how should we allocate people, money, and other resources towards&nbsp;tackling both the causes and consequences of the crisis?&nbsp;<\/p>\n\n\n\n<p>In such situations,&nbsp;we need&nbsp;evidence<em>&nbsp;<\/em>that can inform new policies and&nbsp;evaluate existing ones,&nbsp;whether&nbsp;the&nbsp;public policy of governments or&nbsp;the private policy&nbsp;of organizations.&nbsp;Returning to the link between&nbsp;crises and trafficking, if policy makers do not have access to&nbsp;supporting&nbsp;evidence&nbsp;because it&nbsp;doesn\u2019t exist&nbsp;or cannot be shared, or if they are not persuaded by the weight of evidence in support of&nbsp;the&nbsp;causal relationship,&nbsp;they will not enact policies that ensure appropriate&nbsp;intervention and direct assistance when the time comes.&nbsp;<\/p>\n\n\n\n<p>Policy is the greatest lever we have to save lives and livelihoods at scale. Building technology for evidence-based policy&nbsp;is&nbsp;how we maximize our leverage as we work to&nbsp;make&nbsp;societies&nbsp;more resilient.&nbsp;<\/p>\n\n\n\n<h2 id=\"developing-real-world-evidence\">Developing real-world evidence<\/h2>\n\n\n\n<p>Real-world problems affecting societal resilience&nbsp;leave a trail of&nbsp;\u201creal-world data\u201d&nbsp;(RWD)&nbsp;in their wake. This concept originated in the medical field to differentiate observational&nbsp;data collected for some other purpose (for example, electronic health records and&nbsp;healthcare&nbsp;claims)&nbsp;from&nbsp;experimental&nbsp;data collected through, and for the&nbsp;specific&nbsp;purpose of,&nbsp;a&nbsp;randomized controlled event (like a&nbsp;clinical&nbsp;trial).&nbsp;<\/p>\n\n\n\n<p>The corresponding notion of \u201creal-world evidence\u201d&nbsp;(RWE)&nbsp;similarly emerged in the medical field, defined&nbsp;in&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.law.cornell.edu\/uscode\/text\/21\/355g\" target=\"_blank\" rel=\"noreferrer noopener\">21 U.S. Code \u00a7 355g<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;of the&nbsp;Federal&nbsp;Food, Drug, and Cosmetic&nbsp;Act&nbsp;as \u201cdata regarding the usage, or the potential benefits or risks, of a drug derived from sources other than traditional clinical trials.\u201d While our&nbsp;RWE research is partly inspired by&nbsp;the&nbsp;methods used to&nbsp;derive&nbsp;RWE from RWD in a medical context, we also take a broader view&nbsp;of what&nbsp;counts as&nbsp;evidence&nbsp;for decision making&nbsp;and policy making across unrelated fields.&nbsp;<\/p>\n\n\n\n<p>For&nbsp;problems like human trafficking, for example, it&nbsp;would be unethical to&nbsp;run&nbsp;a randomized controlled trial in which&nbsp;trafficking is&nbsp;allowed to happen.&nbsp;In&nbsp;this&nbsp;case, observational data&nbsp;describing victims of trafficking, collected at the point of assistance, is the&nbsp;next best source of data.&nbsp;Indeed, this kind of positive feedback loop,&nbsp;with&nbsp;direct assistance activities&nbsp;informing evidence-based policy&nbsp;and evidence-based policy&nbsp;informing the allocation of assistance resources, is one of the main&nbsp;ways&nbsp;in which&nbsp;targeted&nbsp;technology&nbsp;development&nbsp;could&nbsp;make a significant difference to real-world outcomes.&nbsp;&nbsp;<\/p>\n\n\n\n<h2 id=\"empowering-domain-experts\">Empowering&nbsp;domain experts<\/h2>\n\n\n\n<p>In practice, however,&nbsp;facilitating&nbsp;positive feedback between assistance and policy activities&nbsp;means dealing with multiple&nbsp;challenges that&nbsp;hinder&nbsp;the progression from data to evidence,&nbsp;to policy,&nbsp;to impact.&nbsp;The people and organizations collecting data&nbsp;on the front line&nbsp;are rarely&nbsp;those&nbsp;responsible for making&nbsp;or evaluating the impact of&nbsp;policy,&nbsp;just as&nbsp;those with the technical expertise to develop evidence are rarely those with the domain expertise needed to interpret and act on&nbsp;that evidence.&nbsp;<\/p>\n\n\n\n<p>To bridge&nbsp;these gaps, we&nbsp;work&nbsp;with&nbsp;domain experts&nbsp;to design&nbsp;tools that democratize the practice of evidence development\u2014reducing reliance on data scientists&nbsp;and other data specialists whose skills&nbsp;are in short supply, especially&nbsp;during a&nbsp;crisis.&nbsp;<\/p>\n\n\n\n<h2 id=\"real-world-evidence-in-action\">Real-world&nbsp;evidence in action<\/h2>\n\n\n\n<p>Over the following sections, we&nbsp;describe tools for developing different kinds of&nbsp;real-world evidence&nbsp;in response to&nbsp;the distinctive characteristics\u2014and challenges\u2014of&nbsp;accessing,&nbsp;analyzing, and acting on&nbsp;real-world data.&nbsp;In each case, we use examples drawn from our&nbsp;efforts to&nbsp;counter&nbsp;human trafficking and modern slavery.<\/p>\n\n\n\n<h3 id=\"developing-evidence-of-correlation-from-private-data\">Developing evidence&nbsp;of correlation&nbsp;from&nbsp;private&nbsp;data<\/h3>\n\n\n\n<h4 id=\"research-challenge\">Research&nbsp;challenge<\/h4>\n\n\n\n<p>When&nbsp;people&nbsp;can\u2019t&nbsp;see&nbsp;the&nbsp;data&nbsp;describing&nbsp;a&nbsp;phenomenon,&nbsp;they&nbsp;can\u2019t make effective policy decisions&nbsp;at any level.&nbsp;However,&nbsp;many&nbsp;real-world datasets&nbsp;relate to&nbsp;individuals and cannot be shared&nbsp;with other organizations because of privacy concerns and data protection regulations.&nbsp;<\/p>\n\n\n\n\n\n<p>This challenge arose&nbsp;when&nbsp;Microsoft&nbsp;participated&nbsp;in&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/techagainsttrafficking.org\/\" target=\"_blank\" rel=\"noreferrer noopener\">Tech Against Trafficking (TAT)<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>\u2014a coalition of technology companies&nbsp;(currently&nbsp;Amazon, BT, Microsoft, and Salesforce) working to combat trafficking with technology. In the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/techagainsttrafficking.org\/accelerating-toward-data-insights-tech-against-trafficking-successfully-concludes-its-pilot-accelerator\/\" target=\"_blank\" rel=\"noreferrer noopener\">2019 TAT Accelerator Program<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, TAT member companies&nbsp;worked together to support the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.ctdatacollaborative.org\/\" target=\"_blank\" rel=\"noreferrer noopener\">Counter Trafficking Data Collaborative (CTDC)<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>\u2014an initiative run by&nbsp;the International Organization for Migration (IOM)&nbsp;that&nbsp;pools data from organizations including, IOM, Polaris, Liberty Shared,&nbsp;OTSH, and A21, to create the world\u2019s largest database&nbsp;of individual survivors&nbsp;of trafficking.&nbsp;&nbsp;<\/p>\n\n\n\n<p>The CTDC data hub makes derivatives of this data openly available&nbsp;as a way of informing evidence-based policy&nbsp;against human trafficking,&nbsp;through&nbsp;data maps, dashboards, and stories that are accessible to&nbsp;policy makers.&nbsp;This raises&nbsp;risks to privacy.