Understanding the Vulnerability of CLIP to Image Compression Article Swipe
Cangxiong Chen
,
Vinay P. Namboodiri
,
Julián Padget
·
YOU?
·
· 2023
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2311.14029
YOU?
·
· 2023
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2311.14029
CLIP is a widely used foundational vision-language model that is used for zero-shot image recognition and other image-text alignment tasks. We demonstrate that CLIP is vulnerable to change in image quality under compression. This surprising result is further analysed using an attribution method-Integrated Gradients. Using this attribution method, we are able to better understand both quantitatively and qualitatively exactly the nature in which the compression affects the zero-shot recognition accuracy of this model. We evaluate this extensively on CIFAR-10 and STL-10. Our work provides the basis to understand this vulnerability of CLIP and can help us develop more effective methods to improve the robustness of CLIP and other vision-language models.
Related Topics
Concepts
Metadata
- Type
- preprint
- Language
- en
- Landing Page
- http://arxiv.org/abs/2311.14029
- https://arxiv.org/pdf/2311.14029
- OA Status
- green
- Cited By
- 1
- Related Works
- 10
- OpenAlex ID
- https://openalex.org/W4389072448
All OpenAlex metadata
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W4389072448Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48550/arxiv.2311.14029Digital Object Identifier
- Title
-
Understanding the Vulnerability of CLIP to Image CompressionWork title
- Type
-
preprintOpenAlex work type
- Language
-
enPrimary language
- Publication year
-
2023Year of publication
- Publication date
-
2023-11-23Full publication date if available
- Authors
-
Cangxiong Chen, Vinay P. Namboodiri, Julián PadgetList of authors in order
- Landing page
-
https://arxiv.org/abs/2311.14029Publisher landing page
- PDF URL
-
https://arxiv.org/pdf/2311.14029Direct link to full text PDF
- Open access
-
YesWhether a free full text is available
- OA status
-
greenOpen access status per OpenAlex
- OA URL
-
https://arxiv.org/pdf/2311.14029Direct OA link when available
- Concepts
-
Computer science, Robustness (evolution), Vulnerability (computing), Artificial intelligence, Image (mathematics), Attribution, Computer vision, Image quality, Computer security, Psychology, Chemistry, Social psychology, Gene, BiochemistryTop concepts (fields/topics) attached by OpenAlex
- Cited by
-
1Total citation count in OpenAlex
- Citations by year (recent)
-
2024: 1Per-year citation counts (last 5 years)
- Related works (count)
-
10Other works algorithmically related by OpenAlex
Full payload
| id | https://openalex.org/W4389072448 |
|---|---|
| doi | https://doi.org/10.48550/arxiv.2311.14029 |
| ids.doi | https://doi.org/10.48550/arxiv.2311.14029 |
| ids.openalex | https://openalex.org/W4389072448 |
| fwci | |
| type | preprint |
| title | Understanding the Vulnerability of CLIP to Image Compression |
| biblio.issue | |
| biblio.volume | |
| biblio.last_page | |
| biblio.first_page | |
| topics[0].id | https://openalex.org/T12357 |
| topics[0].field.id | https://openalex.org/fields/17 |
| topics[0].field.display_name | Computer Science |
| topics[0].score | 0.9782000184059143 |
| topics[0].domain.id | https://openalex.org/domains/3 |
| topics[0].domain.display_name | Physical Sciences |
| topics[0].subfield.id | https://openalex.org/subfields/1707 |
| topics[0].subfield.display_name | Computer Vision and Pattern Recognition |
| topics[0].display_name | Digital Media Forensic Detection |
| topics[1].id | https://openalex.org/T11105 |
| topics[1].field.id | https://openalex.org/fields/17 |
| topics[1].field.display_name | Computer Science |
| topics[1].score | 0.9351000189781189 |
| topics[1].domain.id | https://openalex.org/domains/3 |
| topics[1].domain.display_name | Physical Sciences |
| topics[1].subfield.id | https://openalex.org/subfields/1707 |
