{"as_of":"2026-08-19T08:01:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:2fbc13faaa8c1345316e2f5b50f5abce834387cc27ff8ff6369c1c0d60fc259f","coverage":[{"denominator":33,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":33,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-01T12:00:35.706131Z","state":"measured"},{"denominator":33,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":33,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-19T06:32:44.657259+00:00","state":"measured"},{"denominator":0,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":0,"source":"paper_references, paper_reference_links","source_observed_at":null,"state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"cited_works","source_observed_at":null,"state":"measured"}],"external_citation_measurements":[],"inbound":[],"links":{"evidence":"/evidence","html":"/paper/2607.26631/citation-record","integrity":"/paper/2607.26631/integrity","json":"/paper/2607.26631/citation-record.json","paper":"/paper/2607.26631"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:31.429051Z","title":"A tutorial on human activity recognition using body-worn inertial sensors,","venue":null,"work_id":null,"year":2014},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:31.429051Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:bb81f3cfa21358ad1dc1a6cae56c708386726947771bc81c7e15968c215139c4","observation_id":"840b054f-4b67-401a-a503-b9911eee421a","resolution":{"observed_at":"2026-08-01T12:00:31.429051Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:31.523566Z","title":"A survey on human activity recognition using wearable sensors,","venue":null,"work_id":null,"year":2013},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:31.523566Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:620f734149db2f5c0322f4548da045cedcc2e8cbc9b15a0ad4e7c35bd47e7071","observation_id":"811a218c-b946-4a7c-9922-8f55f9902a83","resolution":{"observed_at":"2026-08-01T12:00:31.523566Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:31.563743Z","title":"Activity recognition from on-body sensors: Accuracy- power trade-off by dynamic sensor selection,","venue":null,"work_id":null,"year":2008},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:31.563743Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:b1780c9766984073aee2d1b98a1c871d9bab762cad770ba9735ff2d4b5f5d1e2","observation_id":"e0c3c64b-034e-4a0d-8170-cbe2adc43611","resolution":{"observed_at":"2026-08-01T12:00:31.563743Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:31.635849Z","title":"Deep convolutional neural networks on multichannel time series for human activity recognition,","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:31.635849Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:6fba641d6bb7eb34c7d3b6eefee1d11a39a6e3dcbca8a33cd5c24a426e4845eb","observation_id":"40262e47-a7db-48a0-b953-e6aeff82b908","resolution":{"observed_at":"2026-08-01T12:00:31.635849Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:31.776288Z","title":"Deep convolutional and lstm recurrent neural networks for multimodal wearable activity recognition,","venue":null,"work_id":null,"year":2016},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:31.776288Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:beeed43de794bf61f61f3b176f047e1c04f8c0925d993e7535eaa2e7672dbc07","observation_id":"37924a58-5f03-43db-b213-6d5ce706e474","resolution":{"observed_at":"2026-08-01T12:00:31.776288Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:31.907218Z","title":"Metier: A deep multi-task learning-based activity and user recognition model using wearable sensors,","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:31.907218Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:3541b51019c1b6e68b095faf9bb6d81e2cf6d41a52d5430cd413ac6832798f2b","observation_id":"8406d1a6-ac55-480c-b247-e7165343ddbf","resolution":{"observed_at":"2026-08-01T12:00:31.907218Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:32.043838Z","title":"Rag-har: Retrieval augmented generation-based human activity recognition,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.043838Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:a1b9f4de0502b185a4a52ed8918f23ea6bc3e133fdb3f1d81c2f761448d082b1","observation_id":"fbabbdf2-7caf-4bb9-a89f-f301c636b04a","resolution":{"observed_at":"2026-08-01T12:00:32.043838Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:32.212915Z","title":"Large language models are zero-shot time series