{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/106243"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/106243","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Stochastic prediction in sequential high-dimensional observation space","abstract":"Predicting the future in real-world settings, particularly from raw sensory observations such as images, is exceptionally challenging. Real-world events can be stochastic and unpredictable, and the high dimensionality and complexity of natural images require the predictive model to build an intricate understanding of the natural world. Many existing predictive methods tackle this problem by making simplifying assumptions about the environment. One common assumption is that the outcome is deterministic and there is only one plausible future. This can lead to low-quality predictions in real-world settings with stochastic dynamics. In this thesis, we study the importance of stochasticity in predicting high-quality predictions of the raw sequential observations. We develop a stochastic variational video prediction method that predicts a different possible future for each sample of its latent variables. We also provide an alternative method based on normalizing flows. To the best of our knowledge, these models are the first to provide an effective stochastic multi-frame prediction for real-world videos. We demonstrate the capability of these methods in predicting detailed future frames of videos on multiple real-world datasets, both action-free and action-conditioned. We also illustrate how such methods can improve the performance of autonomous agents where future prediction is a core required capability. We illustrate how these predictive models can be used for planning in real and simulated robotic tasks as well as improving the sample efficiency in model based reinforcement learning. We also show how similar stochastic techniques can be applied in other areas where stochasticity can be useful such as real-time style transfer.","abstract_html":"Predicting the future in real-world settings, particularly from raw sensory observations such as images, is exceptionally challenging. Real-world events can be stochastic and unpredictable, and the high dimensionality and complexity of natural images require the predictive model to build an intricate understanding of the natural world. Many existing predictive methods tackle this problem by making simplifying assumptions about the environment. One common assumption is that the outcome is deterministic and there is only one plausible future. This can lead to low-quality predictions in real-world settings with stochastic dynamics. In this thesis, we study the importance of stochasticity in predicting high-quality predictions of the raw sequential observations. We develop a stochastic variational video prediction method that predicts a different possible future for each sample of its latent variables. We also provide an alternative method based on normalizing flows. To the best of our knowledge, these models are the first to provide an effective stochastic multi-frame prediction for real-world videos. We demonstrate the capability of these methods in predicting detailed future frames of videos on multiple real-world datasets, both action-free and action-conditioned. We also illustrate how such methods can improve the performance of autonomous agents where future prediction is a core required capability. We illustrate how these predictive models can be used for planning in real and simulated robotic tasks as well as improving the sample efficiency in model based reinforcement learning. We also show how similar stochastic techniques can be applied in other areas where stochasticity can be useful such as real-time style transfer.","abstract_has_math":false,"creators":["Babaeizadeh, Mohammad"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":["Campbell, Roy H.","Smaragdis, Paris","Koyejo, Sanmi","Levine, Sergey","Erhan, Dumitru"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2020,"date_issued":"2020-03-02T21:58:23Z","date_published":"2020-03-02T21:58:23Z","updated_at":"2026-07-22T22:24:45Z","subjects":["video prediction","deep learning","high dimensional prediction","stochastic prediction"],"languages":["en"],"rights":["Copyright 2019 Mohammad Babaeizadeh"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/106243","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Campbell, Roy H.","Smaragdis, Paris","Koyejo, Sanmi","Levine, Sergey","Erhan, Dumitru"]},{"key":"dc:creator","label":"Author","values":["Babaeizadeh, Mohammad"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2020-03-02T21:58:23Z","2019-12-04","2019-12"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["video prediction","deep learning","high dimensional prediction","stochastic prediction"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2019 Mohammad Babaeizadeh"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/106243"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Predicting the future in real-world settings, particularly from raw sensory observations such as images, is exceptionally challenging. Real-world events can be stochastic and unpredictable, and the high dimensionality and complexity of natural images require the predictive model to build an intricate understanding of the natural world. Many existing predictive methods tackle this problem by making simplifying assumptions about the environment. One common assumption is that the