{"id":1107372,"date":"2024-12-06T06:00:00","date_gmt":"2024-12-06T14:00:00","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?p=1107372"},"modified":"2024-12-13T10:17:57","modified_gmt":"2024-12-13T18:17:57","slug":"abstracts-neurips-2024-with-dylan-foster","status":"publish","type":"post","link":"https:\/\/www.microsoft.com\/en-us\/research\/podcast\/abstracts-neurips-2024-with-dylan-foster\/","title":{"rendered":"Abstracts: NeurIPS 2024 with Dylan Foster"},"content":{"rendered":"\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1400\" height=\"788\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788.jpg\" alt=\"Illustrated image of Dylan Foster for the Abstracts series on the Microsoft Research Podcast.\" class=\"wp-image-1108887\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788.jpg 1400w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/11\/Dylan-Foster_Abstracts_Hero_Feature_No_Text_1400x788-1280x720.jpg 1280w\" sizes=\"auto, (max-width: 1400px) 100vw, 1400px\" \/><\/figure>\n\n\n<div class=\"wp-block-msr-podcast-container my-4\">\n\t<iframe loading=\"lazy\" src=\"https:\/\/player.blubrry.com\/?podcast_id=139303363&modern=1\" class=\"podcast-player\" frameborder=\"0\" height=\"164px\" width=\"100%\" scrolling=\"no\" title=\"Podcast Player\"><\/iframe>\n<\/div>\n\n\n\n<p>Members of the research community at Microsoft work continuously to advance their respective fields.<em> Abstracts&nbsp;<\/em>brings its audience to the cutting edge with them through short, compelling conversations about new and noteworthy achievements.&nbsp;<\/p>\n\n\n\n<p>In this episode, Principal Researcher <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/dylanfoster\/\">Dylan Foster<\/a> joins host Amber Tingle to discuss\u00a0the paper <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/publication\/reinforcement-learning-under-latent-dynamics-toward-statistical-and-algorithmic-modularity\/\">\u201cReinforcement Learning Under Latent Dynamics: Toward Statistical and Algorithmic Modularity,\u201d<\/a> an oral presentation at this year\u2019s <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/event\/neurips-2024\/\">Conference on Neural Information Processing Systems<\/a> <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/event\/neurips-2024\/\">(NeurIPS)<\/a>. In the paper, Foster and his coauthors explore whether well-studied RL algorithms for simple problems can be leveraged to solve RL problems with high-dimensional observations and latent dynamics, part of larger efforts to identify algorithm design principles that can enable agents to learn quickly via trial and error in unfamiliar environments.<\/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\"><a data-bi-type=\"button\" class=\"wp-block-button__link wp-element-button\" href=\"https:\/\/www.microsoft.com\/en-us\/research\/publication\/reinforcement-learning-under-latent-dynamics-toward-statistical-and-algorithmic-modularity\/\">Read the paper<\/a><\/div>\n<\/div>\n\n\n<div class=\"wp-block-msr-show-more\">\n\t<div class=\"bg-neutral-100 p-5\">\n\t\t<div class=\"show-more-show-less\" data-mount=\"show-more-show-less\">\n\t\t\t<div>\n\t\t\t\t<span>\n\t\t\t\t\t\n\n<h2 class=\"wp-block-heading\" id=\"transcript\">Transcript<\/h2>\n\n\n\n<p>[MUSIC]<\/p>\n\n\n\n<p><strong>AMBER TINGLE:<\/strong> Welcome to <em>Abstracts<\/em>, a Microsoft Research Podcast that puts the spotlight on world-class research <em>in brief<\/em>. I\u2019m Amber Tingle. In this series, members of the research community at Microsoft give us a quick snapshot\u2014or a <em>podcast abstract<\/em>\u2014of their new and noteworthy papers.<\/p>\n\n\n\n<p>[MUSIC FADES]<\/p>\n\n\n\n<p>Our guest today is Dylan Foster. He is a principal researcher at Microsoft Research and coauthor of a paper called \u201cReinforcement Learning Under Latent Dynamics: Toward Statistical and Algorithmic Modularity.\u201d The work is among the oral presentations at this year&#8217;s Conference on Neural Information Processing Systems, or <em>NeurIPS<\/em>, in Vancouver. Dylan, welcome and thank you for joining us on the podcast!<\/p>\n\n\n\n\t\t\t\t<\/span>\n\t\t\t\t<span id=\"show-more-show-less-toggle-1\" class=\"show-more-show-less-toggleable-content\">\n\t\t\t\t\t\n\n\n\n<p><strong>DYLAN FOSTER:<\/strong> Thanks for having me.<\/p>\n\n\n\n<p><strong>TINGLE: <\/strong>Let&#8217;s start with a brief overview of this paper. Tell us about the problem this work addresses and why the research community should know about it.