&nbsp;For example,&nbsp;if traffickers believe they have identified a victim within published data artifacts, they may assume that this implies collaboration with the authorities in ways that may prompt retaliation. To get around this, CTDC&nbsp;data&nbsp;is&nbsp;de-identified and anonymized using standard approaches. But this is cumbersome,&nbsp;forces a sacrifice of&nbsp;the data\u2019s&nbsp;analytic utility, and&nbsp;may not remove all&nbsp;residual risks to privacy and safety.&nbsp;<\/p>\n\n\n\n\n\n<h4 id=\"research-question\">Research question<\/h4>\n\n\n\n<p>How can we enable policy makers in one organization to view and explore the&nbsp;private&nbsp;data collected and controlled by another in a way that preserves the privacy of&nbsp;groups of&nbsp;data subjects,&nbsp;preserves the utility of datasets,&nbsp;and is accessible to all data stakeholders?&nbsp;<\/p>\n\n\n\n<h4 id=\"enabling-technology\">Enabling&nbsp;technology<\/h4>\n\n\n\n<div class=\"annotations \" data-bi-aN=\"margin-callout\">\n\t<ul class=\"annotations__list card depth-16 bg-body p-4 annotations__list--left\">\n\t\t<li class=\"annotations__list-item\">\n\t\t\t\t\t\t<span class=\"annotations__type d-block text-uppercase font-weight-semibold text-neutral-300 small\">Download<\/span>\n\t\t\t<a href=\"https:\/\/github.com\/microsoft\/synthetic-data-showcase\" target=\"_self\" class=\"annotations__link font-weight-semibold text-decoration-none\" data-bi-type=\"annotated-link\" aria-label=\"Synthetic data showcase\" data-bi-aN=\"margin-callout\" data-bi-cN=\"Synthetic data showcase\">\n\t\t\t\tSynthetic data showcase&nbsp;<span class=\"glyph-append glyph-append-share glyph-append-xsmall\"><\/span>\n\t\t\t<\/a>\n\t\t\t\t\t<\/li>\n\t<\/ul>\n<\/div>\n\n\n\n<p>We developed the concept of a&nbsp;Synthetic Data Showcase&nbsp;as a new mechanism for privacy-preserving data release, now available on&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/microsoft\/synthetic-data-showcase\" target=\"_blank\" rel=\"noreferrer noopener\">GitHub<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;and as an&nbsp;interactive&nbsp;<a href=\"https:\/\/www.microsoft.com\/en-us\/ai\/ai-lab-synthetic-data-showcase\" target=\"_blank\" rel=\"noreferrer noopener\">AI Lab<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>.&nbsp;Synthetic data&nbsp;is&nbsp;generated&nbsp;in a way that reproduces the structure and statistics of a sensitive dataset, but with the guarantee that&nbsp;every combination of attributes in the records appears&nbsp;at least&nbsp;<em>k&nbsp;<\/em>times in the records of the sensitive dataset&nbsp;and therefore cannot be used to isolate any actual groups of individuals smaller than&nbsp;<em>k<\/em>.&nbsp;In other words, we use synthetic data to generalize <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/dataprivacylab.org\/dataprivacy\/projects\/kanonymity\/index3.html\" target=\"_blank\" rel=\"noreferrer noopener\"><em>k<\/em>-anonymity<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> to all attributes of a dataset\u2014not just&nbsp;a subset of&nbsp;attributes&nbsp;determined in advance to be&nbsp;identifying in&nbsp;combination.&nbsp;&nbsp;<\/p>\n\n\n\n<p>Alongside the synthetic data, we also release&nbsp;aggregate data on all short combinations of attributes, to both validate the utility of the synthetic data and&nbsp;to&nbsp;retrieve&nbsp;actual&nbsp;counts (as a&nbsp;multiple of&nbsp;<em>k<\/em>)&nbsp;for official reporting. Finally, we combine both&nbsp;anonymous datasets in an automatically&nbsp;generated Power BI report for an interactive, visual, and accessible form of data&nbsp;exploration.&nbsp;The resulting evidence is at the level of correlation\u2014both across data attributes, as reflected by their joint counts, and across datasets, as reflected by the&nbsp;similarity of counts calculated over the sensitive versus synthetic datasets.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><a data-bi-bhvr=\"14\"  data-bi-cn=\"Interactive dashboard in Power BI showing visual representations of the counts of human trafficking cases with different attributes: year of registration, gender, age, type of labor exploitation, type of sexual exploitation, citizenship, and country of exploitation. Light green frequency bars to the left represent counts of attributes dynamically generated by Power BI. These light green frequency bars are repeated on the right, paired with dark green frequency bars showing actual counts for comparison. The age range 9\u201317 is selected, corresponding to 18,740 estimated cases. This is identical to the actual count shown on the right. The most frequent type of labor exploitation associated with this age range is \u201cother.\u201d\" href=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"2252\" height=\"1277\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final.jpg\" alt=\"Interactive dashboard in Power BI showing visual representations of the counts of human trafficking cases with different attributes: year of registration, gender, age, type of labor exploitation, type of sexual exploitation, citizenship, and country of exploitation. Light green frequency bars to the left represent counts of attributes dynamically generated by Power BI. These light green frequency bars are repeated on the right, paired with dark green frequency bars showing actual counts for comparison. The age range 9\u201317 is selected, corresponding to 18,740 estimated cases. This is identical to the actual count shown on the right. The most frequent type of labor exploitation associated with this age range is \u201cother.\u201d\" class=\"wp-image-776431\" title=\"Use of Synthetic Data Showcase to create and explore the CTDC Global Human Trafficking Synthetic Dataset\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final.jpg 2252w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-300x170.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-1024x581.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-768x435.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-1536x871.jpg 1536w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-2048x1161.jpg 2048w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-343x193.jpg 343w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig1_Global-HumanTraffick_final-240x136.jpg 240w\" sizes=\"(max-width: 2252px) 100vw, 2252px\" \/><\/a><figcaption class=\"wp-element-caption\">In this example, we use Power BI to support privacy-preserving exploration of the anonymous datasets generated by our Synthetic Data Showcase tool. Having selected the records of victims in the age range 9\u201317, we can see the distributions of multiple additional attributes contained in these records: the year the victim was registered, gender, country of citizenship and exploitation, and type of labor or sexual exploitation. All of the counts in these distributions are dynamically generated by Power BI filtering and aggregating records of the synthetic dataset. These &#8220;estimated&#8221; counts are compared on the right with &#8220;actual&#8221; counts precomputed over the sensitive data, showing that the synthetic dataset accurately captures the structure of the sensitive data for the selected age range. For these victims aged 9\u201317, the association with &#8220;typeOfLabourOther&#8221; indicates a potential need to expand the data schema to support more targeted policy design tackling forced labor of children.