{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GUBF6KXYUG72XGWKERQ6GHCT3W","short_pith_number":"pith:GUBF6KXY","schema_version":"1.0","canonical_sha256":"35025f2af8a1bfab9aca2461e31c53dd8e861a42958aef41080e28ac31459d7d","source":{"kind":"arxiv","id":"2505.05086","version":2},"attestation_state":"computed","paper":{"title":"Beyond Low-rank Decomposition: A Shortcut Approach for Efficient On-Device Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ael Quelennec, Enzo Tartaglione, Le-Trung Nguyen, Van-Tam Nguyen","submitted_at":"2025-05-08T09:34:15Z","abstract_excerpt":"On-device learning has emerged as a promising direction for AI development, particularly because of its potential to reduce latency issues and mitigate privacy risks associated with device-server communication, while improving energy efficiency. Despite these advantages, significant memory and computational constraints still represent major challenges for its deployment. Drawing on previous studies on low-rank decomposition methods that address activation memory bottlenecks in backpropagation, we propose a novel shortcut approach as an alternative. Our analysis and experiments demonstrate that"},"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":"2505.05086","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-08T09:34:15Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0e9428de76102133c8a2858da79edd12e46241aafdeeb4dfa5eef461adde8659","abstract_canon_sha256":"34145d58aeb03eaa5395c7721e8b3111e8a2b0c343dc0d203ef5adf0cfba4f11"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:42:29.375129Z","signature_b64":"ji/90cJylUMPfMSqUM9v/Rt31c1mca1HnNr0qg3zhXx6tL1Yg+J3wahDV/NDR5hGKpJ7xFYXSPmaWlH/iFD4AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"35025f2af8a1bfab9aca2461e31c53dd8e861a42958aef41080e28ac31459d7d","last_reissued_at":"2026-07-05T11:42:29.374620Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:42:29.374620Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Beyond Low-rank Decomposition: A Shortcut Approach for Efficient On-Device Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ael Quelennec, Enzo Tartaglione, Le-Trung Nguyen, Van-Tam Nguyen","submitted_at":"2025-05-08T09:34:15Z","abstract_excerpt":"On-device learning has emerged as a promising direction for AI development, particularly because of its potential to reduce latency issues and mitigate privacy risks associated with device-server communication, while improving energy efficiency. Despite these advantages, significant memory and computational constraints still represent major challenges for its deployment. Drawing on previous studies on low-rank decomposition methods that address activation memory bottlenecks in backpropagation, we propose a novel shortcut approach as an alternative. Our analysis and experiments demonstrate that"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.05086","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/2505.05086/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":"2505.05086","created_at":"2026-07-05T11:42:29.374685+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.05086v2","created_at":"2026-07-05T11:42:29.374685+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.05086","created_at":"2026-07-05T11:42:29.374685+00:00"},{"alias_kind":"pith_short_12","alias_value":"GUBF6KXYUG72","created_at":"2026-07-05T11:42:29.374685+00:00"},{"alias_kind":"pith_short_16","alias_value":"GUBF6KXYUG72XGWK","created_at":"2026-07-05T11:42:29.374685+00:00"},{"alias_kind":"pith_short_8","alias_value":"GUBF6KXY","created_at":"2026-07-05T11:42:29.374685+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.07693","citing_title":"CosmosAlign: Adapting a World Foundation Model for Generative Traffic Video Forecasting","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W","json":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W.json","graph_json":"https://pith.science/api/pith-number/GUBF6KXYUG72XGWKERQ6GHCT3W/graph.json","events_json":"https://pith.science/api/pith-number/GUBF6KXYUG72XGWKERQ6GHCT3W/events.json","paper":"https://pith.science/paper/GUBF6KXY"},"agent_actions":{"view_html":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W","download_json":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W.json","view_paper":"https://pith.science/paper/GUBF6KXY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.05086&json=true","fetch_graph":"https://pith.science/api/pith-number/GUBF6KXYUG72XGWKERQ6GHCT3W/graph.json","fetch_events":"https://pith.science/api/pith-number/GUBF6KXYUG72XGWKERQ6GHCT3W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W/action/storage_attestation","attest_author":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W/action/author_attestation","sign_citation":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W/action/citation_signature","submit_replication":"https://pith.science/pith/GUBF6KXYUG72XGWKERQ6GHCT3W/action/replication_record"}},"created_at":"2026-07-05T11:42:29.374685+00:00","updated_at":"2026-07-05T11:42:29.374685+00:00"}