<\/p>\n\n\n\n<p><strong>FOSTER:<\/strong> So this is a, kind of, a theoretical work on reinforcement learning, or RL. When I say reinforcement learning, broadly speaking, this is talking about the question of how can we design AI agents that are capable of, like, interacting with unknown environments and learning how to solve problems through trial and error. So this is part of some broader agenda we&#8217;ve been doing on, kind of, theoretical foundations of RL. And the key questions we&#8217;re looking at here are what are called, like, exploration and sample efficiency. So this just means we&#8217;re trying to understand, like, what are the algorithm design principles that can allow you to explore an unknown environment and learn as quickly as possible? What we&#8217;re doing in this paper is we&#8217;re, kind of, looking at, how can you most efficiently solve reinforcement learning problems where you&#8217;re faced with very high-dimensional observations, but the underlying dynamics of the system you&#8217;re interacting with are simple? So this is a setting that occurs in a lot of natural reinforcement learning and control problems, especially in the context of, like, say, embodied decision-making. So if you think about, say, games like Pong, you know, the state of the game, like, the state of, like, Pong, is extremely simple. It&#8217;s just, you know, what is the position and velocity of the ball, and, like, where are the paddles? But what we&#8217;d like to be able to do is learn to, you know, like, control or, like, solve games like this from raw pixels or, like, images kind of in the same way that a human would, like, just solve them from vision. So if you look at these types of problems, you know, we call these, like, <em>RL with rich observations<\/em> or RL with <em>latent dynamics<\/em>. You know, these are interesting because they, kind of, require you to explore the system, but they also require, you know, representation learning. Like, you want to be able to use neural nets to learn a mapping from, say, the images you see to the latent state of the system. This is a pretty interesting and nontrivial algorithmic problem. And, kind of, what we do in this work is we take a first step towards something like a unified understanding for how to solve these sorts of, like, rich-observation, or latent dynamics, RL problems.<\/p>\n\n\n\n<p><strong>TINGLE:<\/strong> So how did you go about developing this theoretical framework?<\/p>\n\n\n\n<p><strong>FOSTER: <\/strong>Yeah, so if you look at these sort of RL problems with latent dynamics, this is something that&#8217;s actually received a lot of investigation in theory. And a lot of this goes back to, kind of, early work from our lab from, like, 2016, 2017 or so. There&#8217;s some really interesting results here, but progress was largely on a, like, case-by-case basis, meaning, you know, there are many different ways that you can try to model the latent dynamics of your problem, and, you know, each of these somehow leads to a different algorithm, right. So, like, you know, you think very hard about this modeling assumption. You think about, what would an optimal algorithm look like? And you end up, you know, writing an entire paper about it. And there&#8217;s nothing wrong with that per se, but if you want to be able to iterate quickly and, kind of, try different modeling assumptions and see what works in practice, you know, this is not really tenable. It&#8217;s just too slow. And so the starting point for this work was to, kind of, try to take a different and more modular approach. So the idea is, you know, there are many, many different types of, sort of, systems or modeling assumptions for the dynamics that have been already studied extensively and have entire papers about them for the simpler setting in which you can directly see the state of the system. And so what we wanted to ask here is, is it possible to use these existing results in more of, like, a modular fashion? Like, if someone has already written a paper on how to optimally solve a particular type of MDP, or <em>Markov decision process<\/em>, can we just take their algorithm as is and perhaps plug it into some kind of meta-algorithm that can directly, kind of, combine this with representation learning and use it to solve the corresponding rich-observation, or latent dynamics, RL problem?<\/p>\n\n\n\n<p><strong>TINGLE: <\/strong>What were your major findings? What did you learn during this process?<\/p>\n\n\n\n<p><strong>FOSTER:<\/strong> We started by asking the question sort of exactly the way that I just posed it, right. Like, can we take existing algorithms and use them to solve rich-observation RL problems in a modular fashion? And this turned out to be really tricky. Like, there&#8217;s a lot of natural algorithms you might try that seem promising at first but don&#8217;t exactly work out. And what this, kind of, led us to and, sort of, the first main result in this paper is actually a <em>negative<\/em> result. So what we actually showed is most, sort of, well-studied types of systems or, like, MDPs that have been studied in, like, the prior literature on RL, even if they&#8217;re tractable when you&#8217;re able to directly see the state of the system, they can become statistically intractable once you add, sort of, high-dimensional observations to the picture. And statistically tractable here means the amount of interaction that you need, like the amount of, sort of, attempts to explore the system that you need, in order to learn a good