<\/figcaption><\/figure>\n\n\n\n\n\n<div class=\"annotations \" data-bi-aN=\"margin-callout\">\n\t<ul class=\"annotations__list card depth-16 bg-body p-4 annotations__list--left\">\n\t\t<li class=\"annotations__list-item\">\n\t\t\t\t\t\t<span class=\"annotations__type d-block text-uppercase font-weight-semibold text-neutral-300 small\">Download<\/span>\n\t\t\t<a href=\"https:\/\/www.ctdatacollaborative.org\/global-synthetic-dataset\" target=\"_self\" class=\"annotations__link font-weight-semibold text-decoration-none\" data-bi-type=\"annotated-link\" aria-label=\"Global Synthetic Dataset\" data-bi-aN=\"margin-callout\" data-bi-cN=\"Global Synthetic Dataset\">\n\t\t\t\tGlobal Synthetic Dataset&nbsp;<span class=\"glyph-append glyph-append-share glyph-append-xsmall\"><\/span>\n\t\t\t<\/a>\n\t\t\t\t\t<\/li>\n\t<\/ul>\n<\/div>\n\n\n\n<p>Our ability to&nbsp;publish and collaborate on open software&nbsp;enables&nbsp;us&nbsp;to&nbsp;work with IOM on creating a&nbsp;Synthetic Data Showcase&nbsp;for their full, de-identified victim database&nbsp;without ever accessing&nbsp;the highly sensitive data ourselves. Today, IOM has&nbsp;announced the&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.ctdatacollaborative.org\/global-synthetic-dataset\" target=\"_blank\" rel=\"noreferrer noopener\">resulting update<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;to&nbsp;the&nbsp;CTDC&nbsp;website,&nbsp;sharing data&nbsp;on more than three times as many victims as before.&nbsp;This includes several new data columns, with group-level privacy guarantees and utility that anyone can interactively verify.<\/p>\n\n\n\n<p>With the new&nbsp;Global Human Trafficking Synthetic Dataset,&nbsp;Synthetic Data Showcase&nbsp;has&nbsp;enabled&nbsp;IOM and CTDC to share data that couldn\u2019t otherwise be shared, helping&nbsp;address problems that couldn\u2019t otherwise be&nbsp;solved. In the following sections, we show how this dataset can be used to develop additional types of evidence to fight trafficking.&nbsp;<\/p>\n\n\n\n<div class=\"annotations \" data-bi-aN=\"margin-callout\">\n\t<ul class=\"annotations__list card depth-16 bg-body p-4 annotations__list--right\">\n\t\t<li class=\"annotations__list-item\">\n\t\t\t\t\t\t<span class=\"annotations__type d-block text-uppercase font-weight-semibold text-neutral-300 small\">Download<\/span>\n\t\t\t<a href=\"https:\/\/github.com\/UNMigration\/HTCDS\" target=\"_self\" class=\"annotations__link font-weight-semibold text-decoration-none\" data-bi-type=\"annotated-link\" aria-label=\"Human Trafficking Case Data Standard (HTCDS)\" data-bi-aN=\"margin-callout\" data-bi-cN=\"Human Trafficking Case Data Standard (HTCDS)\">\n\t\t\t\tHuman Trafficking Case Data Standard (HTCDS)&nbsp;<span class=\"glyph-append glyph-append-share glyph-append-xsmall\"><\/span>\n\t\t\t<\/a>\n\t\t\t\t\t<\/li>\n\t<\/ul>\n<\/div>\n\n\n\n<p>IOM aims to share the new technique with counter-trafficking organizations worldwide as part of a wider program to improve the production of data and evidence on human trafficking. This includes establishing new international standards and guidance&nbsp;to support governments in producing high-quality administrative data,&nbsp;in partnership with the UN Office on Drugs and Crime, and a&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/UNMigration\/HTCDS\" target=\"_blank\" rel=\"noreferrer noopener\">package of data standards and information management tools<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;for frontline counter-trafficking agencies.<\/p>\n\n\n\n\n\n<p>Further&nbsp;details&nbsp;of the&nbsp;TAT-CTDC-IOM-Microsoft collaboration:&nbsp;<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/group\/societal-resilience\/articles\/case-study-tech-against-trafficking\/\" target=\"_blank\" rel=\"noreferrer noopener\">Societal Resilience case study<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/www.microsoft.com\/en-us\/research\/publication\/design-of-a-privacy-preserving-data-platform-for-collaboration-against-human-trafficking\/\" target=\"_blank\" rel=\"noreferrer noopener\">Design of a Privacy-Preserving Data Platform for Collaboration Against Human Trafficking<\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/republicans-science.house.gov\/sites\/republicans.science.house.gov\/files\/2020-07-28%20Testimony%20Darnton.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">TAT congressional testimony<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/aka.ms\/modernslaveryandhumantrafficking\" target=\"_blank\" rel=\"noreferrer noopener\">Microsoft Modern Slavery&nbsp;and Human Trafficking&nbsp;Statement<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n<\/ul>\n\n\n\n\n\n<h3 id=\"developing-evidence-of-causation-from-observational-data\">Developing&nbsp;evidence&nbsp;of causation&nbsp;from observational data<\/h3>\n\n\n\n<h4 id=\"research-challenge\">Research&nbsp;challenge<\/h4>\n\n\n\n<p>If&nbsp;people&nbsp;can\u2019t&nbsp;see&nbsp;the&nbsp;causation&nbsp;driving&nbsp;a&nbsp;phenomenon,&nbsp;they&nbsp;can\u2019t effectively make strategic policy decisions about where to invest&nbsp;resources&nbsp;over the long-term.&nbsp;However, counts and correlations derived from data&nbsp;cannot by themselves be used to&nbsp;confirm&nbsp;the&nbsp;presence&nbsp;of&nbsp;a&nbsp;causal&nbsp;relationship&nbsp;within&nbsp;a domain, or estimate the size of the causal effect.<\/p>\n\n\n\n\n\n<p>The COVID-19 pandemic&nbsp;triggered&nbsp;an&nbsp;unprecedented effort to&nbsp;identify&nbsp;existing drugs that may reduce mortality and other&nbsp;adverse outcomes of infection. Through the Microsoft Research&nbsp;<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/collaboration\/studies-in-pandemic-preparedness\/\" target=\"_blank\" rel=\"noreferrer noopener\">Studies in Pandemic Preparedness<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> program, scientists at&nbsp;Johns Hopkins University&nbsp;and Stanford have developed <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.frontiersin.org\/articles\/10.3389\/fphar.2021.700776\/full\" target=\"_blank\" rel=\"noreferrer noopener\">new guidelines<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> for performing retrospective analysis of pharmacoepidemiological data in a manner that emulates, to the extent possible, a randomized controlled trial.&nbsp;This&nbsp;capability is&nbsp;valuable whenever it isn\u2019t&nbsp;possible, affordable, or ethical to run a trial&nbsp;for a given treatment.&nbsp;&nbsp;<\/p>\n\n\n\n<p>Microsoft&nbsp;Research&nbsp;also has&nbsp;world-leading&nbsp;experts&nbsp;studying&nbsp;the kind of&nbsp;causal inference&nbsp;needed for trial emulation, as well as the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/Microsoft\/dowhy\" target=\"_blank\" rel=\"noreferrer noopener\">DoWhy<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> and <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/Microsoft\/EconML\" target=\"_blank\" rel=\"noreferrer noopener\">EconML<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> libraries with which to&nbsp;perform it.