outcome is deterministic and there is only one plausible future. This can lead to low-quality predictions in real-world settings with stochastic dynamics. In this thesis, we study the importance of stochasticity in predicting high-quality predictions of the raw sequential observations. We develop a stochastic variational video prediction method that predicts a different possible future for each sample of its latent variables. We also provide an alternative method based on normalizing flows. To the best of our knowledge, these models are the first to provide an effective stochastic multi-frame prediction for real-world videos. We demonstrate the capability of these methods in predicting detailed future frames of videos on multiple real-world datasets, both action-free and action-conditioned. We also illustrate how such methods can improve the performance of autonomous agents where future prediction is a core required capability. We illustrate how these predictive models can be used for planning in real and simulated robotic tasks as well as improving the sample efficiency in model based reinforcement learning. We also show how similar stochastic techniques can be applied in other areas where stochasticity can be useful such as real-time style transfer.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2020-02-28 without embargo terms","The student, Mohammad Babaeizadeh, accepted the attached license on 2019-12-04 at 15:02.","The student, Mohammad Babaeizadeh, submitted this Dissertation for approval on 2019-12-04 at 15:08.","This Dissertation was approved for publication on 2019-12-04 at 17:26.","DSpace SAF Submission Ingestion Package generated from Vireo submission #14696 on 2020-02-28 at 17:15:08","Made available in DSpace on 2020-03-02T21:58:23Z (GMT). No. of bitstreams: 3 BABAEIZADEH-DISSERTATION-2019.pdf: 23515704 bytes, checksum: a509434960acbecc8e0be923d381984e (MD5) LICENSE.txt: 4217 bytes, checksum: 287add544751d662e6066070cd9990af (MD5) PROQUEST_LICENSE.txt: 4563 bytes, checksum: 9caae9b70e11fdb9c850f7e4a0af36bb (MD5) Previous issue date: 2019-12-04"]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Stochastic prediction in sequential high-dimensional observation space"]}]}],"canonical_facts":{"dc:contributor":["Campbell, Roy H.","Smaragdis, Paris","Koyejo, Sanmi","Levine, Sergey","Erhan, Dumitru"],"dc:creator":["Babaeizadeh, Mohammad"],"dc:date":["2020-03-02T21:58:23Z","2019-12-04","2019-12"],"dc:description":["Predicting the future in real-world settings, particularly from raw sensory observations such as images, is exceptionally challenging. Real-world events can be stochastic and unpredictable, and the high dimensionality and complexity of natural images require the predictive model to build an intricate understanding of the natural world. Many existing predictive methods tackle this problem by making simplifying assumptions about the environment. One common assumption is that the outcome is deterministic and there is only one plausible future. This can lead to low-quality predictions in real-world settings with stochastic dynamics. In this thesis, we study the importance of stochasticity in predicting high-quality predictions of the raw sequential observations. We develop a stochastic variational video prediction method that predicts a different possible future for each sample of its latent variables. We also provide an alternative method based on normalizing flows. To the best of our knowledge, these models are the first to provide an effective stochastic multi-frame prediction for real-world videos. We demonstrate the capability of these methods in predicting detailed future frames of videos on multiple real-world datasets, both action-free and action-conditioned. We also illustrate how such methods can improve the performance of autonomous agents where future prediction is a core required capability. We illustrate how these predictive models can be used for planning in real and simulated robotic tasks as well as improving the sample efficiency in model based reinforcement learning. We also show how similar stochastic techniques can be applied in other areas where stochasticity can be useful such as real-time style transfer.","Submission original under an indefinite embargo labeled 'Open Access'. The submission was exported from vireo on 2020-02-28 without embargo terms","The student, Mohammad Babaeizadeh, accepted the attached license on 2019-12-04 at 15:02.","The student, Mohammad Babaeizadeh, submitted this Dissertation for approval on 2019-12-04 at 15:08.","This Dissertation was approved for publication on 2019-12-04 at 17:26.","DSpace SAF Submission Ingestion Package generated from Vireo submission #14696 on 2020-02-28 at 17:15:08","Made available in DSpace on 2020-03-02T21:58:23Z (GMT). No. of bitstreams: 3 BABAEIZADEH-DISSERTATION-2019.pdf: 23515704 bytes, checksum: a509434960acbecc8e0be923d381984e (MD5) LICENSE.txt: 4217 bytes, checksum: 287add544751d662e6066070cd9990af (MD5) PROQUEST_LICENSE.txt: 4563 bytes, checksum: 9caae9b70e11fdb9c850f7e4a0af36bb (MD5) Previous issue date: 2019-12-04"],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/2142/106243"],"dc:language":["en"],"dc:rights":["Copyright 2019 Mohammad Babaeizadeh"],"dc:subject":["video prediction","deep learning","high dimensional prediction","stochastic prediction"],"dc:title":["Stochastic prediction in sequential high-dimensional observation space"],"dc:type":["text"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:24:45Z"}