| topics[1].subfield.display_name | Computer Vision and Pattern Recognition |
| topics[1].display_name | Advanced Image Processing Techniques |
| topics[2].id | https://openalex.org/T10522 |
| topics[2].field.id | https://openalex.org/fields/27 |
| topics[2].field.display_name | Medicine |
| topics[2].score | 0.9293000102043152 |
| topics[2].domain.id | https://openalex.org/domains/4 |
| topics[2].domain.display_name | Health Sciences |
| topics[2].subfield.id | https://openalex.org/subfields/2741 |
| topics[2].subfield.display_name | Radiology, Nuclear Medicine and Imaging |
| topics[2].display_name | Medical Imaging Techniques and Applications |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| concepts[0].id | https://openalex.org/C41008148 |
| concepts[0].level | 0 |
| concepts[0].score | 0.7695344686508179 |
| concepts[0].wikidata | https://www.wikidata.org/wiki/Q21198 |
| concepts[0].display_name | Computer science |
| concepts[1].id | https://openalex.org/C63479239 |
| concepts[1].level | 3 |
| concepts[1].score | 0.643798291683197 |
| concepts[1].wikidata | https://www.wikidata.org/wiki/Q7353546 |
| concepts[1].display_name | Robustness (evolution) |
| concepts[2].id | https://openalex.org/C95713431 |
| concepts[2].level | 2 |
| concepts[2].score | 0.5888738632202148 |
| concepts[2].wikidata | https://www.wikidata.org/wiki/Q631425 |
| concepts[2].display_name | Vulnerability (computing) |
| concepts[3].id | https://openalex.org/C154945302 |
| concepts[3].level | 1 |
| concepts[3].score | 0.5737590789794922 |
| concepts[3].wikidata | https://www.wikidata.org/wiki/Q11660 |
| concepts[3].display_name | Artificial intelligence |
| concepts[4].id | https://openalex.org/C115961682 |
| concepts[4].level | 2 |
| concepts[4].score | 0.534054160118103 |
| concepts[4].wikidata | https://www.wikidata.org/wiki/Q860623 |
| concepts[4].display_name | Image (mathematics) |
| concepts[5].id | https://openalex.org/C143299363 |
| concepts[5].level | 2 |
| concepts[5].score | 0.49379023909568787 |
| concepts[5].wikidata | https://www.wikidata.org/wiki/Q900584 |
| concepts[5].display_name | Attribution |
| concepts[6].id | https://openalex.org/C31972630 |
| concepts[6].level | 1 |
| concepts[6].score | 0.4936846196651459 |
| concepts[6].wikidata | https://www.wikidata.org/wiki/Q844240 |
| concepts[6].display_name | Computer vision |
| concepts[7].id | https://openalex.org/C55020928 |
| concepts[7].level | 3 |
| concepts[7].score | 0.432275652885437 |
| concepts[7].wikidata | https://www.wikidata.org/wiki/Q3813865 |
| concepts[7].display_name | Image quality |
| concepts[8].id | https://openalex.org/C38652104 |
| concepts[8].level | 1 |
| concepts[8].score | 0.11039021611213684 |
| concepts[8].wikidata | https://www.wikidata.org/wiki/Q3510521 |
| concepts[8].display_name | Computer security |
| concepts[9].id | https://openalex.org/C15744967 |
| concepts[9].level | 0 |
| concepts[9].score | 0.0860186517238617 |
| concepts[9].wikidata | https://www.wikidata.org/wiki/Q9418 |
| concepts[9].display_name | Psychology |
| concepts[10].id | https://openalex.org/C185592680 |
| concepts[10].level | 0 |
| concepts[10].score | 0.0 |
| concepts[10].wikidata | https://www.wikidata.org/wiki/Q2329 |
| concepts[10].display_name | Chemistry |
| concepts[11].id | https://openalex.org/C77805123 |
| concepts[11].level | 1 |
| concepts[11].score | 0.0 |
| concepts[11].wikidata | https://www.wikidata.org/wiki/Q161272 |
| concepts[11].display_name | Social psychology |
| concepts[12].id | https://openalex.org/C104317684 |
| concepts[12].level | 2 |
| concepts[12].score | 0.0 |
| concepts[12].wikidata | https://www.wikidata.org/wiki/Q7187 |
| concepts[12].display_name | Gene |
| concepts[13].id | https://openalex.org/C55493867 |
| concepts[13].level | 1 |
| concepts[13].score | 0.0 |
| concepts[13].wikidata | https://www.wikidata.org/wiki/Q7094 |
| concepts[13].display_name | Biochemistry |
| keywords[0].id | https://openalex.org/keywords/computer-science |
