{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/sequential-attend-infer-repeat-generative","title":"Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects","arxiv_id":"1806.01794","date":"2018-06-05","proceeding":"NeurIPS 2018 12","authors":["Adam R. Kosiorek","Hyunjik Kim","Ingmar Posner","Yee Whye Teh"],"abstract":"We present Sequential Attend, Infer, Repeat (SQAIR), an interpretable deep\ngenerative model for videos of moving objects. It can reliably discover and\ntrack objects throughout the sequence of frames, and can also generate future\nframes conditioning on the current frame, thereby simulating expected motion of\nobjects. This is achieved by explicitly encoding object presence, locations and\nappearances in the latent variables of the model. SQAIR retains all strengths\nof its predecessor, Attend, Infer, Repeat (AIR, Eslami et. al., 2016),\nincluding learning in an unsupervised manner, and addresses its shortcomings.\nWe use a moving multi-MNIST dataset to show limitations of AIR in detecting\noverlapping or partially occluded objects, and show how SQAIR overcomes them by\nleveraging temporal consistency of objects. Finally, we also apply SQAIR to\nreal-world pedestrian CCTV data, where it learns to reliably detect, track and\ngenerate walking pedestrians with no supervision.","url_abs":"http://arxiv.org/abs/1806.01794v2","url_pdf":"http://arxiv.org/pdf/1806.01794v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"sequential-attend-infer-repeat-generative","repo_url":"https://github.com/akosiorek/sqair","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.01794","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}