{"as_of":"2026-08-08T12:25:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:660f857a2c30eed72da542aa4b726a333d6244e621dd8b380e47b34e7b64ee41","coverage":[{"denominator":18,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":18,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-01T14:20:29.363059Z","state":"measured"},{"denominator":18,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":18,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-08T06:32:00.761636+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/2607.18802/citation-record","integrity":"/paper/2607.18802/integrity","json":"/paper/2607.18802/citation-record.json","paper":"/paper/2607.18802"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-01T14:20:27.234268Z","title":"MCUNet: Tiny deep learning on IoT devices,","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.234268Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:f61ec09bae444e1ac6973eee995f109476261631b222e737bcd0292adf7d27c2","observation_id":"24e497ea-13a6-45a3-9d73-55f42bae90e1","resolution":{"observed_at":"2026-08-01T14:20:27.234268Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2103.13630","last_updated":"2021-06-21T21:01:12Z","snapshot_observed_at":"2026-07-06T10:53:11.986981Z","submitted_at":"2021-03-25T06:57:11Z","title":"A Survey of Quantization Methods for Efficient Neural Network Inference","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2103.13630","snapshot_observed_at":"2026-08-01T14:20:27.328655Z","title":"Gholami, S","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.328655Z"},"links":{"cited_paper":"/paper/2103.13630","citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:dd044ddcf30565750b139df3d820c1983b2e94c84c87bda74c15091cb541ca6e","observation_id":"a6a1969b-5349-42d7-9add-41d6457898bf","resolution":{"observed_at":"2026-08-01T14:20:27.328655Z","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":"2026-08-01T14:20:27.445559Z","title":"TinyTL: Reduce memory, not parameters for efficient on-device learn- ing,","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.445559Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:375f5242f281fbe544319cb3fda8727ff0a2181879b15c63616e58d16a7b4455","observation_id":"7bc437da-0677-40d9-b4a4-9f99f1b9d589","resolution":{"observed_at":"2026-08-01T14:20:27.445559Z","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":"2026-08-01T14:20:27.543308Z","title":"Memory-efficient patch-based inference for tiny deep learning,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.543308Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:8f112f81b204edb27474a3037b3913f328db6fb1ae5535a6be91b144704093ed","observation_id":"c4a4c895-f551-4630-9668-a12f3b0771ec","resolution":{"observed_at":"2026-08-01T14:20:27.543308Z","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":"2026-08-01T14:20:27.615923Z","title":"TinyOL: TinyML with online-learning on microcontrollers,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.615923Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:e723863053b3f6288582d34c82cda6f568944d283bbc4df40e0315e151f5098b","observation_id":"dc04c3d2-7f62-4896-b83f-6dd0cf4f3e8d","resolution":{"observed_at":"2026-08-01T14:20:27.615923Z","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":"2026-08-01T14:20:27.680585Z","title":"Learning representations by back-propagating errors,","venue":null,"work_id":null,"year":1986},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.680585Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:e1797d5bbcb077ad8e9d34ffad54fc082b58bcaa2972a6e4c24d5a33455233ff","observation_id":"53da0bca-ea95-4d39-8920-1fa0255eccba","resolution":{"observed_at":"2026-08-01T14:20:27.680585Z","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":"2026-08-01T14:20:27.753089Z","title":"Stepping forward on the last mile,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.753089Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:9b8f3443081c81d110a9321c31f089ad1b752283946b03c09b9d214b7b09d299","observation_id":"72d680ff-cbf5-4acb-bbf6-d12a20827dfc","resolution":{"observed_at":"2026-08-01T14:20:27.753089Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1604.06174","last_updated":"2016-04-22T19:21:36Z","snapshot_observed_at":"2026-08-08T09:03:25.135475Z","submitted_at":"2016-04-21T04:15:27Z","title":"Training Deep Nets with Sublinear Memory Cost","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1604.06174","snapshot_observed_at":"2026-08-01T14:20:27.960956Z","title":null,"venue":null,"work_id":null,"year":2016},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:27.960956Z"},"links":{"cited_paper":"/paper/1604.06174","citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:039b7b0a3a0104b1749f7d17bea3f9b08385583f13d31fb2ac74746717d4a555","observation_id":"b39301b7-ac54-4514-be1c-b2e0f0eff17a","resolution":{"observed_at":"2026-08-01T14:20:27.960956Z","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":"2026-08-01T14:20:28.155480Z","title":"POET: Training neural networks on tiny devices with integrated rematerialization and