| keywords[0].score | 0.7695344686508179 |
| keywords[0].display_name | Computer science |
| keywords[1].id | https://openalex.org/keywords/robustness |
| keywords[1].score | 0.643798291683197 |
| keywords[1].display_name | Robustness (evolution) |
| keywords[2].id | https://openalex.org/keywords/vulnerability |
| keywords[2].score | 0.5888738632202148 |
| keywords[2].display_name | Vulnerability (computing) |
| keywords[3].id | https://openalex.org/keywords/artificial-intelligence |
| keywords[3].score | 0.5737590789794922 |
| keywords[3].display_name | Artificial intelligence |
| keywords[4].id | https://openalex.org/keywords/image |
| keywords[4].score | 0.534054160118103 |
| keywords[4].display_name | Image (mathematics) |
| keywords[5].id | https://openalex.org/keywords/attribution |
| keywords[5].score | 0.49379023909568787 |
| keywords[5].display_name | Attribution |
| keywords[6].id | https://openalex.org/keywords/computer-vision |
| keywords[6].score | 0.4936846196651459 |
| keywords[6].display_name | Computer vision |
| keywords[7].id | https://openalex.org/keywords/image-quality |
| keywords[7].score | 0.432275652885437 |
| keywords[7].display_name | Image quality |
| keywords[8].id | https://openalex.org/keywords/computer-security |
| keywords[8].score | 0.11039021611213684 |
| keywords[8].display_name | Computer security |
| keywords[9].id | https://openalex.org/keywords/psychology |
| keywords[9].score | 0.0860186517238617 |
| keywords[9].display_name | Psychology |
| language | en |
| locations[0].id | pmh:oai:arXiv.org:2311.14029 |
| locations[0].is_oa | True |
| locations[0].source.id | https://openalex.org/S4306400194 |
| locations[0].source.issn | |
| locations[0].source.type | repository |
| locations[0].source.is_oa | True |
| locations[0].source.issn_l | |
| locations[0].source.is_core | False |
| locations[0].source.is_in_doaj | False |
| locations[0].source.display_name | arXiv (Cornell University) |
| locations[0].source.host_organization | https://openalex.org/I205783295 |
| locations[0].source.host_organization_name | Cornell University |
| locations[0].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[0].license | |
| locations[0].pdf_url | https://arxiv.org/pdf/2311.14029 |
| locations[0].version | submittedVersion |
| locations[0].raw_type | text |
| locations[0].license_id | |
| locations[0].is_accepted | False |
| locations[0].is_published | False |
| locations[0].raw_source_name | |
| locations[0].landing_page_url | http://arxiv.org/abs/2311.14029 |
| locations[1].id | doi:10.48550/arxiv.2311.14029 |
| locations[1].is_oa | True |
| locations[1].source.id | https://openalex.org/S4306400194 |
| locations[1].source.issn | |
| locations[1].source.type | repository |
| locations[1].source.is_oa | True |
| locations[1].source.issn_l | |
| locations[1].source.is_core | False |
| locations[1].source.is_in_doaj | False |
| locations[1].source.display_name | arXiv (Cornell University) |
| locations[1].source.host_organization | https://openalex.org/I205783295 |
| locations[1].source.host_organization_name | Cornell University |
| locations[1].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[1].license | cc-by |
| locations[1].pdf_url | |
| locations[1].version | |
| locations[1].raw_type | article |
| locations[1].license_id | https://openalex.org/licenses/cc-by |
| locations[1].is_accepted | False |
| locations[1].is_published | |
| locations[1].raw_source_name | |
| locations[1].landing_page_url | https://doi.org/10.48550/arxiv.2311.14029 |
| indexed_in | arxiv, datacite |
| authorships[0].author.id | https://openalex.org/A5072448065 |
| authorships[0].author.orcid | https://orcid.org/0009-0004-5365-2800 |
| authorships[0].author.display_name | Cangxiong Chen |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Chen, Cangxiong |
| authorships[0].is_corresponding | False |
| authorships[1].author.id | https://openalex.org/A5007109424 |
