{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:ITVN3KSEXA6PBH6ZFIPEJQQOS6","short_pith_number":"pith:ITVN3KSE","schema_version":"1.0","canonical_sha256":"44eaddaa44b83cf09fd92a1e44c20e97940f5690d2fcdb6faaf9b6fe1cd42da6","source":{"kind":"arxiv","id":"2306.00817","version":2},"attestation_state":"computed","paper":{"title":"Dilated Convolution with Learnable Spacings: beyond bilinear interpolation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ismail Khalfaoui-Hassani, Thomas Pellegrini, Timoth\\'ee Masquelier","submitted_at":"2023-06-01T15:42:08Z","abstract_excerpt":"Dilated Convolution with Learnable Spacings (DCLS) is a recently proposed variation of the dilated convolution in which the spacings between the non-zero elements in the kernel, or equivalently their positions, are learnable. Non-integer positions are handled via interpolation. Thanks to this trick, positions have well-defined gradients. The original DCLS used bilinear interpolation, and thus only considered the four nearest pixels. Yet here we show that longer range interpolations, and in particular a Gaussian interpolation, allow improving performance on ImageNet1k classification on two stat"},"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":"2306.00817","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-06-01T15:42:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"26680297cbd56adeffd975b675f8b122c3bfa3ed66e0123b9337f50faecb48bf","abstract_canon_sha256":"b3cc304e7f6ae5fa1d5b1a011fd76ecf83712d4bc66a9aeb9bd9f202f644ab51"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:53:38.745838Z","signature_b64":"uah7DwcAy+7WvTOjsSzABAyvqqLSTfOE7BbQsunkUt/4sLO4udW0W20NUgxQbyxJu3pYhmeFcYOfmBntR7XCBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"44eaddaa44b83cf09fd92a1e44c20e97940f5690d2fcdb6faaf9b6fe1cd42da6","last_reissued_at":"2026-07-05T06:53:38.745367Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:53:38.745367Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Dilated Convolution with Learnable Spacings: beyond bilinear interpolation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Ismail Khalfaoui-Hassani, Thomas Pellegrini, Timoth\\'ee Masquelier","submitted_at":"2023-06-01T15:42:08Z","abstract_excerpt":"Dilated Convolution with Learnable Spacings (DCLS) is a recently proposed variation of the dilated convolution in which the spacings between the non-zero elements in the kernel, or equivalently their positions, are learnable. Non-integer positions are handled via interpolation. Thanks to this trick, positions have well-defined gradients. The original DCLS used bilinear interpolation, and thus only considered the four nearest pixels. Yet here we show that longer range interpolations, and in particular a Gaussian interpolation, allow improving performance on ImageNet1k classification on two stat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.00817","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/2306.00817/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":"2306.00817","created_at":"2026-07-05T06:53:38.745429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2306.00817v2","created_at":"2026-07-05T06:53:38.745429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.00817","created_at":"2026-07-05T06:53:38.745429+00:00"},{"alias_kind":"pith_short_12","alias_value":"ITVN3KSEXA6P","created_at":"2026-07-05T06:53:38.745429+00:00"},{"alias_kind":"pith_short_16","alias_value":"ITVN3KSEXA6PBH6Z","created_at":"2026-07-05T06:53:38.745429+00:00"},{"alias_kind":"pith_short_8","alias_value":"ITVN3KSE","created_at":"2026-07-05T06:53:38.745429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2507.01829","citing_title":"mGRADE: Minimal Recurrent Gating Meets Delay Convolutions for Lightweight Sequence Modeling","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6","json":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6.json","graph_json":"https://pith.science/api/pith-number/ITVN3KSEXA6PBH6ZFIPEJQQOS6/graph.json","events_json":"https://pith.science/api/pith-number/ITVN3KSEXA6PBH6ZFIPEJQQOS6/events.json","paper":"https://pith.science/paper/ITVN3KSE"},"agent_actions":{"view_html":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6","download_json":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6.json","view_paper":"https://pith.science/paper/ITVN3KSE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2306.00817&json=true","fetch_graph":"https://pith.science/api/pith-number/ITVN3KSEXA6PBH6ZFIPEJQQOS6/graph.json","fetch_events":"https://pith.science/api/pith-number/ITVN3KSEXA6PBH6ZFIPEJQQOS6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6/action/storage_attestation","attest_author":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6/action/author_attestation","sign_citation":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6/action/citation_signature","submit_replication":"https://pith.science/pith/ITVN3KSEXA6PBH6ZFIPEJQQOS6/action/replication_record"}},"created_at":"2026-07-05T06:53:38.745429+00:00","updated_at":"2026-07-05T06:53:38.745429+00:00"}