Recent searches

No recent searches

Search options

Not available on lingo.lol.
lingo.lol is one of the many independent Mastodon servers you can use to participate in the fediverse.
A place for linguists, philologists, and other lovers of languages.

Administered by:

Server stats:

62
active users

lingo.lol: AboutProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.5

#recsys

0 posts0 participants0 posts today
Michael Ekstrand<p><a href="https://hci.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://hci.social/tags/RecSys2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys2025</span></a> Very excited to welcome as many of you as can come to my old city next year for <a href="https://hci.social/tags/RecSys2026" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys2026</span></a> in Minneapolis!</p>
Michael Ekstrand<p>If you're in Vienna, and are wondering what Victorian romance, Ukrainian progressive metalcore, library science, Brazil (1985), the Pope, and more have to do with ethical, human-driven information access, and whether we should put LLMs into everything, come join us tomorrow evening! <a href="https://hci.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> <a href="https://hci.social/tags/IR" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>IR</span></a> <a href="https://hci.social/@mdekstrand/115191259701673286" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">hci.social/@mdekstrand/1151912</span><span class="invisible">59701673286</span></a></p>
Michael Ekstrand<p>Going to Vienna next week to teach in the RecSys summer school, and also to give this lecture. Very much looking forward to it!</p><p>Why do we do <a href="https://hci.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a>, search, etc.? What are the principles and motivations for those goals? What do those goals imply for system design and evaluation? Join us if you’re in town! <a href="https://caiml.org/news/200/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">caiml.org/news/200/</span><span class="invisible"></span></a></p>
Karl Higley<p>turns out the answer to “substack eta wen?” is probably going to be “about five years later, but ghost actually”</p><p>i didn’t really mean to build an audience the first time (on the twitter of olde), it just kinda happened because i kept posting about <a href="https://recsys.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> things that annoyed me. this time, i think i’m ready give it a real go with like…structure and web design and all that stuff serious people do/have/conjure</p>
Karl Higley<p>tfw someone on linkedin posts your production <a href="https://recsys.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> diagram without attribution to suggest that the problems retrieval-augmented generation is struggling with now have already been solved in recommender systems for 20 years</p><p>that diagram is less than 5 years old, was original work, and mostly no one listened 😂</p>
Michael Ekstrand<p>New paper time! 📄🎉</p><p>With <span class="h-card" translate="no"><a href="https://hci.social/@afsanehrazi" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>afsanehrazi</span></a></span>, Aleksandra Sarcevic, Sole Pera, <span class="h-card" translate="no"><a href="https://hci.social/@rburke" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>rburke</span></a></span>, and Katherine Wright, we write about how <a href="https://hci.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> for social good needs participatory design &amp; evaluation. Coming soon to arXiv &amp; ACM TORS, but you saw it first here! <a href="https://md.ekstrandom.net/pubs/tors-co-recommendation" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">md.ekstrandom.net/pubs/tors-co</span><span class="invisible">-recommendation</span></a></p>
Alan Said<p>Did you recently (or are about to) finish your PhD on <a href="https://recsys.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a>, <a href="https://recsys.social/tags/personalization" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>personalization</span></a>, <a href="https://recsys.social/tags/usermodelling" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>usermodelling</span></a>, or related topic and would like to continue working in these fields? Check out our department's information on MSCA postdocs and reach out to me if you are interested <a href="https://www.gu.se/en/applied-information-technology/msca-postdoctoral-fellowship" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">gu.se/en/applied-information-t</span><span class="invisible">echnology/msca-postdoctoral-fellowship</span></a></p>