| authorships[1].author.orcid | https://orcid.org/0000-0001-5262-9722 |
| authorships[1].author.display_name | Vinay P. Namboodiri |
| authorships[1].author_position | middle |
| authorships[1].raw_author_name | Namboodiri, Vinay P. |
| authorships[1].is_corresponding | False |
| authorships[2].author.id | https://openalex.org/A5005596564 |
| authorships[2].author.orcid | https://orcid.org/0000-0003-1314-2094 |
| authorships[2].author.display_name | Julián Padget |
| authorships[2].author_position | last |
| authorships[2].raw_author_name | Padget, Julian |
| authorships[2].is_corresponding | False |
| has_content.pdf | False |
| has_content.grobid_xml | False |
| is_paratext | False |
| open_access.is_oa | True |
| open_access.oa_url | https://arxiv.org/pdf/2311.14029 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2025-10-10T00:00:00 |
| display_name | Understanding the Vulnerability of CLIP to Image Compression |
| has_fulltext | False |
| is_retracted | False |
| updated_date | 2025-11-06T06:51:31.235846 |
| primary_topic.id | https://openalex.org/T12357 |
| primary_topic.field.id | https://openalex.org/fields/17 |
| primary_topic.field.display_name | Computer Science |
| primary_topic.score | 0.9782000184059143 |
| primary_topic.domain.id | https://openalex.org/domains/3 |
| primary_topic.domain.display_name | Physical Sciences |
| primary_topic.subfield.id | https://openalex.org/subfields/1707 |
| primary_topic.subfield.display_name | Computer Vision and Pattern Recognition |
| primary_topic.display_name | Digital Media Forensic Detection |
| related_works | https://openalex.org/W2035546108, https://openalex.org/W2376361520, https://openalex.org/W2133328864, https://openalex.org/W2093949997, https://openalex.org/W2570200690, https://openalex.org/W2389726244, https://openalex.org/W3030478661, https://openalex.org/W2323536476, https://openalex.org/W2104624653, https://openalex.org/W2128730003 |
| cited_by_count | 1 |
| counts_by_year[0].year | 2024 |
| counts_by_year[0].cited_by_count | 1 |
| locations_count | 2 |
| best_oa_location.id | pmh:oai:arXiv.org:2311.14029 |
| best_oa_location.is_oa | True |
| best_oa_location.source.id | https://openalex.org/S4306400194 |
| best_oa_location.source.issn | |
| best_oa_location.source.type | repository |
| best_oa_location.source.is_oa | True |
| best_oa_location.source.issn_l | |
| best_oa_location.source.is_core | False |
| best_oa_location.source.is_in_doaj | False |
| best_oa_location.source.display_name | arXiv (Cornell University) |
| best_oa_location.source.host_organization | https://openalex.org/I205783295 |
| best_oa_location.source.host_organization_name | Cornell University |
| best_oa_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| best_oa_location.license | |
| best_oa_location.pdf_url | https://arxiv.org/pdf/2311.14029 |
| best_oa_location.version | submittedVersion |
| best_oa_location.raw_type | text |
| best_oa_location.license_id | |
| best_oa_location.is_accepted | False |
| best_oa_location.is_published | False |
| best_oa_location.raw_source_name | |
| best_oa_location.landing_page_url | http://arxiv.org/abs/2311.14029 |
| primary_location.id | pmh:oai:arXiv.org:2311.14029 |
| primary_location.is_oa | True |
| primary_location.source.id | https://openalex.org/S4306400194 |
| primary_location.source.issn | |
| primary_location.source.type | repository |
| primary_location.source.is_oa | True |
| primary_location.source.issn_l | |
| primary_location.source.is_core | False |
| primary_location.source.is_in_doaj | False |
| primary_location.source.display_name | arXiv (Cornell University) |
| primary_location.source.host_organization | https://openalex.org/I205783295 |
| primary_location.source.host_organization_name | Cornell University |
| primary_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| primary_location.license | |
| primary_location.pdf_url | https://arxiv.org/pdf/2311.14029 |
| primary_location.version | submittedVersion |