&nbsp;However, such guidelines and libraries remain inaccessible to&nbsp;experts in other&nbsp;domains&nbsp;who lack&nbsp;expertise in data science and causal inference. This includes people working on anti-trafficking who seek to understand&nbsp;the causal effect of factors,&nbsp;like migration and crises, on the extent and type of trafficking in order to&nbsp;formulate a more effective policy response.<\/p>\n\n\n\n\n\n<h4 id=\"research-question\">Research question<\/h4>\n\n\n\n<p>How can we empower domain experts&nbsp;to answer causal questions&nbsp;using observational data collected for some other purpose&nbsp;in a way that&nbsp;emulates&nbsp;a randomized controlled trial,&nbsp;controls for the inherent biases of the data collection process,&nbsp;and doesn\u2019t require&nbsp;prior&nbsp;expertise in data science or causal inference?&nbsp;<\/p>\n\n\n\n<h4 id=\"enabling-technology\">Enabling&nbsp;technology<\/h4>\n\n\n\n<p>Building on the simplified&nbsp;and&nbsp;structured approach to causal inference&nbsp;promoted by&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/Microsoft\/dowhy\" target=\"_blank\" rel=\"noreferrer noopener\">DoWhy<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, we have developed ShowWhy:&nbsp;an interactive application for guided causal inference over observational data.&nbsp;ShowWhy&nbsp;assumes no prior familiarity with&nbsp;coding or&nbsp;causal inference, yet&nbsp;enables the user to:&nbsp;&nbsp;<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>formulate a causal question<\/li>\n\n\n\n<li>capture&nbsp;relevant&nbsp;domain knowledge<\/li>\n\n\n\n<li>derive corresponding data variables<\/li>\n\n\n\n<li>produce&nbsp;and defend estimates&nbsp;of&nbsp;the&nbsp;average&nbsp;causal effect<\/li>\n<\/ul>\n\n\n\n<p>Behind the scenes, ShowWhy uses&nbsp;a combination of approaches from <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/Microsoft\/dowhy\" target=\"_blank\" rel=\"noreferrer noopener\">DoWhy<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/Microsoft\/EconML\" target=\"_blank\" rel=\"noreferrer noopener\">EconML<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, and <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/uber\/causalml\" target=\"_blank\" rel=\"noreferrer noopener\">CausalML<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> to&nbsp;perform causal inference in Python,&nbsp;while also&nbsp;eliciting the&nbsp;assumptions, decisions, and justifications&nbsp;needed&nbsp;for others to evaluate the standard of evidence&nbsp;represented by the results.&nbsp;Following an analysis, users can export Jupyter Notebooks and other reports that document the end-to-end process&nbsp;in forms suitable for audiences ranging from data scientists&nbsp;evaluating the analysis to decision makers evaluating the appropriate policy response.<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><a data-bi-bhvr=\"14\"  data-bi-cn=\"A user interface for an application titled ShowWhy, with the headline, \u201cFor identified victims of trafficking, does recent natural disaster cause severity of control to increase?\u201d. The page is divided into three horizontal panes: a workflow outline on the left, showing which stages are done and which are still to do, a guidance pane in the center offering guidance about the currently selected workflow step, and a workspace pane to the right where the user completes the current workflow tasks. The workspace shows a causal graph connecting the exposure to the outcome it is hypothesized to cause, along with two kinds of controls: confounders with arrows connecting to both exposure and outcome, and outcome determinants connected to outcome only. A displayed message confirms that it is possible to estimate the causal effect given the overall structure of the causal graph.\" href=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"2253\" height=\"1335\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final.jpg\" alt=\"A user interface for an application titled ShowWhy, with the headline, \u201cFor identified victims of trafficking, does recent natural disaster cause severity of control to increase?\u201d. The page is divided into three horizontal panes: a workflow outline on the left, showing which stages are done and which are still to do, a guidance pane in the center offering guidance about the currently selected workflow step, and a workspace pane to the right where the user completes the current workflow tasks. The workspace shows a causal graph connecting the exposure to the outcome it is hypothesized to cause, along with two kinds of controls: confounders with arrows connecting to both exposure and outcome, and outcome determinants connected to outcome only. A displayed message confirms that it is possible to estimate the causal effect given the overall structure of the causal graph.\" class=\"wp-image-776434\" title=\"Use of ShowWhy to evaluate causal claims about the impact of disasters on human trafficking\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final.jpg 2253w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final-300x178.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final-1024x607.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final-768x455.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final-1536x910.jpg 1536w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final-2048x1214.jpg 2048w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SRBlog_Fig2-ShowWhy_final-240x142.jpg 240w\" sizes=\"(max-width: 2253px) 100vw, 2253px\" \/><span class=\"sr-only\"> (opens in new tab)<\/span><\/a><figcaption class=\"wp-element-caption\">In this example, we use the ShowWhy application to evaluate the causal claim that disasters increase the severity of control experienced by trafficking victims. A prior theory of labor coercion proposes that such severity increases with the lack of alternative options (<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" data-bi-bhvr=\"14\"  data-bi-cn=\"A user interface for an application titled ShowWhy, with the headline, \u201cFor identified victims of trafficking, does recent natural disaster cause severity of control to increase?\u201d. The page is divided into three horizontal panes: a workflow outline on the left, showing which stages are done and which are still to do, a guidance pane in the center offering guidance about the currently selected workflow step, and a workspace pane to the right where the user completes the current workflow tasks. The workspace shows a causal graph connecting the exposure to the outcome it is hypothesized to cause, along with two kinds of controls: confounders with arrows connecting to both exposure and outcome, and outcome determinants connected to outcome only. A displayed message confirms that it is possible to estimate the causal effect given the overall structure of the causal graph.