Please use this identifier to cite or link to this item:
https://hdl.handle.net/10356/171361
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Yang, Jingkang | en_US |
dc.contributor.author | Zhou, Kaiyang | en_US |
dc.contributor.author | Liu, Ziwei | en_US |
dc.date.accessioned | 2023-10-23T01:47:37Z | - |
dc.date.available | 2023-10-23T01:47:37Z | - |
dc.date.issued | 2023 | - |
dc.identifier.citation | Yang, J., Zhou, K. & Liu, Z. (2023). Full-spectrum out-of-distribution detection. International Journal of Computer Vision, 131(10), 2607-2622. https://dx.doi.org/10.1007/s11263-023-01811-z | en_US |
dc.identifier.issn | 0920-5691 | en_US |
dc.identifier.uri | https://hdl.handle.net/10356/171361 | - |
dc.description.abstract | Existing out-of-distribution (OOD) detection literature clearly defines semantic shift as a sign of OOD but does not have a consensus over covariate shift. Samples experiencing covariate shift but not semantic shift from the in-distribution (ID) are either excluded from the test set or treated as OOD, which contradicts the primary goal in machine learning—being able to generalize beyond the training distribution. In this paper, we take into account both shift types and introduce full-spectrum OOD (F-OOD) detection, a more realistic problem setting that considers both detecting semantic shift and being tolerant to covariate shift; and design three benchmarks. These new benchmarks have a more fine-grained categorization of distributions (i.elet@tokeneonedot, training ID, covariate-shifted ID, near-OOD, and far-OOD) for the purpose of more comprehensively evaluating the pros and cons of algorithms. To address the F-OOD detection problem, we propose SEM, a simple feature-based semantics score function. SEM is mainly composed of two probability measures: one is based on high-level features containing both semantic and non-semantic information, while the other is based on low-level feature statistics only capturing non-semantic image styles. With a simple combination, the non-semantic part is canceled out, which leaves only semantic information in SEM that can better handle F-OOD detection. Extensive experiments on the three new benchmarks show that SEM significantly outperforms current state-of-the-art methods. Our code and benchmarks are released in https://github.com/Jingkang50/OpenOOD . | en_US |
dc.description.sponsorship | Agency for Science, Technology and Research (A*STAR) | en_US |
dc.description.sponsorship | Ministry of Education (MOE) | en_US |
dc.description.sponsorship | Nanyang Technological University | en_US |
dc.language.iso | en | en_US |
dc.relation | MOE-T2EP20221-0012 | en_US |
dc.relation.ispartof | International Journal of Computer Vision | en_US |
dc.rights | © 2023 The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature. All rights reserved. | en_US |
dc.subject | Engineering::Computer science and engineering | en_US |
dc.title | Full-spectrum out-of-distribution detection | en_US |
dc.type | Journal Article | en |
dc.contributor.school | School of Computer Science and Engineering | en_US |
dc.contributor.research | S-Lab | en_US |
dc.identifier.doi | 10.1007/s11263-023-01811-z | - |
dc.identifier.scopus | 2-s2.0-85163080052 | - |
dc.identifier.issue | 10 | en_US |
dc.identifier.volume | 131 | en_US |
dc.identifier.spage | 2607 | en_US |
dc.identifier.epage | 2622 | en_US |
dc.subject.keywords | AI Safety | en_US |
dc.subject.keywords | Model Trustworthy | en_US |
dc.description.acknowledgement | This study is supported by the Ministry of Education, Singapore, under its MOE AcRF Tier 2 (MOE-T2EP20221-0012), NTU NAP, and under the RIE2020 Industry Alignment Fund - Industry Collaboration Projects (IAF-ICP) Funding Initiative, as well as cash and in-kind contribution from the industry partner(s). | en_US |
item.grantfulltext | none | - |
item.fulltext | No Fulltext | - |
Appears in Collections: | SCSE Journal Articles |
SCOPUSTM
Citations
20
19
Updated on Dec 6, 2024
Page view(s)
116
Updated on Dec 9, 2024
Google ScholarTM
Check
Altmetric
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.