{"683596":{"#nid":"683596","#data":{"type":"event","title":"Machine Learning Seminar Series Spring 2026 | Learning How to Reduce Uncertainty","body":[{"value":"\u003Cdiv\u003E\u003Ch2\u003E\u003Cstrong\u003EPLEASE NOTE SEMINAR IS TUESDAY THE 21st!\u003C\/strong\u003E\u003C\/h2\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EAbstract: \u003C\/strong\u003EHow should we act under incomplete information?\u0026nbsp;We must integrate over missing values to assess the risk of a downstream decision, or the expected risk reduction from gathering further information.\u0026nbsp; Applications include NLP annotation, LLM workflows, clinical medicine, experimental design, and many others.\u0026nbsp; The classical approach to predicting missing values involves fitting some generative model to the incomplete data.\u0026nbsp; Alas, classical methods are slow and often inaccurate due to local optima, poor mixing, or model misspecification.\u0026nbsp; We propose a low-fuss, efficient alternative with fewer assumptions: given incomplete data, train a neural network to directly predict masked values. Our generic network architecture resembles BERT: it constructs a representation for each random variable\u0027s posterior marginal distribution and iteratively refines it through attention on other random variables and entities.\u0026nbsp; We find that this posterior marginal Transformer (\u0022Marformer\u0022) works much better than EM and almost as well as MCMC even when those methods know the true data generating model, and it outperforms them on naturally occurring data. We outline future methods for adaptive masking and a future design for managing data collection efforts.\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EBio: \u003C\/strong\u003E\u003Ca href=\u0022https:\/\/www.cs.jhu.edu\/~jason\/index.html\u0022\u003EJason Eisner\u003C\/a\u003E is Professor of Computer Science at Johns Hopkins University and a Fellow of the Association for Computational Linguistics. At Johns Hopkins, he is also affiliated with the \u003Ca href=\u0022https:\/\/clsp.jhu.edu\/\u0022\u003ECenter for Language and Speech Processing\u003C\/a\u003E, the \u003Ca href=\u0022https:\/\/minds.jhu.edu\/\u0022\u003EMathematical Institute for Data Science\u003C\/a\u003E, the \u003Ca href=\u0022https:\/\/cogsci.jhu.edu\/\u0022\u003ECognitive Science Department\u003C\/a\u003E, and the \u003Ca href=\u0022https:\/\/ai.jhu.edu\/\u0022\u003EData Science and AI Institute\u003C\/a\u003E. His goal is to develop the probabilistic modeling, inference, and learning techniques needed for a unified model of all kinds of linguistic structure, and to connect existing models (such as LLMs) to commonsense reasoning, formal reasoning, and downstream applications. His 180+ papers have presented various algorithms for parsing, machine translation, and weighted finite-state machines; formalizations, algorithms, theorems, and empirical results in computational phonology; unsupervised or semi-supervised learning methods for syntax, morphology, and word-sense disambiguation; and principled methods for conversational AI, including neural language modeling and semantic parsing. From 2019-2024 he was Director of Research at Microsoft Semantic Machines, which developed new approaches to conversational AI. He is also the lead designer of Dyna, a declarative programming language that provides an infrastructure for AI algorithms. He has received 3 school-wide awards for excellence in teaching, most recently in 2025, as well as recent\u0026nbsp;\u003C\/p\u003E\u003Ch2\u003E\u003Ca href=\u0022https:\/\/nam12.safelinks.protection.outlook.com\/?url=https%3A%2F%2Fgatech.zoom.us%2Fj%2F97755001285%3Fpwd%3DtVbLPUl4mWarolJ3gY7b36atmUYFk6.1\u0026amp;data=05%7C02%7Cchrista.ernst%40research.gatech.edu%7C4efe1ffb335b471c7cb508de9a3eb9dd%7C482198bbae7b4b258b7a6d7f32faa083%7C1%7C0%7C639117791347005758%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C\u0026amp;sdata=WnVXcGt68r3twltHh0JeGi%2ByZFSjJKJM2iqIOdOPzMw%3D\u0026amp;reserved=0\u0022\u003E\u0026nbsp;Zoom Link\u003C\/a\u003E\u003C\/h2\u003E\u003Ch3\u003EMeeting ID: 977 5500 1285 \u0026nbsp;| Passcode: 247306\u0026nbsp;\u003C\/h3\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cem\u003EFor more information, or for CODA guest access, please contact \u003C\/em\u003E\u003Ca href=\u0022mailto:shatcher8@gatech.edu\u0022 title=\u0022mailto:shatcher8@gatech.edu\u0022\u003E\u003Cem\u003Eshatcher8@gatech.edu\u003C\/em\u003E\u003C\/a\u003E\u003Cem\u003E at least 2 business days prior to the event.\u003C\/em\u003E\u003C\/p\u003E\u003C\/div\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003E\u003Cstrong\u003EPLEASE NOTE SEMINAR IS TUESDAY THE 21st!\u003C\/strong\u003E\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Featuring | Jason Eisner - Professor of Computer Science, Johns Hopkins University"}],"uid":"27863","created_gmt":"2025-08-06 17:20:58","changed_gmt":"2026-04-15 17:31:14","author":"Christa Ernst","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2026-04-21T12:00:00-04:00","event_time_end":"2026-04-21T13:00:00-04:00","event_time_end_last":"2026-04-21T13:00:00-04:00","gmt_time_start":"2026-04-21 16:00:00","gmt_time_end":"2026-04-21 17:00:00","gmt_time_end_last":"2026-04-21 17:00:00","rrule":null,"timezone":"America\/New_York"},"location":"CODA Building 9th floor Atrium","extras":[],"groups":[{"id":"1278","name":"College of Sciences"},{"id":"545781","name":"Institute for Data Engineering and Science"}],"categories":[],"keywords":[{"id":"9167","name":"machine learning"},{"id":"654","name":"College of Computing"},{"id":"187023","name":"go-data"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[{"id":"78761","name":"Faculty\/Staff"},{"id":"177814","name":"Postdoc"},{"id":"78771","name":"Public"},{"id":"174045","name":"Graduate students"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003Echrista.ernst@research.gatech.edu\u003C\/p\u003E","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}