LensKit RecSys Toolkit<p>Spent some time today updating the list of <a href="https://recsys.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> research using LensKit, and we're now at 105 known research products using LensKit for Python! <a href="https://lenskit.org/research/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">lenskit.org/research/</span><span class="invisible"></span></a></p><p>As a friendly reminder, if you use LensKit in your research, please cite the CIKM'20 paper (<a href="https://md.ekstrandom.net/pubs/lkpy" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">md.ekstrandom.net/pubs/lkpy</span><span class="invisible"></span></a>), as that makes tracking usage easier. Also, drop Michael an e-mail with your paper!</p>
Karl Higley<p>Not entirely hypothetical <a href="https://recsys.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> thought experiment:</p><p>You need to choose one or more models to be deployed in a live recommender system that serves recs to real people. Because the system is new and has a limited pool of users, you’re not able to use either accuracy metrics (like NDCG) or user behavior metrics (like CTR.)</p><p>How do you compare and evaluate models described in the literature for their applicability and suitability in this scenario?</p>
Karl Higley<p>In case you missed it, Yandex recently released a large scale music recommendation dataset with 5B events, audio embeddings, and an organic interaction flag:</p><p><a href="https://arxiv.org/abs/2505.22238" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">arxiv.org/abs/2505.22238</span><span class="invisible"></span></a><br><a href="https://huggingface.co/datasets/yandex/yambda" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">huggingface.co/datasets/yandex</span><span class="invisible">/yambda</span></a></p><p><a href="https://recsys.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a></p>
Michael Ekstrand<p>This from the Netflix team is a very interesting read on modern <a href="https://hci.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> — building a "foundation model" for recommendation from thoughtful principles. I find myself frustrated by a lot of transformers/LLMs/GPT for RecSys work, but this is genuinely interesting. <a href="https://netflixtechblog.com/foundation-model-for-personalized-recommendation-1a0bd8e02d39" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">netflixtechblog.com/foundation</span><span class="invisible">-model-for-personalized-recommendation-1a0bd8e02d39</span></a></p>
Michael Ekstrand<p>New at my blog: I'd been bothered for a few years now in teaching <a href="https://hci.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> by the need for a biased lift computation for informative product-relatedness measures that aren't dominated by low-information items. So this term, I sat down to solve the problem. <a href="https://md.ekstrandom.net/blog/2025/01/biased-lift" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">md.ekstrandom.net/blog/2025/01</span><span class="invisible">/biased-lift</span></a></p>
Michael Ekstrand<p>Do you appreciate news? Do you want to help <a href="https://hci.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> research? POPROX News is open for subscriptions!</p><p>The Platform for OPen Recommendation and Online eXpermentation is an <a href="https://hci.social/tags/NSFfunded" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>NSFfunded</span></a> project to help scientists and journalists explore better ways to select and present news stories. Our MVP is a daily personalized newsletter — if you'd like to receive that, and periodic surveys, sign up at <a href="https://user.poprox.ai/enroll?source=ff&amp;subsource=memd" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">user.poprox.ai/enroll?source=f</span><span class="invisible">f&amp;subsource=memd</span></a></p><p>After signup, you'll receive a confirmation email — check spam if needed.</p>
Michael Ekstrand<p><a href="https://hci.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> fail: this is the second time — at least — LinkedIn only showed me the post(s) advertising an event I wanted to attend after registration closed.</p>
Michael Ekstrand<p>might have developed a new <a href="https://hci.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> hyperparameter search strategy in class last night.</p><p>CalvinSearch: try hyperparameter settings until The Complete Calvin and Hobbes is the top recommendation.</p>
GVogeler<p>At the <a href="https://hcommons.social/tags/dayofdh2024" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>dayofdh2024</span></a> introducing students on the basics of <a href="https://hcommons.social/tags/textencoding" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>textencoding</span></a> (♥️<a href="https://hcommons.social/tags/unicode" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>unicode</span></a>, <a href="https://hcommons.social/tags/xml" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>xml</span></a>, <span class="h-card" translate="no"><a href="https://hcommons.social/@TEIConsortium" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>TEIConsortium</span></a></span>), <a