forecasters,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.212915Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:13cef556dcff9467bea688189195156aae7face103a03707735212afdc239931","observation_id":"e3bd38b3-3853-4ce9-b444-9caa2799838a","resolution":{"observed_at":"2026-08-01T12:00:32.212915Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2310.03589","last_updated":"2024-05-27T23:34:15Z","snapshot_observed_at":"2026-08-16T14:54:46.227012Z","submitted_at":"2023-10-05T15:14:00Z","title":"TimeGPT-1","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2310.03589","snapshot_observed_at":"2026-08-01T12:00:32.380856Z","title":"Timegpt- 1,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.380856Z"},"links":{"cited_paper":"/paper/2310.03589","citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:638f1fcbfc7a4763d86d7ba7c53ce149901b39de02e702ed745db11d075bbc39","observation_id":"39a0fd8d-d4f0-4266-89e8-1fa2e6507838","resolution":{"observed_at":"2026-08-01T12:00:32.380856Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:32.554601Z","title":"Time-llm: Time series forecasting by reprogramming large language models,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.554601Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:c2fa96e447cb5f59fbe2f7d5d3ca15be9926da8d9ad0403a6c59adc1d0476e2f","observation_id":"47ac68f9-fa12-4cf5-a1e0-f57cb29c8d97","resolution":{"observed_at":"2026-08-01T12:00:32.554601Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:32.713345Z","title":"Sensorllm: Aligning large language models with motion sensors for human activity recognition,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.713345Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:11e6afdee88b693e4ee2418bb7e4b861b25c96f1acd7ae5aea764ac0da883f95","observation_id":"23047bdc-d3fa-42ce-881c-48a4ed0af195","resolution":{"observed_at":"2026-08-01T12:00:32.713345Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:32.846843Z","title":"Llm4har: Generalizable on-device human activity recognition with pretrained llms,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.846843Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:e25af8bbbdaf15892da31365bf06491503263bbb6a2730b0ead5e3c16db2980b","observation_id":"4f263870-9e37-49aa-b736-06909feb8f0f","resolution":{"observed_at":"2026-08-01T12:00:32.846843Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:32.996255Z","title":"A personal health large language model for sleep and fitness coaching,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:32.996255Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:e447c5ffd00a0558827d00aae4d637d796c775aefd6dad279a3bbd5040bd8328","observation_id":"4ac4cdf1-aa32-4f1b-9b09-ad590fc1e4b2","resolution":{"observed_at":"2026-08-01T12:00:32.996255Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:33.119709Z","title":"Contrastive self- supervised learning for sensor-based human activity recognition,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:33.119709Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:b995f717f3896c5760c1a4be393d23300170c5eac8d6b3a1e04652d2a11c7448","observation_id":"07e3c33d-7e17-4c51-a553-2f11f49c32cd","resolution":{"observed_at":"2026-08-01T12:00:33.119709Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:33.283114Z","title":"Towards llm- powered ambient sensor based multi-person human activity recognition,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:33.283114Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:c910adf91b8a3273ec19b83051317bc16fe3f8af50e041f08fbe1f4634020798","observation_id":"221dc6e3-940d-4eba-b156-ab55deb33591","resolution":{"observed_at":"2026-08-01T12:00:33.283114Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2306.02140","last_updated":"2023-06-03T15:41:59Z","snapshot_observed_at":"2026-08-16T15:26:53.249481Z","submitted_at":"2023-06-03T15:41:59Z","title":"Unsupervised Human Activity Recognition through Two-stage Prompting with ChatGPT","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2306.02140","snapshot_observed_at":"2026-08-01T12:00:33.413295Z","title":"Unsupervised human activity recognition through two-stage prompting with chatgpt,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:33.413295Z"},"links":{"cited_paper":"/paper/2306.02140","citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:8f486c866741eb20daa1b8f61da1353fe5dac46680eb3abe10a7856020f7185d","observation_id":"b59ca483-fb21-4ac0-b189-c2cbee593997","resolution":{"observed_at":"2026-08-01T12:00:33.413295Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:33.577285Z","title":"Hargpt: Are llms zero-shot human activity recognizers?","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:33.577285Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:206b619cc25fe52ada95f2e15ad104016cd2947cdeee36ec8e2c4dececb9ba51","observation_id":"2a47cb31-daa4-4d82-8a7f-0c78f1d144a5","resolution":{"observed_at":"2026-08-01T12:00:33.577285Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:33.766323Z","title":"Tsfel: Time series feature extraction library,","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:33.766323Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:45dbd08bbd0fd0f8909dab517b799b7046b745241e128cc421c0ee523fb72074","observation_id":"17d22593-af1f-46bd-b380-099b2e75312e","resolution":{"observed_at":"2026-08-01T12:00:33.766323Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:33.964735Z","title":"Deep triplet networks with attention for sensor-based human activity recognition,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:33.964735Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:edef8c70f12ac176658bd8a93be9e956afaab0f14fc9c35f1a8a0e7996dac755","observation_id":"d461d297-b4fe-4ff8-a7eb-e9e652b531a0","resolution":{"observed_at":"2026-08-01T12:00:33.964735Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:34.115770Z","title":"Adversarial deep feature extraction network for user independent human activity recognition,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:34.115770Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:dcc8332cdf9857364503d90f36022326bf210b316edfde96a2496fde698e6e36","observation_id":"1672188e-1b88-439d-9e10-90c4d572f4be","resolution":{"observed_at":"2026-08-01T12:00:34.115770Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:34.222813Z","title":"Alae-tae-cutmix+: Beyond the state-of-the- art for human activity recognition using wearable sensors,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:34.222813Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:79cb8373f32fc37ca9ed948719299b5259b6fc87110312e63684592a86007917","observation_id":"ce24a229-8c78-4a8e-9b34-e5bac419645a","resolution":{"observed_at":"2026-08-01T12:00:34.222813Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:34.361259Z","title":"Investigating enhancements to contrastive predictive coding for human activity recognition,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:34.361259Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:dbb6c1801a034c3e103b98be1b7205ab415506748ee2ba6c7e19d972c04a635f","observation_id":"851b853f-b1ec-4147-b6d8-4210fe7878a2","resolution":{"observed_at":"2026-08-01T12:00:34.361259Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:34.542572Z","title":"Usc-had: A daily activity dataset for ubiquitous activity recognition using wearable sensors,","venue":null,"work_id":null,"year":2012},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:34.542572Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:c32f2803a08ba12f89333a621053b17dd363d84b4db103b724a03bf471d97f1d","observation_id":"f106b973-3350-4284-9ee8-74e7d4fd1a7b","resolution":{"observed_at":"2026-08-01T12:00:34.542572Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:34.688221Z","title":"Introducing a new benchmarked dataset for activity monitoring,","venue":null,"work_id":null,"year":2012},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:34.688221Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:44173161308c6ca7de27173873a3a123726364e7e4c0979fb57ad2fdd56aaa19","observation_id":"f85e8eef-ecf8-41b9-a589-28c7e5ae92ae","resolution":{"observed_at":"2026-08-01T12:00:34.688221Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:34.866632Z","title":"mhealthdroid: A novel framework for agile development of mobile health applications,","venue":null,"work_id":null,"year":2014},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:34.866632Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:c00c82e0a11a36a184b2da44b152a7475d9374cef4ca6f72ace41e9247e1afbd","observation_id":"a41225ad-d948-47b3-b864-9d55c85187e5","resolution":{"observed_at":"2026-08-01T12:00:34.866632Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.017939Z","title":"Activity recognition using wearable sensors for tracking the elderly,","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.017939Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:1d53c06a13a0b7dcaec23f96ca655f14fd278471a9e4fe806debae3e1fd3ad09","observation_id":"e254a0ee-c0c3-4eba-82f8-ae3cfd3b45d1","resolution":{"observed_at":"2026-08-01T12:00:35.017939Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.122705Z","title":"Smart