decision-making policy becomes, like, very, very large, like much, much larger than the corresponding, sort of, complexity if you were able to directly see the states of the system. You know, you could look at this and say, I guess we&#8217;re out of luck. You know, maybe there&#8217;s just no hope of solving these sorts of problems. But that&#8217;s perhaps a little too pessimistic. You know, really the way you should interpret this result is just that you need <em>more assumptions<\/em>. And that&#8217;s precisely what the, sort of, second result we have in this paper is. So our second result shows that you can, sort of, bypass this impossibility result and, you know, achieve truly modular algorithms under a couple different types of additional assumptions.<\/p>\n\n\n\n<p><strong>TINGLE: <\/strong>Dylan, I&#8217;d like to know\u2014and I&#8217;m sure our audience would, too\u2014what this work means when it comes to real-world application. What impact will this have on the research community?<\/p>\n\n\n\n<p><strong>FOSTER:<\/strong> Yeah, so maybe I&#8217;ll answer that, um, with two different points. The first one is a broader point, which is, why is it important to understand this problem of exploration and sample efficiency in reinforcement learning? If you look at the, sort of, setting we study in this paper\u2014you know, this, like, RL or decision-making with high-dimensional observations\u2014on the empirical side, people have made a huge amount of progress on this problem through deep reinforcement learning. This was what kind of led to these amazing breakthroughs in solving games like Atari in the last decade. But if you look at these results, the gains are somehow more coming from the, like, inductive bias or the, like, generalization abilities of deep learning and not necessarily from the specific algorithms. So, like, current algorithms do not actually explore very deliberately, and so their sample efficiency is very high. Like, it&#8217;s hard to draw a one-to-one comparison, but you can argue they need, like, far more experience than a human would to solve these sorts of problems. So it&#8217;s not clear that we&#8217;re really anywhere near the ceiling of what can be achieved in terms of, like, how efficiently can you have, you know, an agent learn to solve new problems from trial and error. And I think better algorithms here could potentially be, like, transformative in a lot of different domains. To get into this specific work, I think there&#8217;s a couple of important takeaways for researchers. One is that by giving this impossibility result that shows that RL with latent dynamics is impossible without further assumptions, we&#8217;re kind of narrowing down the search space where other researchers can look for efficient algorithms. The second takeaway is, you know, we are showing that this problem becomes tractable when you make additional assumptions. But I view these more as, like, a proof of concept. Like, we&#8217;re kind of, showing for the first time that it <em>is<\/em> possible to do something nontrivial, but I think a lot more work and research will be required in order to like, you know, build on this and take this to something that can lead to, like, practical algorithms.<s><\/s><\/p>\n\n\n\n<p><strong>TINGLE: <\/strong>Well, Dylan Foster, thank you for joining us today to discuss your paper on reinforcement learning under latent dynamics. We certainly appreciate it.<\/p>\n\n\n\n<p><strong>FOSTER: <\/strong>Thanks a lot. Thanks for having me.<\/p>\n\n\n\n<p>[MUSIC]<\/p>\n\n\n\n<p><strong>TINGLE: <\/strong>And to our listeners, thank you all for tuning in. If you&#8217;d like to read Dylan&#8217;s paper, you may find a link at aka.ms\/abstracts. You can also find the paper on arXiv and on the NeurIPS conference website. I&#8217;m Amber Tingle from Microsoft Research, and we hope you&#8217;ll join us next time on <em>Abstracts<\/em>!<\/p>\n\n\n\n<p>[MUSIC FADES]<\/p>\n\n\t\t\t\t<\/span>\n\t\t\t<\/div>\n\t\t\t<button\n\t\t\t\tclass=\"action-trigger glyph-prepend mt-2 mb-0 show-more-show-less-toggle\"\n\t\t\t\taria-expanded=\"false\"\n\t\t\t\tdata-show-less-text=\"Show less\"\n\t\t\t\ttype=\"button\"\n\t\t\t\taria-controls=\"show-more-show-less-toggle-1\"\n\t\t\t\taria-label=\"Show more content\"\n\t\t\t\tdata-alternate-aria-label=\"Show less content\">\n\t\t\t\tShow more\t\t\t<\/button>\n\t\t<\/div>\n\t<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>Can existing algorithms designed for simple reinforcement learning problems be used to solve more complex RL problems? Researcher Dylan Foster discusses the modular approach he and his coauthors explored in their 2024 NeurIPS paper on RL under latent dynamics.