{"as_of":"2026-08-08T04:44:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:4a44814ccbd56c97d8450e66f30e799647b272b2789364ee0651dde71a2dda9c","coverage":[{"denominator":36,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":36,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-07T13:42:42.235277Z","state":"measured"},{"denominator":37,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":37,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-07T06:34:17.273281+00:00","state":"measured"},{"denominator":1,"lane":"inbound_itemization","note":"Pith citing papers itemized under the disclosed page cap.","records_observed":1,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-07T13:42:39.417023Z","state":"measured"},{"denominator":1,"lane":"external_citation_measurements","note":"A source-named dated measurement, never combined with another source.","records_observed":0,"source":"pith","source_observed_at":"2026-08-07T13:42:43.278084Z","state":"measured"}],"external_citation_measurements":[],"inbound":[{"citation":{"cited_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"cited_work":{"arxiv_id":"2505.21156","doi":null,"metadata_source":"pith","pith_arxiv_id":"2505.21156","snapshot_observed_at":"2026-08-07T13:42:43.278084Z","title":"Model as Loss: A Self-Consistent Training Paradigm","venue":"cs.SD","work_id":"d1423426-eeeb-487c-b04e-d00a4005f586","year":2025},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.417023Z"},"links":{"cited_paper":"/paper/2505.21156","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:b538672d5191727e4c4d737355a5811098f38453393eb082385102d941a52db6","observation_id":"150fbb1c-0708-405f-8fed-afe245e4acbc","resolution":{"observed_at":"2026-08-07T13:42:43.386551Z","resolver_source":"local_arxiv","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}}],"links":{"evidence":"/evidence","html":"/paper/2505.21156/citation-record","integrity":"/paper/2505.21156/integrity","json":"/paper/2505.21156/citation-record.json","paper":"/paper/2505.21156"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:46.798646Z","title":"A critical component in training enhancement models is the choice of loss function, which directly influences the quality and generaliza- tion of enhanced output [6, 7]","venue":null,"work_id":"3ccf4751-347b-4b16-96e4-76c85d849c8a","year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":1,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.302808Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:794507996b6159716c571cdd1fa121c1ab56bb49e498cafe31f2dfc2b8499567","observation_id":"0e687cf0-439d-4192-b42c-ccf3fb14be10","resolution":{"observed_at":"2026-08-07T13:42:46.878072Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"cited_work":{"arxiv_id":"2505.21156","doi":null,"metadata_source":"pith","pith_arxiv_id":"2505.21156","snapshot_observed_at":"2026-08-07T13:42:43.278084Z","title":"Model as Loss: A Self-Consistent Training Paradigm","venue":"cs.SD","work_id":"d1423426-eeeb-487c-b04e-d00a4005f586","year":2025},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":2,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.417023Z"},"links":{"cited_paper":"/paper/2505.21156","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:b538672d5191727e4c4d737355a5811098f38453393eb082385102d941a52db6","observation_id":"150fbb1c-0708-405f-8fed-afe245e4acbc","resolution":{"observed_at":"2026-08-07T13:42:43.386551Z","resolver_source":"local_arxiv","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:46.662617Z","title":"It has an encoder that extracts relevant fea- tures and passes them into a first-stage decoder","venue":null,"work_id":"6d176514-d941-4ffc-a2e2-017d1e6bb95e","year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":3,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.484316Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:d984cb622872c9b59a2aa674904e4ad3eed98d7480d9512956589d9854d9c1ce","observation_id":"bd4daa8e-3902-4d8a-8f4d-024057d3f9e2","resolution":{"observed_at":"2026-08-07T13:42:46.746243Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:46.507287Z","title":"Oursmal−dynamic achieves the best performance across all NISQA metrics, while Ours mal−frozen leads in all intrusive metrics","venue":null,"work_id":"a135e8e6-f304-43c0-bce4-5d9c00385eac","year":2025},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":4,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.535589Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:44cb5159ea8814a8b75a568a136348c2c4fd489542c4aea2e6051b34d2fcde2d","observation_id":"b2057606-70ce-4716-bf5e-5c4417b66d81","resolution":{"observed_at":"2026-08-07T13:42:46.577644Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:46.327997Z","title":"By