\" href=\"https:\/\/onlinelibrary.wiley.com\/doi\/abs\/10.3982\/ECTA8963\" target=\"_blank\" rel=\"noreferrer noopener\">Acemoglu&nbsp;and&nbsp;Wolitzky, 2011<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>), and crisis situations could rapidly remove all such options available to affected individuals. Starting with the CTDC Global Human Trafficking Synthetic Dataset as real-world data on the experiences of individual victims, we can use ShowWhy to model the causal structure of the domain. This includes identifying potential confounders that must be controlled for, i.e., factors assumed to have a causal influence on both the exposure and outcome. One such example is the &#8220;rule of law&#8221; in the affected country, which may exacerbate the human consequences of any natural disaster. By using a variety of models, together with a variety of question definitions and statistical estimators, ShowWhy can produce a range of causal estimates for each combination of assumptions, offering a broad base of evidence for decision making and policy response.<\/figcaption><\/figure>\n\n\n\n<p>One of the most challenging aspects of causal inference&nbsp;is&nbsp;knowing&nbsp;which of multiple reasonable decisions to make at each step of the process. This includes how to define the population, exposure, and outcome of interest, how to model the causal structure of the domain,&nbsp;which estimation approach to use, and so on.&nbsp;Nonexperts&nbsp;deal with&nbsp;even greater uncertainty&nbsp;about whether the final result may hinge on some arbitrary decision,&nbsp;like the precise value of a threshold or the contents of a&nbsp;query.&nbsp;<\/p>\n\n\n\n<p>To&nbsp;address&nbsp;this uncertainty&nbsp;and&nbsp;counter any claims of selective reporting&nbsp;in support of a preferred hypothesis, ShowWhy&nbsp;enables&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.nature.com\/articles\/s41562-020-0912-z\" target=\"_blank\" rel=\"noreferrer noopener\">specification curve analysis<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, in which all reasonable specifications of the&nbsp;causal&nbsp;inference task can&nbsp;be estimated, refuted, and jointly analyzed for&nbsp;significance.&nbsp;While a single&nbsp;estimate of the&nbsp;causal&nbsp;effect can inspire both overconfidence and&nbsp;underconfidence\u2014depending on the prior beliefs of the audience\u2014ShowWhy promotes a more balanced discussion&nbsp;about the overall&nbsp;strength&nbsp;(and contingency)&nbsp;of&nbsp;a&nbsp;much broader&nbsp;body of&nbsp;evidence.&nbsp;ShowWhy shifts the&nbsp;focus&nbsp;to&nbsp;where it matters: from a theoretical debate about the validity of&nbsp;a given result to&nbsp;a practical&nbsp;debate about&nbsp;the&nbsp;validity&nbsp;of&nbsp;decisions that&nbsp;make a meaningful difference to such results&nbsp;in practice.<\/p>\n\n\n\n\n\n<p>ShowWhy will be released open-source on GitHub&nbsp;in late 2021.&nbsp;While this tool&nbsp;can be used to answer a wide range of causal questions across domains, we are particularly interested in how it can help our partners understand the drivers of human trafficking and exploitation.&nbsp;<\/p>\n\n\n\n<p>Just as ShowWhy&nbsp;aims to make&nbsp;the practice of&nbsp;causal&nbsp;inference accessible to many for the first time, the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.ctdatacollaborative.org\/global-synthetic-dataset\" target=\"_blank\" rel=\"noreferrer noopener\">Global Human Trafficking Synthetic Dataset<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> similarly makes available, for the first time, rich microdata describing all&nbsp;victims of trafficking&nbsp;identified by CTDC data contributors.&nbsp;At the&nbsp;upcoming <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/researchsummit.microsoft.com\/\" target=\"_blank\" rel=\"noreferrer noopener\">Microsoft Research Summit<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, October 19\u201321, 2021,&nbsp;we\u2019ll&nbsp;show how this combination of resources can enable community-generated evidence in support of, for example, the&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/publications.iom.int\/system\/files\/pdf\/addressing_human_trafficking_dec2015.pdf\" target=\"_blank\" rel=\"noreferrer noopener\">link between crises and&nbsp;trafficking<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;reported by IOM.&nbsp;&nbsp;<\/p>\n\n\n\n<p>While such&nbsp;evidence generated from&nbsp;synthetic data is suggestive rather than conclusive, it allows the community to conceive and evaluate causal questions that would otherwise be&nbsp;inconceivable.&nbsp;And because ShowWhy enables simple reproduction of analyses over alternative datasets,&nbsp;organizations like IOM&nbsp;that control highly sensitive data&nbsp;can&nbsp;easily&nbsp;use&nbsp;their own data to validate any external analyses performed using synthetic data.<\/p>\n\n\n\n\n\n<p>Partner use&nbsp;of retrospective&nbsp;pharmacoepidemiological&nbsp;analysis in support of off-label drug discovery in the fight against COVID-19 and&nbsp;other respiratory infections:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2021.637647\/full\" target=\"_blank\" rel=\"noreferrer noopener\">The Association Between Alpha-1 Adrenergic Receptor Antagonists and In-Hospital Mortality From COVID-19<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.medrxiv.org\/content\/10.1101\/2021.04.08.21255148v1\" target=\"_blank\" rel=\"noreferrer noopener\">COVID-19 outcomes among&nbsp;hospitalized&nbsp;men with or without exposure to alpha-1-adrenergic receptor blocking agents<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/elifesciences.org\/articles\/61700\" target=\"_blank\" rel=\"noreferrer noopener\">Alpha-1 adrenergic receptor antagonists to prevent hyperinflammation and death from lower respiratory tract infection<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.jci.org\/articles\/view\/139642\" target=\"_blank\" rel=\"noreferrer noopener\">Preventing cytokine storm syndrome in COVID-19 using \u03b1-1 adrenergic receptor antagonists<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n<\/ul>\n\n\n\n\n\n<h3 id=\"developing-evidence-of-change-from-temporal-data\">Developing&nbsp;evidence&nbsp;of change&nbsp;from&nbsp;temporal&nbsp;data<\/h3>\n\n\n\n<h4 id=\"research-challenge\">Research&nbsp;challenge<\/h4>\n\n\n\n<p>If&nbsp;people&nbsp;can\u2019t see the&nbsp;structure&nbsp;and dynamics&nbsp;of&nbsp;a&nbsp;phenomenon,&nbsp;they&nbsp;can\u2019t make effective policy&nbsp;decisions&nbsp;at the&nbsp;tactical&nbsp;level.&nbsp;However, in situations with substantial variability in how data observations&nbsp;occur&nbsp;over time, it can be&nbsp;difficult to separate meaningful changes&nbsp;from&nbsp;the&nbsp;noise.&nbsp;<\/p>\n\n\n\n\n\n<p>This challenge arose&nbsp;in&nbsp;the&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/techagainsttrafficking.org\/accelerating-the-use-of-technology-to-combat-human-trafficking\/\" target=\"_blank\" rel=\"noreferrer noopener\">2021 TAT Accelerator Program<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>,&nbsp;currently in progress, with <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.unseenuk.org\/\" target=\"_blank\" rel=\"noreferrer noopener\">Unseen UK<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;and&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.seattleagainstslavery.org\/\" target=\"_blank\" rel=\"noreferrer noopener\">Seattle Against Slavery<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;as participating organizations. For Unseen,&nbsp;one major&nbsp;challenge&nbsp;is&nbsp;identifying hidden patterns and emerging trends&nbsp;within case records generated&nbsp;through calls&nbsp;to the UK Modern Slavery and Exploitation helpline.