{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:26UDCRM7IYHGQVLC7FYRBDUR4F","short_pith_number":"pith:26UDCRM7","schema_version":"1.0","canonical_sha256":"d7a831459f460e685562f971108e91e1428a37e0678b1534ab334a7c672e3729","source":{"kind":"arxiv","id":"2404.12599","version":2},"attestation_state":"computed","paper":{"title":"QUTE: Quantifying Uncertainty in TinyML with Early-exit-assisted ensembles for model-monitoring","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Nikhil P Ghanathe, Steven J E Wilton","submitted_at":"2024-04-19T03:06:50Z","abstract_excerpt":"Uncertainty quantification (UQ) provides a resource-efficient solution for on-device monitoring of tinyML models deployed without access to true labels. However, existing UQ methods impose significant memory and compute demands, making them impractical for ultra-low-power, KB-sized TinyML devices. Prior work has attempted to reduce overhead by using early-exit ensembles to quantify uncertainty in a single forward pass, but these approaches still carry prohibitive costs. To address this, we propose QUTE, a novel resource-efficient early-exit-assisted ensemble architecture optimized for tinyML m"},"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":"2404.12599","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-19T03:06:50Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"3df2da8717b1755ca621051dd0e3621781e8799067fee8d0f73dc001fc55590e","abstract_canon_sha256":"70ea2979dc1297bac03beb8ffd5b6192ee7e694b216dcb3a28e648582bc93bca"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:36:10.948588Z","signature_b64":"44ePFEY0nTmE1Zf9W7ltBmja2Fleks67aY6BdY22Qia8MswbcvKr1jJVu389nXJTsHjHA/Yh70iCBzXUOAXXAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7a831459f460e685562f971108e91e1428a37e0678b1534ab334a7c672e3729","last_reissued_at":"2026-07-05T09:36:10.948177Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:36:10.948177Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"QUTE: Quantifying Uncertainty in TinyML with Early-exit-assisted ensembles for model-monitoring","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Nikhil P Ghanathe, Steven J E Wilton","submitted_at":"2024-04-19T03:06:50Z","abstract_excerpt":"Uncertainty quantification (UQ) provides a resource-efficient solution for on-device monitoring of tinyML models deployed without access to true labels. However, existing UQ methods impose significant memory and compute demands, making them impractical for ultra-low-power, KB-sized TinyML devices. Prior work has attempted to reduce overhead by using early-exit ensembles to quantify uncertainty in a single forward pass, but these approaches still carry prohibitive costs. To address this, we propose QUTE, a novel resource-efficient early-exit-assisted ensemble architecture optimized for tinyML m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.12599","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/2404.12599/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":"2404.12599","created_at":"2026-07-05T09:36:10.948238+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.12599v2","created_at":"2026-07-05T09:36:10.948238+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.12599","created_at":"2026-07-05T09:36:10.948238+00:00"},{"alias_kind":"pith_short_12","alias_value":"26UDCRM7IYHG","created_at":"2026-07-05T09:36:10.948238+00:00"},{"alias_kind":"pith_short_16","alias_value":"26UDCRM7IYHGQVLC","created_at":"2026-07-05T09:36:10.948238+00:00"},{"alias_kind":"pith_short_8","alias_value":"26UDCRM7","created_at":"2026-07-05T09:36:10.948238+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":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F","json":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F.json","graph_json":"https://pith.science/api/pith-number/26UDCRM7IYHGQVLC7FYRBDUR4F/graph.json","events_json":"https://pith.science/api/pith-number/26UDCRM7IYHGQVLC7FYRBDUR4F/events.json","paper":"https://pith.science/paper/26UDCRM7"},"agent_actions":{"view_html":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F","download_json":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F.json","view_paper":"https://pith.science/paper/26UDCRM7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.12599&json=true","fetch_graph":"https://pith.science/api/pith-number/26UDCRM7IYHGQVLC7FYRBDUR4F/graph.json","fetch_events":"https://pith.science/api/pith-number/26UDCRM7IYHGQVLC7FYRBDUR4F/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F/action/timestamp_anchor","attest_storage":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F/action/storage_attestation","attest_author":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F/action/author_attestation","sign_citation":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F/action/citation_signature","submit_replication":"https://pith.science/pith/26UDCRM7IYHGQVLC7FYRBDUR4F/action/replication_record"}},"created_at":"2026-07-05T09:36:10.948238+00:00","updated_at":"2026-07-05T09:36:10.948238+00:00"}