onr:"swepub:oai:research.chalmers.se:8e3efae8-7e02-4e98-a700-aed382738fab"
Search: onr:"swepub:oai:research.chalmers.se:8e3efae8-7e02-4e98-a700-aed382738fab" > You can have your e...
- 1 of 1
- Previous record
- Next record
- To hitlist
- Meding, IsakZenseact AB,Zenseact, Sweden (author)
You can have your ensemble and run it too - Deep Ensembles Spread Over Time
- Article/chapterEnglish2023
Publisher, publication year, extent ...
- IEEE COMPUTER SOC,2023
Numbers
- LIBRIS-ID:oai:research.chalmers.se:8e3efae8-7e02-4e98-a700-aed382738fab
- https://doi.org/10.1109/ICCVW60793.2023.00434DOI
- https://research.chalmers.se/publication/539627URI
- https://lup.lub.lu.se/record/9bf0dbfe-e6ab-4b88-81de-42478b0d290cURI
- https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-201872URI
- https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-199065URI
Supplementary language notes
- Language:English
- Summary in:English
Part of subdatabase
- SwePubSwePub
Classification
Notes
- Funding Agencies|Wallenberg AI, Autonomous Systems, and Software Programme (WASP)
- Ensembles of independently trained deep neural networks yield uncertainty estimates that rival Bayesian networks in performance. They also offer sizable improvements in terms of predictive performance over single models. However, deep ensembles are not commonly used in environments with limited computational budget - such as autonomous driving - since the complexity grows linearly with the number of ensemble members. An important observation that can be made for robotics applications, such as autonomous driving, is that data is typically sequential. For instance, when an object is to be recognized, an autonomous vehicle typically observes a sequence of images, rather than a single image. This raises the question, could the deep ensemble be spread over time?In this work, we propose and analyze Deep Ensembles Spread Over Time (DESOT). The idea is to apply only a single ensemble member to each data point in the sequence, and fuse the predictions over a sequence of data points. We implement and experiment with DESOT for traffic sign classification, where sequences of tracked image patches are to be classified. We find that DESOT obtains the benefits of deep ensembles, in terms of predictive and uncertainty estimation performance, while avoiding the added computational cost. Moreover, DESOT is simple to implement and does not require sequences during training. Finally, we find that DESOT, like deep ensembles, outperform single models for out-of-distribution detection.
Subject headings and genre
- NATURVETENSKAP Data- och informationsvetenskap Systemvetenskap, informationssystem och informatik hsv//swe
- NATURAL SCIENCES Computer and Information Sciences Information Systems hsv//eng
- NATURVETENSKAP Data- och informationsvetenskap Datavetenskap hsv//swe
- NATURAL SCIENCES Computer and Information Sciences Computer Sciences hsv//eng
- NATURVETENSKAP Data- och informationsvetenskap Datorseende och robotik hsv//swe
- NATURAL SCIENCES Computer and Information Sciences Computer Vision and Robotics hsv//eng
- traffic sign recognition
- uncertainty estimation
- out of distribution detection
- ensemble
- ensemble
- out of distribution detection
- traffic sign recognition
- uncertainty estimation
Added entries (persons, corporate bodies, meetings, titles ...)
- Bodin, AlexanderZenseact AB,Zenseact, Sweden (author)
- Tonderski, AdamLund University,Lunds universitet,LTH profilområde: AI och digitalisering,LTH profilområden,Lunds Tekniska Högskola,Datorseende och maskininlärning,Forskargrupper vid Lunds universitet,LTH Profile Area: AI and Digitalization,LTH Profile areas,Faculty of Engineering, LTH,Computer Vision and Machine Learning,Lund University Research Groups,Zenseact AB,Zenseact, Sweden; Lund Univ, Sweden,Zenseact, Sweden; Lund University, Sweden(Swepub:lu)ad6358to (author)
- Johnander Faxén, JoakimLinköping University,Linköpings universitet,Datorseende,Tekniska fakulteten,Zenseact, Sweden(Swepub:liu)joajo88 (author)
- Petersson, Christoffer,1979Chalmers University of Technology,Chalmers tekniska högskola,Zenseact, Sweden; Chalmers Univ Technol, Sweden,Zenseact, Sweden; Chalmers University of Technology, Sweden(Swepub:cth)chrpet (author)
- Svensson, Lennart,1976Chalmers University of Technology,Chalmers tekniska högskola,Chalmers Univ Technol, Sweden,Chalmers University of Technology, Sweden(Swepub:cth)pale (author)
- Zenseact ABZenseact, Sweden (creator_code:org_t)
Related titles
- In:Proceedings - 2023 IEEE/CVF International Conference on Computer Vision Workshops, ICCVW 2023: IEEE COMPUTER SOC, s. 4022-4031
Internet link
- http://dx.doi.org/10.1109/ICCVW60793.2023.00434FULLTEXT
- https://doi.org/10.1109/ICCVW60793.2023.00434
- https://research.chalmers.se/publication/539627
- https://lup.lub.lu.se/record/9bf0dbfe-e6ab-4b88-81de-42478b0d290c
- https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-201872
- https://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-199065
To the university's database
- 1 of 1
- Previous record
- Next record
- To hitlist
Find more in SwePub
- By the author/editor
- Meding, Isak
- Bodin, Alexander
- Tonderski, Adam
- Johnander Faxén, ...
- Petersson, Chris ...
- Svensson, Lennar ...
- About the subject
-
- NATURAL SCIENCES
- NATURAL SCIENCES
- and Computer and Inf ...
- and Information Syst ...
-
- NATURAL SCIENCES
- NATURAL SCIENCES
- and Computer and Inf ...
- and Computer Science ...
-
- NATURAL SCIENCES
- NATURAL SCIENCES
- and Computer and Inf ...
- and Computer Vision ...
- Articles in the publication
- Proceedings - 20 ...
- By the university
- Chalmers University of Technology
- Lund University
- Linköping University
Search outside SwePub
- Extend your search to:
- Google Book Search
- Google Scholar
Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.
- Copyright © LIBRIS - National Library Systems
- LIBRIS.kb.se