align- ing the loss with the model’s task-specific feature space,MAL overcomes the limitations of traditional handcrafted and pre- trained deep feature losses","venue":null,"work_id":"2f7db0f1-a1f6-4481-aa0d-6253b15cedd8","year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":5,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.564426Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:ad62ab3d0ce1fdb4b2d3ba7025bb521ee215f7139c114e6aae70887d3934d1d3","observation_id":"adc5ea9c-781c-4e2c-8526-d1ed84ddb60b","resolution":{"observed_at":"2026-08-07T13:42:46.409394Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:46.207693Z","title":"We also thank Sai Dhawal Phaye for discussions during the early stages of MAL, and Kanav Sabharwal for his feedback on the writing","venue":null,"work_id":"7d0fbdad-065b-4e88-a145-368e74b2075e","year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":6,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.600549Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:845765ca2b9351f62ce4c7ae8110303bcb2f6c91bafddc7a6694fb10be2dc910","observation_id":"da2cb6ea-fbfe-4965-8815-eb2b6cd21079","resolution":{"observed_at":"2026-08-07T13:42:46.262251Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:46.041898Z","title":"Benesty, S","venue":null,"work_id":"920477be-ca45-40e7-95e0-2616f51c5802","year":2006},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":7,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.625352Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:cd67fa4b50769f7eb66a6178121cae5eb735e76be2b40e654f148ef9033a4215","observation_id":"db361ab7-6d80-4eb1-8980-21fff662b73f","resolution":{"observed_at":"2026-08-07T13:42:46.099919Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2005.13981","last_updated":"2020-10-18T04:36:21Z","snapshot_observed_at":"2026-07-31T23:56:07.410341Z","submitted_at":"2020-05-16T23:48:37Z","title":"The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, Subjective Testing Framework, and Challenge Results","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2005.13981","snapshot_observed_at":"2026-08-07T13:42:39.659525Z","title":"The interspeech 2020 deep noise suppression challenge: Datasets, subjective testing framework, and challenge results,","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":8,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.659525Z"},"links":{"cited_paper":"/paper/2005.13981","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:96c4767d63ae33736311e03bb1e0e41251ada67b7a3bc4b351c4014ca9e04d3d","observation_id":"5a19d199-cecf-4d1e-99a9-df58fa5eaf18","resolution":{"observed_at":"2026-08-07T13:42:39.659525Z","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":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:45.885372Z","title":"FINALLY: fast and universal speech enhancement with studio- like quality,","venue":null,"work_id":"1b9086ec-23d3-45b5-a23d-e5ab851aa5be","year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":9,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.687723Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:34c368850a81bef01f2a034429cda75c6fbe175cb28987975ff0f2976e588433","observation_id":"59f19fad-0952-4a46-97a4-a0f158c95a2b","resolution":{"observed_at":"2026-08-07T13:42:45.961420Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:45.711926Z","title":"Icassp 2023 deep noise suppression challenge,","venue":null,"work_id":"428cbb7e-cefd-4348-bdf4-4d7d613a52bc","year":2023},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":10,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.718277Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:f02c63f6c228b75d6520783e3609a5def2a919fcf51d003e11884ed0f63604b6","observation_id":"91f9ca76-1716-46ee-bd05-0ca97149b6b8","resolution":{"observed_at":"2026-08-07T13:42:45.814651Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:40.099883Z","title":"Wavlm: Large-scale self-supervised pre-training for full stack speech processing,","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":11,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.099883Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:4877e18841c3c4bcd3ad4b40c1d34281cc5a29407fd5539cc765a8672eafb81c","observation_id":"0cfe1aad-1d20-49ea-99d6-fa17a36c6706","resolution":{"observed_at":"2026-08-07T13:42:40.099883Z","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":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:45.550698Z","title":"DeepFilterNet2: Towards real-time speech enhancement on em- bedded devices for full-band audio,","venue":null,"work_id":"afd3dc24-a4cf-42b1-836c-ce00563b27ec","year":2022},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":12,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.793782Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:979f0a0cbe16709c3d96e15acefdc0ca66db676f770cd5758404b02956dac3a7","observation_id":"c54b29f9-6209-4705-8ce6-54c93371eb01","resolution":{"observed_at":"2026-08-07T13:42:45.637746Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:45.406703Z","title":"Speech denoising with deep feature losses,","venue":null,"work_id":"9e77422b-980b-4fff-8031-45402eb9b793","year":2019},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":13,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.850451Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:57a5ff6d50ac22826ef07518b886402fc5f1edd406a72ade68fa2289245a185f","observation_id":"cd0e01e3-40a0-46ae-bca0-6769c5d08127","resolution":{"observed_at":"2026-08-07T13:42:45.457857Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:39.901561Z","title":"A consolidated view of loss functions for supervised deep learning-based speech enhancement,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":14,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.901561Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:9f60d9a3973d3ad95310478a920175f0158fd98cabd0864a80af9fe99f93b224","observation_id":"edc8d6a3-73ac-4ff2-b995-3ade4958300a","resolution":{"observed_at":"2026-08-07T13:42:39.901561Z","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":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:45.239494Z","title":"auraloss: Audio focused loss functions in PyTorch,","venue":null,"work_id":"b7f1d4f7-a607-47cd-8312-c21dca929ef4","year":2020},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":15,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.937356Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:def1c9254645de98ea7851b026b68d79a39a325da8b28ace3f04ff9eba059cf4","observation_id":"42003d0f-b1e2-4356-abe8-327c4ef1f3d6","resolution":{"observed_at":"2026-08-07T13:42:45.308332Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:45.086820Z","title":"A deep learning loss function based on the perceptual evaluation of the speech quality,","venue":null,"work_id":"154e38c7-1b88-4153-b70a-0cb50a9e93e0","year":2018},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":16,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.990681Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:1f3ac7afe8a190aa81c21121cf647f14a87e0168cef780d7d88107062e4ed2d9","observation_id":"ebdb0d5a-8ba9-4425-9e52-04e3060ee408","resolution":{"observed_at":"2026-08-07T13:42:45.148533Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:44.871914Z","title":"On loss func- tions for supervised monaural time-domain speech enhancement,","venue":null,"work_id":"f25853cf-03b1-43e1-8e85-4c773e6dd374","year":2020},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":17,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.047743Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:e3f1ca3b58b478525a751599de94ab3895129d1c9c6ffc72d36008078d7eed0c","observation_id":"d7db6738-659c-470b-b7db-e16ae999cec3","resolution":{"observed_at":"2026-08-07T13:42:44.974447Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:41.253378Z","title":"Nisqa: A deep cnn-self-attention model for multidimensional speech quality prediction with crowdsourced datasets,","venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":18,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.253378Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:1c50a7917f08a2104d88bbe3495b05e8cb192c5d8e49983d79997a9f855a131b","observation_id":"fcc36071-f261-4f1b-bea1-bbb06ebcea90","resolution":{"observed_at":"2026-08-07T13:42:41.253378Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1904.05862","last_updated":"2019-09-11T08:19:49Z","snapshot_observed_at":"2026-08-06T05:33:37.746688Z","submitted_at":"2019-04-11T17:29:30Z","title":"wav2vec: Unsupervised Pre-training for Speech Recognition","version":4},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1904.05862","snapshot_observed_at":"2026-08-07T13:42:40.142844Z","title":"wav2vec: Unsupervised pre-training for speech