paging,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.155480Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:628bf42ea087bf741b0d530f2d63db86fdf0f8e3041c44042325a894b96719a2","observation_id":"8f9a1e06-8d7d-4d6a-a975-f3ad74aca479","resolution":{"observed_at":"2026-08-01T14:20:28.155480Z","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":"2026-08-01T14:20:28.310267Z","title":"Random gradient-free minimization of convex functions,","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.310267Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:15513e32e6a0617e1c3e718b7ffb81274cf231a74b7123f8d52ff3a88aa4b232","observation_id":"1c1dfeea-d008-4c52-baaa-7b666df60bd5","resolution":{"observed_at":"2026-08-01T14:20:28.310267Z","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":"2026-08-01T14:20:28.445159Z","title":"Multivariate stochastic approximation using a simultaneous perturbation gradient approximation,","venue":null,"work_id":null,"year":1992},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.445159Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:5f20fff7ab78bbc4726b3363c3755ab3d13e0ed8636c7ec10b5c2fc5f0f94869","observation_id":"9223ed58-602c-48df-b619-836b6c8005a1","resolution":{"observed_at":"2026-08-01T14:20:28.445159Z","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":"2026-08-01T14:20:28.563098Z","title":"On-device training under 256KB memory,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.563098Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:2f1d2db657cd4708421562f72a54c34b8f9595d5f6ffb511a1f4b33f8bf6ba36","observation_id":"7c2c27e5-3b55-4a5c-b54b-30868d89a6b4","resolution":{"observed_at":"2026-08-01T14:20:28.563098Z","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":"2026-08-01T14:20:28.639797Z","title":"Fine-tuning language models with just forward passes,","venue":null,"work_id":null,"year":2023},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.639797Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:71ce5a227553bd82a8ba2cd1aa2a64fa24a3d28de48e7aff3106877d3dcfaaf6","observation_id":"28ceef89-b1a8-4998-8385-1c7be35b316f","resolution":{"observed_at":"2026-08-01T14:20:28.639797Z","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":"2026-08-01T14:20:28.718262Z","title":"Adam: A method for stochas- tic optimization,","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.718262Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:5b37ed9f6b618a73a870c71175683d2b4c24af017792d627beb6c9286c67cfe1","observation_id":"0b0cff78-55f8-483a-a86d-4943ca8e0dd2","resolution":{"observed_at":"2026-08-01T14:20:28.718262Z","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":"2026-08-01T14:20:28.897072Z","title":"Eu- rosat: A novel dataset and deep learning benchmark for land use and land cover classification,","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:28.897072Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:6db53f5838e2a36ddbb0e39e73f832f44dcd48475b1918cd9940d6ebba19ec1f","observation_id":"820ae85e-d78c-44d9-9253-4ea1a3101221","resolution":{"observed_at":"2026-08-01T14:20:28.897072Z","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":"2026-08-01T14:20:29.117411Z","title":"An analysis of single- layer networks in unsupervised feature learning,","venue":null,"work_id":null,"year":2011},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:29.117411Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:7cc8d73e5636683f1a68b8da9b4f1dbfd38453a91136dec7b785cfb89ed4d0d1","observation_id":"ce28588b-20da-47b1-ab32-7cc8dcbca1fc","resolution":{"observed_at":"2026-08-01T14:20:29.117411Z","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":"2026-08-01T14:20:29.262564Z","title":"Practical bayesian optimization of machine learning hyperparam- eters,","venue":null,"work_id":null,"year":2012},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:29.262564Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:612c90692467c9efa08ad8f534abc59d6b61e2cb156793a164afdca6acfcce3d","observation_id":"42506206-7bcb-48fd-8165-e9ed0c560197","resolution":{"observed_at":"2026-08-01T14:20:29.262564Z","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":"2026-08-01T14:20:29.363059Z","title":null,"venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-01T14:20:29.363059Z"},"links":{"citing_paper":"/paper/2607.18802"},"observation_digest":"sha256:c9e65ed852ca54b62cd08b3d0cd7ece041ce1f00ebb623c50bcb1d3fbd8399b9","observation_id":"fe9fe6d8-1d56-4c6f-a444-c36ef0a78bef","resolution":{"observed_at":"2026-08-01T14:20:29.363059Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2607.18802","last_updated":"2026-07-21T07:30:08Z","latest_version":1,"primary_category":"cs.LG","snapshot_observed_at":"2026-08-03T07:22:34.007883Z","submitted_at":"2026-07-21T07:30:08Z","title":"QScheduler: Adaptive Gradient Sampling for Zeroth-Order On-Device Training on INT8 NPUs"},"reference_resolution":{"displayed":18,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":18,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":18},"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":"2026-08-08T06:32:00.761636+00:00","source":"crossref"},{"observed_at":"2026-08-08T06:31:55.24221+00:00","source":"retraction_watch"}],"thesis":"As of 8 August 2026, this Paper Citation Record lists 18 of 18 outbound references and 0 inbound Pith citation observations for arXiv:2607.18802."}