{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3VKGSFE4USNPQA2KY26RXAYH2S","short_pith_number":"pith:3VKGSFE4","schema_version":"1.0","canonical_sha256":"dd5469149ca49af8034ac6bd1b8307d48fc9fa724df61de6b17a16f86f186fc6","source":{"kind":"arxiv","id":"2507.07949","version":1},"attestation_state":"computed","paper":{"title":"TinierHAR: Towards Ultra-Lightweight Deep Learning Models for Efficient Human Activity Recognition on Edge Devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Daniel Geissler, Mengxi Liu, Paul Lukowicz, Sizhen Bian, Vitor Fortes Rey","submitted_at":"2025-07-10T17:33:52Z","abstract_excerpt":"Human Activity Recognition (HAR) on resource-constrained wearable devices demands inference models that harmonize accuracy with computational efficiency. This paper introduces TinierHAR, an ultra-lightweight deep learning architecture that synergizes residual depthwise separable convolutions, gated recurrent units (GRUs), and temporal aggregation to achieve SOTA efficiency without compromising performance. Evaluated across 14 public HAR datasets, TinierHAR reduces Parameters by 2.7x (vs. TinyHAR) and 43.3x (vs. DeepConvLSTM), and MACs by 6.4x and 58.6x, respectively, while maintaining the aver"},"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":"2507.07949","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-07-10T17:33:52Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"521a3cc4cc18fca856fdb9913416f420fdd96d666e2712e770b8afbe1892ddbb","abstract_canon_sha256":"2330320505ebe2f4fc23591cb6c73159a569452cb75d839de3d7de61e1062e91"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:35:07.465155Z","signature_b64":"4vnGuaLPtEzql7ZIqtzXE5YmocxLWiZMR929PpVTNZLN/hfkZV9cUgKQlFRT12I64g0/bf80/qLVMAI7EkHeDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dd5469149ca49af8034ac6bd1b8307d48fc9fa724df61de6b17a16f86f186fc6","last_reissued_at":"2026-07-05T11:35:07.464686Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:35:07.464686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TinierHAR: Towards Ultra-Lightweight Deep Learning Models for Efficient Human Activity Recognition on Edge Devices","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Daniel Geissler, Mengxi Liu, Paul Lukowicz, Sizhen Bian, Vitor Fortes Rey","submitted_at":"2025-07-10T17:33:52Z","abstract_excerpt":"Human Activity Recognition (HAR) on resource-constrained wearable devices demands inference models that harmonize accuracy with computational efficiency. This paper introduces TinierHAR, an ultra-lightweight deep learning architecture that synergizes residual depthwise separable convolutions, gated recurrent units (GRUs), and temporal aggregation to achieve SOTA efficiency without compromising performance. Evaluated across 14 public HAR datasets, TinierHAR reduces Parameters by 2.7x (vs. TinyHAR) and 43.3x (vs. DeepConvLSTM), and MACs by 6.4x and 58.6x, respectively, while maintaining the aver"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.07949","kind":"arxiv","version":1},"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/2507.07949/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":"2507.07949","created_at":"2026-07-05T11:35:07.464745+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.07949v1","created_at":"2026-07-05T11:35:07.464745+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.07949","created_at":"2026-07-05T11:35:07.464745+00:00"},{"alias_kind":"pith_short_12","alias_value":"3VKGSFE4USNP","created_at":"2026-07-05T11:35:07.464745+00:00"},{"alias_kind":"pith_short_16","alias_value":"3VKGSFE4USNPQA2K","created_at":"2026-07-05T11:35:07.464745+00:00"},{"alias_kind":"pith_short_8","alias_value":"3VKGSFE4","created_at":"2026-07-05T11:35:07.464745+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.05572","citing_title":"Viveka: Context-Aware Sensing for Energy Efficiency in Smart Wearables","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S","json":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S.json","graph_json":"https://pith.science/api/pith-number/3VKGSFE4USNPQA2KY26RXAYH2S/graph.json","events_json":"https://pith.science/api/pith-number/3VKGSFE4USNPQA2KY26RXAYH2S/events.json","paper":"https://pith.science/paper/3VKGSFE4"},"agent_actions":{"view_html":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S","download_json":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S.json","view_paper":"https://pith.science/paper/3VKGSFE4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.07949&json=true","fetch_graph":"https://pith.science/api/pith-number/3VKGSFE4USNPQA2KY26RXAYH2S/graph.json","fetch_events":"https://pith.science/api/pith-number/3VKGSFE4USNPQA2KY26RXAYH2S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S/action/storage_attestation","attest_author":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S/action/author_attestation","sign_citation":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S/action/citation_signature","submit_replication":"https://pith.science/pith/3VKGSFE4USNPQA2KY26RXAYH2S/action/replication_record"}},"created_at":"2026-07-05T11:35:07.464745+00:00","updated_at":"2026-07-05T11:35:07.464745+00:00"}