| primary_location.raw_type | text |
| primary_location.license_id | |
| primary_location.is_accepted | False |
| primary_location.is_published | False |
| primary_location.raw_source_name | |
| primary_location.landing_page_url | http://arxiv.org/abs/2311.14029 |
| publication_date | 2023-11-23 |
| publication_year | 2023 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 2 |
| abstract_inverted_index.We | 20, 73 |
| abstract_inverted_index.an | 40 |
| abstract_inverted_index.in | 28, 61 |
| abstract_inverted_index.is | 1, 9, 24, 36 |
| abstract_inverted_index.of | 70, 90, 104 |
| abstract_inverted_index.on | 77 |
| abstract_inverted_index.to | 26, 51, 86, 100 |
| abstract_inverted_index.us | 95 |
| abstract_inverted_index.we | 48 |
| abstract_inverted_index.Our | 81 |
| abstract_inverted_index.and | 15, 56, 79, 92, 106 |
| abstract_inverted_index.are | 49 |
| abstract_inverted_index.can | 93 |
| abstract_inverted_index.for | 11 |
| abstract_inverted_index.the | 59, 63, 66, 84, 102 |
| abstract_inverted_index.CLIP | 0, 23, 91, 105 |
| abstract_inverted_index.This | 33 |
| abstract_inverted_index.able | 50 |
| abstract_inverted_index.both | 54 |
| abstract_inverted_index.help | 94 |
| abstract_inverted_index.more | 97 |
| abstract_inverted_index.that | 8, 22 |
| abstract_inverted_index.this | 45, 71, 75, 88 |
| abstract_inverted_index.used | 4, 10 |
| abstract_inverted_index.work | 82 |
| abstract_inverted_index.Using | 44 |
| abstract_inverted_index.basis | 85 |
| abstract_inverted_index.image | 13, 29 |
| abstract_inverted_index.model | 7 |
| abstract_inverted_index.other | 16, 107 |
| abstract_inverted_index.under | 31 |
| abstract_inverted_index.using | 39 |
| abstract_inverted_index.which | 62 |
| abstract_inverted_index.better | 52 |
| abstract_inverted_index.change | 27 |
| abstract_inverted_index.model. | 72 |
| abstract_inverted_index.nature | 60 |
| abstract_inverted_index.result | 35 |
| abstract_inverted_index.tasks. | 19 |
| abstract_inverted_index.widely | 3 |
| abstract_inverted_index.STL-10. | 80 |
| abstract_inverted_index.affects | 65 |
| abstract_inverted_index.develop | 96 |
| abstract_inverted_index.exactly | 58 |
| abstract_inverted_index.further | 37 |
| abstract_inverted_index.improve | 101 |
| abstract_inverted_index.method, | 47 |
| abstract_inverted_index.methods | 99 |
| abstract_inverted_index.models. | 109 |
| abstract_inverted_index.quality | 30 |
| abstract_inverted_index.CIFAR-10 | 78 |
| abstract_inverted_index.accuracy | 69 |
| abstract_inverted_index.analysed | 38 |
| abstract_inverted_index.evaluate | 74 |
| abstract_inverted_index.provides | 83 |
| abstract_inverted_index.alignment | 18 |
| abstract_inverted_index.effective | 98 |
| abstract_inverted_index.zero-shot | 12, 67 |
| abstract_inverted_index.Gradients. | 43 |
| abstract_inverted_index.image-text | 17 |
| abstract_inverted_index.robustness | 103 |
| abstract_inverted_index.surprising | 34 |
| abstract_inverted_index.understand | 53, 87 |
| abstract_inverted_index.vulnerable | 25 |
| abstract_inverted_index.attribution | 41, 46 |
| abstract_inverted_index.compression | 64 |
| abstract_inverted_index.demonstrate | 21 |
| abstract_inverted_index.extensively | 76 |
| abstract_inverted_index.recognition | 14, 68 |
| abstract_inverted_index.compression. | 32 |
| abstract_inverted_index.foundational | 5 |
| abstract_inverted_index.qualitatively | 57 |
| abstract_inverted_index.vulnerability | 89 |
| abstract_inverted_index.quantitatively | 55 |
| abstract_inverted_index.vision-language | 6, 108 |
| abstract_inverted_index.method-Integrated | 42 |
| cited_by_percentile_year | |
| countries_distinct_count | 0 |
| institutions_distinct_count | 3 |
| sustainable_development_goals[0].id | https://metadata.un.org/sdg/10 |
| sustainable_development_goals[0].score | 0.5600000023841858 |
| sustainable_development_goals[0].display_name | Reduced inequalities |
| citation_normalized_percentile |