{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:2QKAH5L63EOZXCSEHCDVHPIOIY","short_pith_number":"pith:2QKAH5L6","schema_version":"1.0","canonical_sha256":"d41403f57ed91d9b8a44388753bd0e463c612fd775612ec02616e0d9602546d1","source":{"kind":"arxiv","id":"2001.08248","version":1},"attestation_state":"computed","paper":{"title":"How Much Position Information Do Convolutional Neural Networks Encode?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Md Amirul Islam, Neil D. B. Bruce, Sen Jia","submitted_at":"2020-01-22T19:44:43Z","abstract_excerpt":"In contrast to fully connected networks, Convolutional Neural Networks (CNNs) achieve efficiency by learning weights associated with local filters with a finite spatial extent. An implication of this is that a filter may know what it is looking at, but not where it is positioned in the image. Information concerning absolute position is inherently useful, and it is reasonable to assume that deep CNNs may implicitly learn to encode this information if there is a means to do so. In this paper, we test this hypothesis revealing the surprising degree of absolute position information that is encoded"},"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":"2001.08248","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2020-01-22T19:44:43Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"35674fcf098b35c7356c69bf1ab6bd66d6b3c3276a81a49666c57288d0a5283b","abstract_canon_sha256":"70ffeedcc25f52fb94afb905d2ed383e6b9c76165b66107281ee203636b16c0f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:35:11.295923Z","signature_b64":"GS6fM1ERKjTrECZJ4zoz+0uALMEi7JT3vgqaD/1NuuDLHH5nzGnIUWUpH4p80E+MSBmnbNXoZpB7RoitHDP1Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d41403f57ed91d9b8a44388753bd0e463c612fd775612ec02616e0d9602546d1","last_reissued_at":"2026-07-05T00:35:11.295479Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:35:11.295479Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Much Position Information Do Convolutional Neural Networks Encode?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Md Amirul Islam, Neil D. B. Bruce, Sen Jia","submitted_at":"2020-01-22T19:44:43Z","abstract_excerpt":"In contrast to fully connected networks, Convolutional Neural Networks (CNNs) achieve efficiency by learning weights associated with local filters with a finite spatial extent. An implication of this is that a filter may know what it is looking at, but not where it is positioned in the image. Information concerning absolute position is inherently useful, and it is reasonable to assume that deep CNNs may implicitly learn to encode this information if there is a means to do so. In this paper, we test this hypothesis revealing the surprising degree of absolute position information that is encoded"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2001.08248","kind":"arxiv","version":1},"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/2001.08248/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":"2001.08248","created_at":"2026-07-05T00:35:11.295548+00:00"},{"alias_kind":"arxiv_version","alias_value":"2001.08248v1","created_at":"2026-07-05T00:35:11.295548+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2001.08248","created_at":"2026-07-05T00:35:11.295548+00:00"},{"alias_kind":"pith_short_12","alias_value":"2QKAH5L63EOZ","created_at":"2026-07-05T00:35:11.295548+00:00"},{"alias_kind":"pith_short_16","alias_value":"2QKAH5L63EOZXCSE","created_at":"2026-07-05T00:35:11.295548+00:00"},{"alias_kind":"pith_short_8","alias_value":"2QKAH5L6","created_at":"2026-07-05T00:35:11.295548+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27745","citing_title":"Panoramic Scene Understanding: A Survey from Distortion-Aware Engineering to Sphere-Native Modeling","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15684","citing_title":"ElasticDiT: Efficient Diffusion Transformers via Elastic Architecture and Sparse Attention for High-Resolution Image Generation on Mobile Devices","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2410.10629","citing_title":"SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12491","citing_title":"Elastic Attention Cores for Scalable Vision Transformers","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2104.09864","citing_title":"RoFormer: Enhanced Transformer with Rotary Position Embedding","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY","json":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY.json","graph_json":"https://pith.science/api/pith-number/2QKAH5L63EOZXCSEHCDVHPIOIY/graph.json","events_json":"https://pith.science/api/pith-number/2QKAH5L63EOZXCSEHCDVHPIOIY/events.json","paper":"https://pith.science/paper/2QKAH5L6"},"agent_actions":{"view_html":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY","download_json":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY.json","view_paper":"https://pith.science/paper/2QKAH5L6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2001.08248&json=true","fetch_graph":"https://pith.science/api/pith-number/2QKAH5L63EOZXCSEHCDVHPIOIY/graph.json","fetch_events":"https://pith.science/api/pith-number/2QKAH5L63EOZXCSEHCDVHPIOIY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY/action/storage_attestation","attest_author":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY/action/author_attestation","sign_citation":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY/action/citation_signature","submit_replication":"https://pith.science/pith/2QKAH5L63EOZXCSEHCDVHPIOIY/action/replication_record"}},"created_at":"2026-07-05T00:35:11.295548+00:00","updated_at":"2026-07-05T00:35:11.295548+00:00"}