&nbsp;&nbsp;<\/p>\n\n\n\n<p>While it is easy to notice a dramatic spike in any one of the many attributes that describe a trafficking case (for example, an increase in reports linked to a particular location, industry,&nbsp;<em>or&nbsp;<\/em>age range), it is much harder to identify <em>unusual<\/em> or unusually frequent combinations of attributes (for example,&nbsp;a particular location, industry,&nbsp;<em>and<\/em>&nbsp;age range) that may represent an underlying change in real-world trafficking activity.&nbsp;&nbsp;<\/p>\n\n\n\n<p>Compared with statistics on individual attributes, attribute combinations describe actual cases in ways that can directly inform targeted policy responses. The problem is that the number of attribute combinations grows combinatorially, each combination having&nbsp;a maximum frequency at some point in time, and only a small proportion of these maxima representing&nbsp;a meaningful change.<\/p>\n\n\n\n\n\n<h4 id=\"research-question\">Research question<\/h4>\n\n\n\n<p>How can we&nbsp;detect&nbsp;meaningful changes&nbsp;within noisy data&nbsp;streams, in a way that&nbsp;accounts&nbsp;for&nbsp;intrinsic&nbsp;variability over time,&nbsp;reveals&nbsp;emerging&nbsp;groups&nbsp;of interrelated&nbsp;records&nbsp;and attribute values,&nbsp;and&nbsp;enables&nbsp;differentiated&nbsp;policy response?&nbsp;<\/p>\n\n\n\n<h4 id=\"enabling-technology\">Enabling technology<\/h4>\n\n\n\n<p>For this problem, we&nbsp;are collaborating with&nbsp;the School of&nbsp;Mathematics&nbsp;Research at the University of Bristol&nbsp;on how to apply their&nbsp;recent advances in graph statistics to the analysis of human trafficking data. With the CTDC <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.ctdatacollaborative.org\/global-synthetic-dataset\" target=\"_blank\" rel=\"noreferrer noopener\">Global Human Trafficking Synthetic Dataset<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;as representative&nbsp;data,&nbsp;we can connect pairs of attributes based on the number of records sharing both attributes in each time period of interest (for example, for each year of victim registration).&nbsp;&nbsp;<\/p>\n\n\n\n<p>Given this time series of graphs, we can use&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2007.10455\" target=\"_blank\" rel=\"noreferrer noopener\">Unfolded Adjacency Spectral Embedding (UASE)<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> to map all attributes over all time periods into a single embedded space with the <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2106.01282\" target=\"_blank\" rel=\"noreferrer noopener\">strong stability guarantee<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;that constant positions in this space represent constant patterns of behavior.&nbsp;The more similar the behavior of two nodes, the closer their positions in the embedding.&nbsp;By&nbsp;applying&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2106.01260\" target=\"_blank\" rel=\"noreferrer noopener\">new insights into the measurement of relatedness<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;within embedded spaces, we can identify&nbsp;groups&nbsp;of attribute nodes \u201cconverging\u201d towards one another in a given time period, with respect to all other periods, as a measure of meaningful change normalized over all attributes and periods.&nbsp;<\/p>\n\n\n\n<p>To date,&nbsp;we&nbsp;have informally&nbsp;observed&nbsp;that&nbsp;combinations of converging attributes typically&nbsp;coincide with the maximum absolute or relative&nbsp;frequency&nbsp;of the detected attribute combination over all time\u2014something that would immediately be understood as an \u201cinsight.\u201d&nbsp;Due to&nbsp;the graph method used to generate it,&nbsp;these insights are&nbsp;both structurally and&nbsp;statistically meaningful. However, understanding whether this represents&nbsp;a meaningful change in the real world demands domain knowledge beyond the data. This is why, as with <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/microsoft\/synthetic-data-showcase\" target=\"_blank\" rel=\"noreferrer noopener\">Synthetic Data Showcase<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, we&nbsp;use Power BI to create visual interfaces for interactively exploring and explaining&nbsp;sets&nbsp;of candidate insights&nbsp;in the context of other real-world data sources,&nbsp;like&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/unstats.un.org\/sdgs\/indicators\/database\/\" target=\"_blank\" rel=\"noreferrer noopener\">UN SDG indicators<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;on established causes of human trafficking.&nbsp;<\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><a data-bi-bhvr=\"14\"  data-bi-cn=\"Interactive dashboard in Power BI titled \u201cExplore converging attribute patterns in human trafficking case records over time\u201d. To the left, a histogram of pattern count by year, with a maximum of 140 patterns in 2019. The year 2019 is also selected, showing a table of the 140 patterns linked to 2019. The patterns are ranked by a salience score, with the top pattern having a salience score of 5, a length of 11 attributes, and a link to 100 cases. This pattern is selected and contains a range of control methods for citizens of Micronesia exploited in the US. Below the table, there is a time series titled \u201cPattern salience\u201d that shows a sole spike in 2019. To the right of this time series is another time series drawn from data on UN SDG indicators, showing a rise in the number of persons directly affected by disaster in Micronesia in 2019. Below these time series is an explanation of the pattern. This explanation notes that the pattern detected in 2019 corresponds to the historic maximum count of that attribute combination, and that it is one of 37 detected patterns of length 11, out of 735,860 distinct combinations with this length, representing a pattern detection rate of 0.01%.\" href=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"2040\" height=\"1145\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final.jpg\" alt=\"Interactive dashboard in Power BI titled \u201cExplore converging attribute patterns in human trafficking case records over time\u201d. To the left, a histogram of pattern count by year, with a maximum of 140 patterns in 2019. The year 2019 is also selected, showing a table of the 140 patterns linked to 2019. The patterns are ranked by a salience score, with the top pattern having a salience score of 5, a length of 11 attributes, and a link to 100 cases. This pattern is selected and contains a range of control methods for citizens of Micronesia exploited in the US. Below the table, there is a time series titled \u201cPattern salience\u201d that shows a sole spike in 2019. To the right of this time series is another time series drawn from data on UN SDG indicators, showing a rise in the number of persons directly affected by disaster in Micronesia in 2019. Below these time series is an explanation of the pattern. This explanation notes that the pattern detected in 2019 corresponds to the historic maximum count of that attribute combination, and that it is one of 37 detected patterns of length 11, out of 735,860 distinct combinations with this length, representing a pattern detection rate of 0.01%.