href="https://hcommons.social/tags/geneticediting" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>geneticediting</span></a> , <a href="https://hcommons.social/tags/criticalApparatus" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>criticalApparatus</span></a> , managing <a href="https://dhinfra.at" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">dhinfra.at</span><span class="invisible"></span></a> (<a href="https://hcommons.social/tags/gpu" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>gpu</span></a> clusters for the Austrian <a href="https://hcommons.social/tags/dh" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>dh</span></a> community), talking about scholarly editing Ceija <a href="https://hcommons.social/tags/Stojka" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Stojka</span></a> notebooks <a href="https://www.ceijastojka.org/12924460-the-notebooks" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://www.</span><span class="ellipsis">ceijastojka.org/12924460-the-n</span><span class="invisible">otebooks</span></a> and chatting with Dominik Kowald from KNOW Center Graz about <a href="https://hcommons.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> for DH. Now to preparing my presentation on <a href="https://gams.uni-graz.at/o:voccod" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">gams.uni-graz.at/o:voccod</span><span class="invisible"></span></a> the <a href="https://hcommons.social/tags/SKOS" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>SKOS</span></a> version Denis Muzerelle's Vocabulaire de la <a href="https://hcommons.social/tags/Codicologie" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>Codicologie</span></a> .</p>
Alan Said<p>Still trying to figure out how to scratch your <a href="https://recsys.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a> itch after this year's <span class="h-card" translate="no"><a href="https://hci.social/@Recsys" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>Recsys</span></a></span>? Have a look at the <a href="https://recsys.social/tags/ACMTORS" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ACMTORS</span></a> Call for papers for the Special Issue on <a href="https://recsys.social/tags/RecommenderSystems" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecommenderSystems</span></a> for Good. Will you be be submitting a paper as an xmas gift to the RecSys community? // <span class="h-card" translate="no"><a href="https://idf.social/@Nava" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>Nava</span></a></span> et al.<br><a href="https://dl.acm.org/pb-assets/static_journal_pages/tors/pdf/TORS_SI-Recommender-Systems-for-Good-1721933817383.pdf" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">dl.acm.org/pb-assets/static_jo</span><span class="invisible">urnal_pages/tors/pdf/TORS_SI-Recommender-Systems-for-Good-1721933817383.pdf</span></a></p>
Michael Ekstrand<p>What if <a href="https://hci.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> put as much energy into explicitly anti-fascist recommendations as it does into viewpoint-diverse recommendations?</p><p>What if the goal isn’t to model tolerance for diversity, but to identify items that will increase the reader’s empathy and understanding of their neighbors and their needs?</p><p>This paper provides some thoughts applicable to that direction. <a href="https://doi.org/10.1108/JD-01-2020-0002" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">doi.org/10.1108/JD-01-2020-000</span><span class="invisible">2</span></a></p>
FAccTRec Workshop<p>Come join us on Monday, Oct. 14 for the 7th FAccTRec workshop at <a href="https://recsys.social/tags/RecSys2024" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys2024</span></a>! We've got a great program of papers, a keynote from <span class="h-card" translate="no"><a href="https://recsys.social/@jvinagre" class="u-url mention" rel="nofollow noopener" target="_blank">@<span>jvinagre</span></a></span>, and discussion about fair and responsble <a href="https://recsys.social/tags/recsys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>recsys</span></a>. <a href="https://facctrec.github.io/facctrec2024/program/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="ellipsis">facctrec.github.io/facctrec202</span><span class="invisible">4/program/</span></a></p>
AltRecSys Workshop<p>Very pleased to announce the AltRecSys program at <a href="https://recsys.social/tags/RecSys2024" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys2024</span></a>! Come join us for an afternoon of talks and discussion about the big open questions in <a href="https://recsys.social/tags/RecSys" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>RecSys</span></a> and how we should meet them. <a href="https://altrecsys.github.io/program/" rel="nofollow noopener" translate="no" target="_blank"><span class="invisible">https://</span><span class="">altrecsys.github.io/program/</span><span class="invisible"></span></a></p>
TrendingLive feeds
About