{"as_of":"2026-08-09T14:24:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:4b2af71bc4ee0be011db570c2c963b331d1a3f60cb99fa3924d279d1737c3701","coverage":[{"denominator":12,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":12,"source":"paper_references, paper_reference_links","source_observed_at":"2026-07-12T14:06:47.857777Z","state":"measured"},{"denominator":12,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":12,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-09T06:31:02.800959+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/2606.14427/citation-record","integrity":"/paper/2606.14427/integrity","json":"/paper/2606.14427/citation-record.json","paper":"/paper/2606.14427"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"An evaluation of edge tpu accelerators for convolutional neural networks,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:225fa8a0ce1fdbf7b7d635fd3efb4eb68bf249ae4da85ad6a917be529f748ca8","observation_id":"b3938e1a-a69e-4f35-8376-ec29a2872ad9","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","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":"2026-07-12T14:06:47.857777Z","title":"Xnor neural engine: A hardware accelerator ip for 21.6-fj/op binary neural network inference,","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:5fabc19bb5517f23b54b37abcf706ba71a2f4a523ddb55be6a662967898f29c1","observation_id":"db55d524-59de-4311-996c-74dc78bf3002","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2403.09441","last_updated":"2026-05-28T08:31:23Z","snapshot_observed_at":"2026-07-06T17:44:37.894023Z","submitted_at":"2024-03-14T14:34:25Z","title":"An Empirical Study of the Influence of Adversarial Fine-Tuning on Compressed Neural Networks","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2403.09441","snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"Adversarial fine-tuning of compressed neural networks for joint improvement of robustness and efficiency,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"cited_paper":"/paper/2403.09441","citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:db8a983a5e39025673d23a88cf3795b1c46873a607e1db41050b0d4c788011af","observation_id":"9f81d80e-b45d-47e3-8046-5c86a8ad210b","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","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":"2026-07-12T14:06:47.857777Z","title":"David and goliath: An empirical evaluation of attacks and defenses for qnns at the deep edge,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:3bb864eae913f0f96f8cedb4c509ff6e4a81a2314123d1a068329ae5176ce334","observation_id":"c93c8359-5a9f-4b1f-9fe3-64d490fe550e","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1605.07277","last_updated":"2016-05-24T03:27:48Z","snapshot_observed_at":"2026-07-06T04:57:19.691523Z","submitted_at":"2016-05-24T03:27:48Z","title":"Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samples","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1605.07277","snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"Transferability in ma- chine learning: from phenomena to black-box attacks using adversarial samples,","venue":null,"work_id":null,"year":2016},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"cited_paper":"/paper/1605.07277","citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:ad1265e0a962b09890ff1acab3b38328a6224e404fdcbf7a80f36fc50decb6a6","observation_id":"d7b81524-ae33-46f0-b95d-ee120e0b40df","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","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":"2026-07-12T14:06:47.857777Z","title":"Prada: protecting against dnn model stealing attacks,","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:88a5d0c52501ddaf63b8ac45da55341f78b574c88bf70ea672f1fede523b55a1","observation_id":"d47579e4-ccb6-463d-854b-cf5d7049a346","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","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":"2026-07-12T14:06:47.857777Z","title":"Knockoff nets: Stealing function- ality of black-box models,","venue":null,"work_id":null,"year":2019},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:48939112f4af286a266ac9e609eb0743bba814a024f0ab55470bc8d5a2ac44d5","observation_id":"3a3e9d38-ce59-4c4b-a5e8-a58cbac4fde5","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","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":"2026-07-12T14:06:47.857777Z","title":"Copycat cnn: Stealing knowledge by persuading confession with random non-labeled data,","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:f6f59209c4bde0caa6b46226418ada3ddb1ca9bfdc9fa4656ad0c3d6def733ec","observation_id":"8bb9d968-179d-421c-bcb5-f165670cf340","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2106.07597","last_updated":"2021-08-24T19:12:29Z","snapshot_observed_at":"2026-08-03T14:46:30.239773Z","submitted_at":"2021-06-14T17:05:17Z","title":"MLPerf Tiny Benchmark","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2106.07597","snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"Mlperf tiny bench- mark,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"cited_paper":"/paper/2106.07597","citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:6dfd2aae037d1735707d8222498b2104332038691a40eb36005318cd9b634393","observation_id":"2fa0aaf9-1cb5-45be-9346-41cfdfe39fde","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1412.6572","last_updated":"2015-03-20T20:19:16Z","snapshot_observed_at":"2026-07-06T04:04:16.777653Z","submitted_at":"2014-12-20T01:17:12Z","title":"Explaining and Harnessing Adversarial Examples","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1412.6572","snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"Explaining and harnessing adversarial examples,","venue":null,"work_id":null,"year":2014},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"cited_paper":"/paper/1412.6572","citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:0a7b77af015eb93b88d9ff487db54784fda6e6ea82a015e37f1bb7074f9cb555","observation_id":"12fb4f4c-5048-4d0e-9ed1-12b214c4389a","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1706.06083","last_updated":"2019-09-04T18:53:10Z","snapshot_observed_at":"2026-08-07T14:27:46.872660Z","submitted_at":"2017-06-19T17:53:11Z","title":"Towards Deep Learning Models Resistant to Adversarial Attacks","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1706.06083","snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"Towards deep learning models resistant to adversarial attacks,","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"cited_paper":"/paper/1706.06083","citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:57a577ba3c1ce64793d5445d08e81e2a216d2dc569474e70273e9d46d51d5d6f","observation_id":"738eb42e-0e50-4032-ae27-fa8e1594153a","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2010.09670","last_updated":"2021-10-31T20:03:39Z","snapshot_observed_at":"2026-08-09T09:16:24.377074Z","submitted_at":"2020-10-19T17:06:18Z","title":"RobustBench: a standardized adversarial robustness benchmark","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2010.09670","snapshot_observed_at":"2026-07-12T14:06:47.857777Z","title":"Robustbench: a standardized adversarial robustness benchmark,","venue":null,"work_id":null,"year":2010},"citing_paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-07-12T14:06:47.857777Z"},"links":{"cited_paper":"/paper/2010.09670","citing_paper":"/paper/2606.14427"},"observation_digest":"sha256:87d3903c6972141d15289e95c4e11c6bc138405f43b2f36e6b3ef0598f0c8026","observation_id":"b6f084ff-dac2-4ac7-8c52-5cc71819e445","resolution":{"observed_at":"2026-07-12T14:06:47.857777Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2606.14427","last_updated":"2026-06-12T13:05:34Z","latest_version":1,"primary_category":"cs.CR","snapshot_observed_at":"2026-08-04T19:13:08.620726Z","submitted_at":"2026-06-12T13:05:34Z","title":"Breaking TinyML: Why Quantized Neural Networks Need Domain-Specific Security Analysis"},"reference_resolution":{"displayed":12,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":12,"verified_exact":0,"verified_fuzzy":0},"total_outbound_references":12},"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":"2026-08-09T06:31:02.800959+00:00","source":"crossref"},{"observed_at":"2026-08-09T06:30:57.326959+00:00","source":"retraction_watch"}],"thesis":"As of 9 August 2026, this Paper Citation Record lists 12 of 12 outbound references and 0 inbound Pith citation observations for arXiv:2606.14427."}