\" class=\"wp-image-776437\" title=\"Use of Unfolded Adjacency Spectral Embedding (UASE) to detect salient patterns within human trafficking victim case records\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final.jpg 2040w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-300x168.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-1024x575.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-768x431.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-1536x862.jpg 1536w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-343x193.jpg 343w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/SR-Blog_Fig3_UnfoldedSpectralEmbedding_Final-960x540.jpg 960w\" sizes=\"(max-width: 2040px) 100vw, 2040px\" \/><span class=\"sr-only\"> (opens in new tab)<\/span><\/a><figcaption class=\"wp-element-caption\">In this example, we use Power BI to review patterns detected in the CTDC Global Human Trafficking Synthetic Dataset. Having selected 2019 as a year of interest, we can see the most salient pattern in that period describes 100 citizens of Micronesia exploited in the US. Comparing this spike with UN SDG indicators, we observe that 2019 saw a substantial increase in the number of people directly affected by disaster in Micronesia: around 30% of the population. Web search revealed that Micronesia was so badly impacted by Typhoon&nbsp;Wutip&nbsp;in February 2019 that it required a&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" data-bi-bhvr=\"14\"  data-bi-cn=\"Interactive dashboard in Power BI titled \u201cExplore converging attribute patterns in human trafficking case records over time\u201d. To the left, a histogram of pattern count by year, with a maximum of 140 patterns in 2019. The year 2019 is also selected, showing a table of the 140 patterns linked to 2019. The patterns are ranked by a salience score, with the top pattern having a salience score of 5, a length of 11 attributes, and a link to 100 cases. This pattern is selected and contains a range of control methods for citizens of Micronesia exploited in the US. Below the table, there is a time series titled \u201cPattern salience\u201d that shows a sole spike in 2019. To the right of this time series is another time series drawn from data on UN SDG indicators, showing a rise in the number of persons directly affected by disaster in Micronesia in 2019. Below these time series is an explanation of the pattern. This explanation notes that the pattern detected in 2019 corresponds to the historic maximum count of that attribute combination, and that it is one of 37 detected patterns of length 11, out of 735,860 distinct combinations with this length, representing a pattern detection rate of 0.01%.\" href=\"https:\/\/humanitariancompendium.iom.int\/appeals\/micronesia-2019\" target=\"_blank\" rel=\"noreferrer noopener\">humanitarian response by IOM<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>. Use of the dataset and tool therefore revealed a plausible connection between a localized disaster event and a trafficking victim cluster that might warrant further investigation and potential policy response.<\/figcaption><\/figure>\n\n\n\n\n\n<p>With the 2021 TAT accelerator still at an early stage,&nbsp;the&nbsp;CTDC&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.ctdatacollaborative.org\/global-synthetic-dataset\" target=\"_blank\" rel=\"noreferrer noopener\">Global Human Trafficking Synthetic Dataset<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> provided a realistic and&nbsp;recognizable&nbsp;foundation for&nbsp;demonstrating new analytic capabilities that can help address real, emerging (or previously unobserved) forms of the problem.&nbsp;As the accelerator proceeds, we will work to enable&nbsp;each organization to apply these capabilities to their own data,&nbsp;progressing&nbsp;through a series of external demonstrations on open data, mock data,&nbsp;and de-identified data,&nbsp;toward&nbsp;internal testing on actual sensitive data and integration with real-world tools and workflows.&nbsp;<\/p>\n\n\n\n<p>At the&nbsp;2021&nbsp;Microsoft&nbsp;Research Summit, we will share&nbsp;our TAT accelerator work in progress&nbsp;and demonstrate our new dynamic graph analysis capabilities using the newly&nbsp;available CTDC&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.ctdatacollaborative.org\/global-synthetic-dataset\" target=\"_blank\" rel=\"noreferrer noopener\">Global Human Trafficking Synthetic Dataset<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>, which contains evidence of twice as many attribute relationships as the previous <em>k<\/em>-anonymous data release. Following the summit, we will also update our open-source <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/microsoft\/graspologic\" target=\"_blank\" rel=\"noreferrer noopener\">graspologic<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> graph statistics package to include all of the&nbsp;methods&nbsp;from our collaborators at University of Bristol supporting our anti-trafficking efforts. Examples of end-to-end&nbsp;use of these methods, complete with interactive Power BI reports, will be made available following the conclusion of the TAT accelerator in early 2022.<\/p>\n\n\n\n\n\n<p>Partner research into dynamic graph embedding used in this&nbsp;case study and our&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/github.com\/microsoft\/graspologic\" target=\"_blank\" rel=\"noreferrer noopener\">graspologic<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>&nbsp;graph statistics package:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2007.10455\" target=\"_blank\" rel=\"noreferrer noopener\">The multilayer random dot product graph<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2106.01282\" target=\"_blank\" rel=\"noreferrer noopener\">Spectral embedding for dynamic networks with stability guarantees<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/arxiv.org\/abs\/2106.01260\" target=\"_blank\" rel=\"noreferrer noopener\">Matrix&nbsp;factorisation&nbsp;and the interpretation of geodesic distance<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n\n\n\n<li><a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/ieeexplore.ieee.org\/abstract\/document\/8215766\" target=\"_blank\" rel=\"noreferrer noopener\">A Central Limit Theorem for an Omnibus Embedding of Multiple Random Dot Product Graphs<span class=\"sr-only\"> (opens in new tab)<\/span><\/a><\/li>\n<\/ul>\n\n\n\n\n\n<h2 id=\"towards-combinatorial-impact\">Towards&nbsp;combinatorial impact<\/h2>\n\n\n\n<p>For Synthetic&nbsp;Data&nbsp;Showcase,&nbsp;attribute combinations&nbsp;represent&nbsp;privacy to be preserved. For ShowWhy, they represent bias to be controlled.&nbsp;And for our dynamic graph capabilities,&nbsp;they&nbsp;represent insights to be revealed. The result in each case&nbsp;is a form of evidence that wouldn\u2019t otherwise exist, used to tackle problems that couldn\u2019t otherwise be solved.