{"id":11893,"date":"2026-03-18T10:34:00","date_gmt":"2026-03-18T08:34:00","guid":{"rendered":"https:\/\/etti.upb.ro\/?p=11893"},"modified":"2026-10-16T02:46:16","modified_gmt":"2026-10-15T23:46:16","slug":"bucharest-deep-learning-meetup-circumventing-shortcuts-in-audio-visual-deepfake-detection-datasets-2026","status":"publish","type":"post","link":"https:\/\/etti.upb.ro\/en\/2026\/03\/18\/bucharest-deep-learning-meetup-circumventing-shortcuts-in-audio-visual-deepfake-detection-datasets-2026\/","title":{"rendered":"Bucharest Deep Learning Meetup \u2013 Avoiding Shortcuts in Audio-Visual Deepfake Detection Datasets"},"content":{"rendered":"<style>.etti-catalog-article{max-width:100%;line-height:1.65;overflow-wrap:anywhere}.etti-catalog-article .etti-archive-note{font-size:.9rem;color:#4b5563;border-left:3px solid #235c83;padding:.5rem .8rem;margin:0 0 1.2rem}.etti-catalog-article .etti-archive-note p{margin:.2rem 0}.etti-catalog-article dl{margin:1rem 0}.etti-catalog-article dt{font-weight:700;margin:1rem 0 .15rem}.etti-catalog-article dd{margin:0 0 .65rem}.etti-catalog-article ul{padding-left:1.4rem}.etti-catalog-article figure{margin:1.4rem 0;max-width:100%}.etti-catalog-article figure img{display:block;max-width:100%;height:auto;margin:0 auto}.etti-catalog-article figcaption{font-size:.9rem;color:#4b5563;margin-top:.5rem;text-align:center}.etti-catalog-article h2{font-size:1.6rem!important;line-height:1.3;margin-top:1.6rem}.etti-catalog-article .etti-event-state{border-left:3px solid #b46b15;background:#fff8ed;padding:.7rem .9rem}<\/style>\n<article class=\"etti-catalog-article\"><aside class=\"etti-archive-note\"><p>Activity Archive \u00b7 2026<\/p><p><a href=\"\/en\/activitati-si-oportunitati\/\">Activities and Opportunities<\/a> \/ <a href=\"\/en\/activitati-si-oportunitati\/cercetare-colaborari\/\">Research and Collaborations<\/a><\/p><p>The information and terms of participation apply to the edition described.<\/p><\/aside><p>Presentation by \u0218tefan Smeu on deepfake detection, the tendency of AI models to learn shortcuts from datasets, and the use of audio-video synchronization for robust detection on new examples. The topic is based on a paper accepted at CVPR 2025.<\/p>\n<h2>Details<\/h2>\n<dl><dt>Organizer<\/dt><dd>Bucharest Deep Learning Meetup; guest speaker: \u0218tefan Smeu.<\/dd><dt>Activity Schedule<\/dt><dd>March 26, 2026, at 6:30 p.m.<\/dd><dt>Location and Format<\/dt><dd>Noua cl\u0103dire FMI, Politehnica Business Tower, Iuliu Maniu 15G, etaj 5, sala 503.<\/dd><dt>Participation<\/dt><dd>Locuri limitate.<\/dd><\/dl>\n<\/article>","protected":false},"excerpt":{"rendered":"<p>Presentation by \u0218tefan Smeu on deepfake detection, the tendency of AI models to learn shortcuts from datasets, and the use of audio-video synchronization for robust detection on new examples. The topic is based on a paper accepted at CVPR 2025.<\/p>","protected":false},"author":5,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_coblocks_attr":"","_coblocks_dimensions":"","_coblocks_responsive_height":"","_coblocks_accordion_ie_support":"","_eb_attr":"","_uag_custom_page_level_css":"","footnotes":""},"categories":[236,240,237],"tags":[246],"class_list":["post-11893","post","type-post","status-publish","format-standard","hentry","category-activitati-si-oportunitati","category-activitati-cercetare-colaborari","category-activitati-cursuri-scoli-concursuri","tag-activitati-2026"],"featured_image_src":null,"author_info":{"display_name":"Cr\u0103ciunescu R\u0103zvan-Eusebiu","author_link":"https:\/\/etti.upb.ro\/en\/author\/razvan-craciunescu\/"},"acf":[],"uagb_featured_image_src":{"full":false,"thumbnail":false,"medium":false,"medium_large":false,"large":false,"elegant_blocks_team_1":false,"elegant_blocks_blog_1":false,"elegant_blocks_slider_1":false,"1536x1536":false,"2048x2048":false,"trp-custom-language-flag":false,"bizberg_medium":false,"bizberg_gallery":false,"bizberg_blog_list":false,"bizberg_detail_image":false,"bizberg_portfolio_homepage":false,"bizberg_blog_list_no_sidebar_1":false,"bizberg_detail_image_no_sidebar":false},"uagb_author_info":{"display_name":"Cr\u0103ciunescu R\u0103zvan-Eusebiu","author_link":"https:\/\/etti.upb.ro\/en\/author\/razvan-craciunescu\/"},"uagb_comment_info":0,"uagb_excerpt":"Prezentare a lui \u0218tefan Smeu despre detec\u021bia deepfake, tendin\u021ba modelelor AI de a \u00eenv\u0103\u021ba scurt\u0103turi din seturile de date \u0219i folosirea sincroniz\u0103rii audio-video pentru detec\u021bie robust\u0103 pe exemple noi. Tema este bazat\u0103 pe o lucrare acceptat\u0103 la CVPR 2025.","_links":{"self":[{"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/posts\/11893","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/users\/5"}],"replies":[{"embeddable":true,"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/comments?post=11893"}],"version-history":[{"count":1,"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/posts\/11893\/revisions"}],"predecessor-version":[{"id":12108,"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/posts\/11893\/revisions\/12108"}],"wp:attachment":[{"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/media?parent=11893"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/categories?post=11893"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/etti.upb.ro\/en\/wp-json\/wp\/v2\/tags?post=11893"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}