<\/p>\n","protected":false},"author":38004,"featured_media":1110249,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"msr-url-field":"https:\/\/player.blubrry.com\/id\/139303363","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"categories":[240054],"tags":[],"research-area":[13556],"msr-region":[],"msr-event-type":[],"msr-locale":[268875],"msr-post-option":[269148,269142,243990],"msr-impact-theme":[],"msr-promo-type":[],"msr-podcast-series":[268128],"class_list":["post-1107372","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-msr-podcast","msr-research-area-artificial-intelligence","msr-locale-en_us","msr-post-option-approved-for-river","msr-post-option-include-in-river","msr-post-option-podcast-featured","msr-podcast-series-abstracts"],"msr_event_details":{"start":"","end":"","location":""},"podcast_url":"https:\/\/player.blubrry.com\/id\/139303363","podcast_episode":"","msr_research_lab":[],"msr_impact_theme":[],"related-publications":[],"related-downloads":[],"related-videos":[],"related-academic-programs":[],"related-groups":[],"related-projects":[],"related-events":[1088157],"related-researchers":[{"type":"user_nicename","value":"Amber Tingle","user_id":42681,"display_name":"Amber Tingle","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/ambertingle\/\" aria-label=\"Visit the profile page for Amber Tingle\">Amber Tingle<\/a>","is_active":false,"last_first":"Tingle, Amber","people_section":0,"alias":"ambertingle"},{"type":"user_nicename","value":"Dylan Foster","user_id":40330,"display_name":"Dylan Foster","author_link":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/dylanfoster\/\" aria-label=\"Visit the profile page for Dylan Foster\">Dylan Foster<\/a>","is_active":false,"last_first":"Foster, Dylan","people_section":0,"alias":"dylanfoster"}],"msr_type":"Post","featured_image_thumbnail":"<img width=\"960\" height=\"540\" src=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-960x540.jpg\" class=\"img-object-cover\" alt=\"Illustrated image of Dylan Foster for the Abstracts series on the Microsoft Research Podcast.\" decoding=\"async\" loading=\"lazy\" srcset=\"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-960x540.jpg 960w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-300x169.jpg 300w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-1024x576.jpg 1024w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-768x432.jpg 768w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-1066x600.jpg 1066w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-655x368.jpg 655w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-240x135.jpg 240w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-640x360.jpg 640w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1-1280x720.jpg 1280w, https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2024\/12\/Dylan-Foster_Abstracts_Hero_Feature_River_No_Text_1400x788-1.jpg 1400w\" sizes=\"auto, (max-width: 960px) 100vw, 960px\" \/>","byline":"<a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/ambertingle\/\" title=\"Go to researcher profile for Amber Tingle\" aria-label=\"Go to researcher profile for Amber Tingle\" data-bi-type=\"byline author\" data-bi-cN=\"Amber Tingle\">Amber Tingle<\/a> and <a href=\"https:\/\/www.microsoft.com\/en-us\/research\/people\/dylanfoster\/\" title=\"Go to researcher profile for Dylan Foster\" aria-label=\"Go to researcher profile for Dylan Foster\" data-bi-type=\"byline author\" data-bi-cN=\"Dylan Foster\">Dylan Foster<\/a>","formattedDate":"December 6, 2024","formattedExcerpt":"Can existing algorithms designed for simple reinforcement learning problems be used to solve more complex RL problems? Researcher Dylan Foster discusses the modular approach he and his coauthors explored in their 2024 NeurIPS paper on RL under latent dynamics.","locale":{"slug":"en_us","name":"English","native":"","english":"English"},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1107372","targetHints":{"allow":["GET"]}}],"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\/38004"}],"replies":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/comments?post=1107372"}],"version-history":[{"count":24,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1107372\/revisions"}],"predecessor-version":[{"id":1112223,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/posts\/1107372\/revisions\/1112223"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/1110249"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=1107372"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/categories?post=1107372"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/tags?post=1107372"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=1107372"},{"taxonomy":"msr-region","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-region?post=1107372"},{"taxonomy":"msr-event-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-event-type?post=1107372"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=1107372"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=1107372"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=1107372"},{"taxonomy":"msr-promo-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-promo-type?post=1107372"},{"taxonomy":"msr-podcast-series","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-podcast-series?post=1107372"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}