&nbsp;And while we have&nbsp;addressed the private, observational, and&nbsp;temporal&nbsp;nature of real-world data separately, the reality is that&nbsp;many&nbsp;datasets&nbsp;across domains&nbsp;share all these qualities. By making our collection of real-world evidence&nbsp;tools available&nbsp;open-source, we hope to maximize the ability of organizations around the world to contribute to a shared evidence base in their problem domain, for any and all problems of societal significance, both today and into the future.&nbsp;<\/p>\n\n\n\n<p>As a next step&nbsp;for our&nbsp;anti-trafficking&nbsp;work in particular, we are&nbsp;also&nbsp;excited to announce&nbsp;that Microsoft&nbsp;has joined&nbsp;<a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.tellfinderalliance.com\/\" target=\"_blank\" rel=\"noreferrer noopener\">TellFinder Alliance<span class=\"sr-only\"> (opens in new tab)<\/span><\/a>\u2014a global&nbsp;network of partners&nbsp;working to combat&nbsp;human trafficking using ephemeral web data. The associated <a class=\"msr-external-link glyph-append glyph-append-open-in-new-tab glyph-append-xsmall\" href=\"https:\/\/www.tellfinderalliance.com\/tellfinder-technology\" target=\"_blank\" rel=\"noreferrer noopener\">TellFinder<span class=\"sr-only\"> (opens in new tab)<\/span><\/a> application already helps investigators and analysts develop evidence&nbsp;at the&nbsp;case level, leading to&nbsp;the prosecution&nbsp;of&nbsp;both individual traffickers and organized&nbsp;trafficking networks. By applying our tools with partners through both Tech Against Trafficking and TellFinder Alliance, we hope to develop the evidence&nbsp;that&nbsp;will help&nbsp;shape future&nbsp;interventions at the policy level\u2014disrupting the mechanisms&nbsp;by which trafficking takes&nbsp;place and leaving no room in society for any kind of slavery or exploitation.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>From the&nbsp;intense&nbsp;shock of the&nbsp;COVID-19 pandemic&nbsp;to the effects of climate change, our global society has never faced greater risk.&nbsp;The&nbsp;Societal Resilience team at Microsoft Research was&nbsp;established&nbsp;in recognition of this risk and tasked with&nbsp;developing&nbsp;open technologies that enable&nbsp;a&nbsp;scalable response in times of crisis.&nbsp;And just as we think about scalability in a holistic way\u2014scaling across different forms of&nbsp;common&nbsp;problems, for different [&hellip;]<\/p>\n","protected":false},"author":39507,"featured_media":777835,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"categories":[1],"tags":[],"research-area":[13556,13563],"msr-region":[],"msr-event-type":[],"msr-locale":[268875],"msr-post-option":[243984],"msr-impact-theme":[],"msr-promo-type":[],"msr-podcast-series":[],"class_list":["post-776320","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-research-blog","msr-research-area-artificial-intelligence","msr-research-area-data-platform-analytics","msr-locale-en_us","msr-post-option-blog-homepage-featured"],"msr_event_details":{"start":"","end":"","location":""},"podcast_url":"","podcast_episode":"","msr_research_lab":[],"msr_impact_theme":[],"related-publications":[],"related-downloads":[],"related-videos":[],"related-academic-programs":[],"related-groups":[734455,901101],"related-projects":[911022],"related-events":[],"related-researchers":[{"type":"user_nicename","value":"Darren Edge","user_id":31509,"display_name":"Darren Edge","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/daedge\/\" aria-label=\"Visit the profile page for Darren Edge\">Darren Edge<\/a>","is_active":false,"last_first":"Edge, Darren","people_section":0,"alias":"daedge"},{"type":"user_nicename","value":"Jonathan Larson","user_id":32385,"display_name":"Jonathan Larson","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/jolarso\/\" aria-label=\"Visit the profile page for Jonathan Larson\">Jonathan Larson<\/a>","is_active":false,"last_first":"Larson, Jonathan","people_section":0,"alias":"jolarso"}],"msr_type":"Post","featured_image_thumbnail":"<img width=\"960\" height=\"540\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-scaled-960x540.jpg\" class=\"img-object-cover\" alt=\"chart: CTDC global dataset on victims of trafficking\" decoding=\"async\" loading=\"lazy\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-scaled-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-1536x865.jpg 1536w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-2048x1153.jpg 2048w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-343x193.jpg 343w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-1280x720.jpg 1280w, https:\/\/www.microsoft.com\/en-us\/research\/uploads\/prod\/2021\/09\/1400x788_Resilience_no_logo_still-1920x1080.jpg 1920w\" sizes=\"(max-width: 960px) 100vw, 960px\" \/>","byline":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/daedge\/\" title=\"Go to researcher profile for Darren Edge\" aria-label=\"Go to researcher profile for Darren Edge\" data-bi-type=\"byline author\" data-bi-cN=\"Darren Edge\">Darren Edge<\/a> and <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/jolarso\/\" title=\"Go to researcher profile for Jonathan Larson\" aria-label=\"Go to researcher profile for Jonathan Larson\" data-bi-type=\"byline author\" data-bi-cN=\"Jonathan Larson\">Jonathan Larson<\/a>","formattedDate":"September 23, 2021","formattedExcerpt":"From the&nbsp;intense&nbsp;shock of the&nbsp;COVID-19 pandemic&nbsp;to the effects of climate change, our global society has never faced greater risk.&nbsp;The&nbsp;Societal Resilience team at Microsoft Research was&nbsp;established&nbsp;in recognition of this risk and tasked with&nbsp;developing&nbsp;open technologies that enable&nbsp;a&nbsp;scalable response in times of crisis.&nbsp;And just as we think about scalability&hellip;","locale":{"slug":"en_us","name":"English","native":"","english":"English"},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/776320"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/users\/39507"}],"replies":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/comments?post=776320"}],"version-history":[{"count":33,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/776320\/revisions"}],"predecessor-version":[{"id":927360,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/776320\/revisions\/927360"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/777835"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=776320"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/categories?post=776320"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/tags?post=776320"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=776320"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=776320"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=776320"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=776320"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=776320"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=776320"},{"taxonomy":"msr-promo-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-promo-type?post=776320"},{"taxonomy":"msr-podcast-series","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-podcast-series?post=776320"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}