devices are different: Assessing and mitigating mobile sensing heterogeneities for activity recognition,","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":27,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.122705Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:ce518125727eed6780182241cd2f85310fd002901b9cff15ff10ffc2111d6c00","observation_id":"ad32ead9-c981-4ef7-aa06-595548aa8311","resolution":{"observed_at":"2026-08-01T12:00:35.122705Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.235458Z","title":"Evaluating large language models as virtual annotators for time-series physical sensing data,","venue":null,"work_id":null,"year":2025},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":28,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.235458Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:979ea21a59ece62caec3d05784abd7af7ab7e9517176c6089dfb3a3975b95605","observation_id":"39a3da11-b6c4-4ac5-a9c6-7ff912bca4bb","resolution":{"observed_at":"2026-08-01T12:00:35.235458Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.304587Z","title":"GPT-5 mini Model,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.304587Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:38c0e1dcb8086b246c905f6f12891050fad578a0b0143673989b6b26a5dbed6c","observation_id":"f446d6e8-e92c-4a36-9687-af8fb68b8eda","resolution":{"observed_at":"2026-08-01T12:00:35.304587Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.417803Z","title":"Large language models meet wearable sensing: A demo of training-free har with rag-har,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":30,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.417803Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:d2e64a8ef3cabca0a62b0e0bb41c7cbde8d6a35a03c082973abfcdf94fc4b7db","observation_id":"8cdedf85-43a5-485e-b8e1-8b5a6e86472a","resolution":{"observed_at":"2026-08-01T12:00:35.417803Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.533631Z","title":"Flutter: Build apps for any screen,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":31,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.533631Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:b0fb335e641f8f83508bc6ecf8873ec2a15dc39777b3f3528145a67028af3bf3","observation_id":"1f9f3bb1-8db7-4b0d-a99e-405bb989f081","resolution":{"observed_at":"2026-08-01T12:00:35.533631Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.614030Z","title":"Milvus vector database,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":32,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.614030Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:2eab4c2a683df6decb34b63d9be522954a2af82302352c01e3c38eab59417631","observation_id":"3ccff169-dc8e-4e8b-98b7-6d929ce118c1","resolution":{"observed_at":"2026-08-01T12:00:35.614030Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T12:00:35.706131Z","title":"A comprehensive survey on synthetic network traffic generation,","venue":null,"work_id":null,"year":2026},"citing_paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment","version":1},"reference_index":33,"source":"pdf_text","source_observed_at":"2026-08-01T12:00:35.706131Z"},"links":{"citing_paper":"/paper/2607.26631"},"observation_digest":"sha256:ae70f5f7fff318ecdcdd1582ec85e0f6389b99baf69e54807077b419d74d7a42","observation_id":"de12d64b-aac9-4c40-90ed-82fd15b21e1e","resolution":{"observed_at":"2026-08-01T12:00:35.706131Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2607.26631","last_updated":"2026-07-29T08:57:03Z","latest_version":1,"primary_category":"cs.LG","snapshot_observed_at":"2026-08-09T23:20:54.359739Z","submitted_at":"2026-07-29T08:57:03Z","title":"RAG-HAR+: Towards Cost-Efficient LLM-Based Human Activity Recognition for Edge Deployment"},"reference_resolution":{"displayed":33,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":33,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":33},"refusal":"A citation records a reference. It does not transfer a finding from one paper to another.","schema":"pith.paper-citation-record.v1","standing_sources":[{"observed_at":"2026-08-19T06:32:44.657259+00:00","source":"crossref"},{"observed_at":"2026-08-19T06:32:39.956319+00:00","source":"retraction_watch"}],"thesis":"As of 19 August 2026, this Paper Citation Record lists 33 of 33 outbound references and 0 inbound Pith citation observations for arXiv:2607.26631."}