{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:JOWYC7KCYVPBXHB5FX4VDQYWAF","short_pith_number":"pith:JOWYC7KC","schema_version":"1.0","canonical_sha256":"4bad817d42c55e1b9c3d2df951c31601634737f52e4ff2a2b3b45a5c966ae8f6","source":{"kind":"arxiv","id":"2106.06998","version":4},"attestation_state":"computed","paper":{"title":"XConv: Low-memory stochastic backpropagation for convolutional layers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NA","math.NA"],"primary_cat":"cs.LG","authors_text":"Ali Siahkoohi, Anirudh Thatipelli, Felix J. Herrmann, Jeffrey Sam, Mathias Louboutin, Rongrong Wang","submitted_at":"2021-06-13T13:54:02Z","abstract_excerpt":"Training convolutional neural networks at scale demands substantial memory, largely because intermediate activations must be stored for backpropagation. Existing remedies (checkpointing, invertible architectures, or gradient-approximation methods such as randomized automatic differentiation) either add significant computation, impose architectural constraints, or require non-trivial code changes. We propose XConv, a near-drop-in replacement for standard 2D and 3D convolutional layers that addresses all three: it preserves standard backpropagation, imposes no architectural constraints, and inte"},"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":"2106.06998","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-06-13T13:54:02Z","cross_cats_sorted":["cs.NA","math.NA"],"title_canon_sha256":"b4e35fde61d6b3ae5f6541b48d74821cd7a408902dd0889b78084084aed0cf64","abstract_canon_sha256":"75aec6179d91e96c6d59ef58fada4eb8b5f123923d4d5a7ea6602a01871e4db0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:14:26.642919Z","signature_b64":"kM9UgJjJP2R8o16j1RlqUcOQYyrdBrPXL/YCSFvr3P4s2ZqeJJkxZQuY50Sxe4mV1FDiLbq02R7T9eLa/d8KDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4bad817d42c55e1b9c3d2df951c31601634737f52e4ff2a2b3b45a5c966ae8f6","last_reissued_at":"2026-06-23T03:14:26.642402Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:14:26.642402Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"XConv: Low-memory stochastic backpropagation for convolutional layers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.NA","math.NA"],"primary_cat":"cs.LG","authors_text":"Ali Siahkoohi, Anirudh Thatipelli, Felix J. Herrmann, Jeffrey Sam, Mathias Louboutin, Rongrong Wang","submitted_at":"2021-06-13T13:54:02Z","abstract_excerpt":"Training convolutional neural networks at scale demands substantial memory, largely because intermediate activations must be stored for backpropagation. Existing remedies (checkpointing, invertible architectures, or gradient-approximation methods such as randomized automatic differentiation) either add significant computation, impose architectural constraints, or require non-trivial code changes. We propose XConv, a near-drop-in replacement for standard 2D and 3D convolutional layers that addresses all three: it preserves standard backpropagation, imposes no architectural constraints, and inte"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.06998","kind":"arxiv","version":4},"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/2106.06998/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":"2106.06998","created_at":"2026-06-23T03:14:26.642480+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.06998v4","created_at":"2026-06-23T03:14:26.642480+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.06998","created_at":"2026-06-23T03:14:26.642480+00:00"},{"alias_kind":"pith_short_12","alias_value":"JOWYC7KCYVPB","created_at":"2026-06-23T03:14:26.642480+00:00"},{"alias_kind":"pith_short_16","alias_value":"JOWYC7KCYVPBXHB5","created_at":"2026-06-23T03:14:26.642480+00:00"},{"alias_kind":"pith_short_8","alias_value":"JOWYC7KC","created_at":"2026-06-23T03:14:26.642480+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF","json":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF.json","graph_json":"https://pith.science/api/pith-number/JOWYC7KCYVPBXHB5FX4VDQYWAF/graph.json","events_json":"https://pith.science/api/pith-number/JOWYC7KCYVPBXHB5FX4VDQYWAF/events.json","paper":"https://pith.science/paper/JOWYC7KC"},"agent_actions":{"view_html":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF","download_json":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF.json","view_paper":"https://pith.science/paper/JOWYC7KC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.06998&json=true","fetch_graph":"https://pith.science/api/pith-number/JOWYC7KCYVPBXHB5FX4VDQYWAF/graph.json","fetch_events":"https://pith.science/api/pith-number/JOWYC7KCYVPBXHB5FX4VDQYWAF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF/action/storage_attestation","attest_author":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF/action/author_attestation","sign_citation":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF/action/citation_signature","submit_replication":"https://pith.science/pith/JOWYC7KCYVPBXHB5FX4VDQYWAF/action/replication_record"}},"created_at":"2026-06-23T03:14:26.642480+00:00","updated_at":"2026-06-23T03:14:26.642480+00:00"}