{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:CYTZQ3YTQU4LP6SWL6OSNF4DPE","short_pith_number":"pith:CYTZQ3YT","schema_version":"1.0","canonical_sha256":"1627986f138538b7fa565f9d269783790e0f889b1448edf2fa340d496f9ba184","source":{"kind":"arxiv","id":"2305.07772","version":2},"attestation_state":"computed","paper":{"title":"Monitoring and Adapting ML Models on Mobile Devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Asaf Cidon, Chengzhi Mao, Junfeng Yang, Lauren Hong, Lingxiao Li, Nader Karayanni, Wei Hao, Zixi Wang","submitted_at":"2023-05-12T21:33:26Z","abstract_excerpt":"ML models are increasingly being pushed to mobile devices, for low-latency inference and offline operation. However, once the models are deployed, it is hard for ML operators to track their accuracy, which can degrade unpredictably (e.g., due to data drift). We design the first end-to-end system for continuously monitoring and adapting models on mobile devices without requiring feedback from users. Our key observation is that often model degradation is due to a specific root cause, which may affect a large group of devices. Therefore, once the system detects a consistent degradation across a l"},"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":"2305.07772","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-05-12T21:33:26Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"d662697d2c877f52fcaf7dd28cd15906886ddca0454fc58436711fdc9f9b933c","abstract_canon_sha256":"1590082b2df6d06465b608b9835196a93e55abdd8b6ff1a3340a0a64c29d249f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:11:05.307980Z","signature_b64":"tYcd6/Hn+KCbLKNBwK1FPyZIxg6+keung6Gcbx54rhe4rIfmZCXwSqqIxW+p1YLbhdICkzUn3g7ctHipl4ZBBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1627986f138538b7fa565f9d269783790e0f889b1448edf2fa340d496f9ba184","last_reissued_at":"2026-07-05T06:11:05.307564Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:11:05.307564Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Monitoring and Adapting ML Models on Mobile Devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Asaf Cidon, Chengzhi Mao, Junfeng Yang, Lauren Hong, Lingxiao Li, Nader Karayanni, Wei Hao, Zixi Wang","submitted_at":"2023-05-12T21:33:26Z","abstract_excerpt":"ML models are increasingly being pushed to mobile devices, for low-latency inference and offline operation. However, once the models are deployed, it is hard for ML operators to track their accuracy, which can degrade unpredictably (e.g., due to data drift). We design the first end-to-end system for continuously monitoring and adapting models on mobile devices without requiring feedback from users. Our key observation is that often model degradation is due to a specific root cause, which may affect a large group of devices. Therefore, once the system detects a consistent degradation across a l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.07772","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/2305.07772/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":"2305.07772","created_at":"2026-07-05T06:11:05.307620+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.07772v2","created_at":"2026-07-05T06:11:05.307620+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.07772","created_at":"2026-07-05T06:11:05.307620+00:00"},{"alias_kind":"pith_short_12","alias_value":"CYTZQ3YTQU4L","created_at":"2026-07-05T06:11:05.307620+00:00"},{"alias_kind":"pith_short_16","alias_value":"CYTZQ3YTQU4LP6SW","created_at":"2026-07-05T06:11:05.307620+00:00"},{"alias_kind":"pith_short_8","alias_value":"CYTZQ3YT","created_at":"2026-07-05T06:11:05.307620+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.10692","citing_title":"DEBUG-HD: Debugging TinyML models on-device using Hyper-Dimensional computing","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE","json":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE.json","graph_json":"https://pith.science/api/pith-number/CYTZQ3YTQU4LP6SWL6OSNF4DPE/graph.json","events_json":"https://pith.science/api/pith-number/CYTZQ3YTQU4LP6SWL6OSNF4DPE/events.json","paper":"https://pith.science/paper/CYTZQ3YT"},"agent_actions":{"view_html":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE","download_json":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE.json","view_paper":"https://pith.science/paper/CYTZQ3YT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.07772&json=true","fetch_graph":"https://pith.science/api/pith-number/CYTZQ3YTQU4LP6SWL6OSNF4DPE/graph.json","fetch_events":"https://pith.science/api/pith-number/CYTZQ3YTQU4LP6SWL6OSNF4DPE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE/action/storage_attestation","attest_author":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE/action/author_attestation","sign_citation":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE/action/citation_signature","submit_replication":"https://pith.science/pith/CYTZQ3YTQU4LP6SWL6OSNF4DPE/action/replication_record"}},"created_at":"2026-07-05T06:11:05.307620+00:00","updated_at":"2026-07-05T06:11:05.307620+00:00"}