{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:YVULIHPNN3522GYXATL2ZAZKHA","short_pith_number":"pith:YVULIHPN","schema_version":"1.0","canonical_sha256":"c568b41ded6efbad1b1704d7ac832a381dd83d7ab3c797e2e781d6c5c30b267f","source":{"kind":"arxiv","id":"2003.07064","version":2},"attestation_state":"computed","paper":{"title":"On Translation Invariance in CNNs: Convolutional Layers can Exploit Absolute Spatial Location","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.IV"],"primary_cat":"cs.CV","authors_text":"Jan C. van Gemert, Osman Semih Kayhan","submitted_at":"2020-03-16T08:00:06Z","abstract_excerpt":"In this paper we challenge the common assumption that convolutional layers in modern CNNs are translation invariant. We show that CNNs can and will exploit the absolute spatial location by learning filters that respond exclusively to particular absolute locations by exploiting image boundary effects. Because modern CNNs filters have a huge receptive field, these boundary effects operate even far from the image boundary, allowing the network to exploit absolute spatial location all over the image. We give a simple solution to remove spatial location encoding which improves translation invarianc"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2003.07064","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2020-03-16T08:00:06Z","cross_cats_sorted":["cs.LG","eess.IV"],"title_canon_sha256":"3cd1c30864fd3b681ff508dc2401a9d68ec881fa53eefb6f600272a405abe4d4","abstract_canon_sha256":"fba2500f19a6353ccdff33c953cbb0abc44671a88f7131d9001110ee62982fe5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:06:53.829186Z","signature_b64":"MwZCGwWPrjoJ9VoMBOzN0KsDWEVsTAkQnpzmzeoLVBJaycOFywFic3kWDJkrhUlKgljfBmHOzNFroctWfT16BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c568b41ded6efbad1b1704d7ac832a381dd83d7ab3c797e2e781d6c5c30b267f","last_reissued_at":"2026-07-05T01:06:53.828684Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:06:53.828684Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On Translation Invariance in CNNs: Convolutional Layers can Exploit Absolute Spatial Location","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.IV"],"primary_cat":"cs.CV","authors_text":"Jan C. van Gemert, Osman Semih Kayhan","submitted_at":"2020-03-16T08:00:06Z","abstract_excerpt":"In this paper we challenge the common assumption that convolutional layers in modern CNNs are translation invariant. We show that CNNs can and will exploit the absolute spatial location by learning filters that respond exclusively to particular absolute locations by exploiting image boundary effects. Because modern CNNs filters have a huge receptive field, these boundary effects operate even far from the image boundary, allowing the network to exploit absolute spatial location all over the image. We give a simple solution to remove spatial location encoding which improves translation invarianc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2003.07064","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2003.07064/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2003.07064","created_at":"2026-07-05T01:06:53.828739+00:00"},{"alias_kind":"arxiv_version","alias_value":"2003.07064v2","created_at":"2026-07-05T01:06:53.828739+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2003.07064","created_at":"2026-07-05T01:06:53.828739+00:00"},{"alias_kind":"pith_short_12","alias_value":"YVULIHPNN352","created_at":"2026-07-05T01:06:53.828739+00:00"},{"alias_kind":"pith_short_16","alias_value":"YVULIHPNN3522GYX","created_at":"2026-07-05T01:06:53.828739+00:00"},{"alias_kind":"pith_short_8","alias_value":"YVULIHPN","created_at":"2026-07-05T01:06:53.828739+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26894","citing_title":"Modeling Local, Global, and Cross-Modal Context in Multimodal 3D MRI","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA","json":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA.json","graph_json":"https://pith.science/api/pith-number/YVULIHPNN3522GYXATL2ZAZKHA/graph.json","events_json":"https://pith.science/api/pith-number/YVULIHPNN3522GYXATL2ZAZKHA/events.json","paper":"https://pith.science/paper/YVULIHPN"},"agent_actions":{"view_html":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA","download_json":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA.json","view_paper":"https://pith.science/paper/YVULIHPN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2003.07064&json=true","fetch_graph":"https://pith.science/api/pith-number/YVULIHPNN3522GYXATL2ZAZKHA/graph.json","fetch_events":"https://pith.science/api/pith-number/YVULIHPNN3522GYXATL2ZAZKHA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA/action/storage_attestation","attest_author":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA/action/author_attestation","sign_citation":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA/action/citation_signature","submit_replication":"https://pith.science/pith/YVULIHPNN3522GYXATL2ZAZKHA/action/replication_record"}},"created_at":"2026-07-05T01:06:53.828739+00:00","updated_at":"2026-07-05T01:06:53.828739+00:00"}