recognition,","venue":null,"work_id":null,"year":1904},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":19,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.142844Z"},"links":{"cited_paper":"/paper/1904.05862","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:adb1c4480074426b5ab41fa184ee216756faeb872e5767988bf9eb1f15d504fe","observation_id":"6c47dc4f-ccf0-487f-9318-a04e7d3f7167","resolution":{"observed_at":"2026-08-07T13:42:40.142844Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2006.11477","last_updated":"2020-10-22T06:09:10Z","snapshot_observed_at":"2026-08-07T05:13:16.889179Z","submitted_at":"2020-06-20T02:35:02Z","title":"wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations","version":3},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2006.11477","snapshot_observed_at":"2026-08-07T13:42:40.191267Z","title":"wav2vec 2.0: A framework for self-supervised learning of speech representations,","venue":null,"work_id":null,"year":2006},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":20,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.191267Z"},"links":{"cited_paper":"/paper/2006.11477","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:d3d3dc97b5c7187b121a65d39c5496900a0560ebcefce5b3389cc066e915772a","observation_id":"303cc871-b196-4e99-a144-c27aab83368c","resolution":{"observed_at":"2026-08-07T13:42:40.191267Z","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":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:44.715192Z","title":"Speechlmscore: Evaluating speech generation using speech language model,","venue":null,"work_id":"85e8cc91-8e7b-49dc-b168-b67f130d5adb","year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":21,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.261946Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:eb2be1b37a958fbb82278ec312703a9bc1c94f73026604c0895995be9502d45c","observation_id":"44abafa9-faad-436e-97c9-c881514f32e3","resolution":{"observed_at":"2026-08-07T13:42:44.788952Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:44.242712Z","title":"Enhancing lora reception with generative models: Channel-aware denoising of loraphy signals,","venue":null,"work_id":"00dac5c4-f1f2-4f53-a332-fc247de38480","year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":22,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.803264Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:8ded8bb0ca4300a375e5281bbc6ded406095a6712a6692474483cacdc4ccc479","observation_id":"dd060110-7b75-4eb9-adb8-ec375a6cfb94","resolution":{"observed_at":"2026-08-07T13:42:44.314166Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:40.490769Z","title":"Icassp 2022 deep noise suppression challenge,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":23,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.490769Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:7dd53a0840f9f9c634166f3d324ec06eb11906a0be89334d2ce27c4da5b6154f","observation_id":"4464a80b-d1d0-4913-b132-de41569bc7ce","resolution":{"observed_at":"2026-08-07T13:42:40.490769Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2202.13288","last_updated":"2022-02-27T04:28:57Z","snapshot_observed_at":"2026-08-01T20:28:17.535698Z","submitted_at":"2022-02-27T04:28:57Z","title":"ICASSP 2022 Deep Noise Suppression Challenge","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2202.13288","snapshot_observed_at":"2026-08-07T13:42:40.643307Z","title":"Available: https://arxiv.org/abs/2202.13288","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":24,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.643307Z"},"links":{"cited_paper":"/paper/2202.13288","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:3a8ffb9322c7e73767df51e77440d3e39396cbd5034ab3c4d0243b5983d7e0d3","observation_id":"be6f4b9e-2860-4bbe-ae6c-4c34b7946efc","resolution":{"observed_at":"2026-08-07T13:42:40.643307Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2406.04660","last_updated":"2024-06-07T05:59:01Z","snapshot_observed_at":"2026-07-06T18:26:58.940425Z","submitted_at":"2024-06-07T05:59:01Z","title":"URGENT Challenge: Universality, Robustness, and Generalizability For Speech Enhancement","version":1},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2406.04660","snapshot_observed_at":"2026-08-07T13:42:40.758830Z","title":"Urgent challenge: Universality, robustness, and generalizability for speech enhance- ment,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":25,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.758830Z"},"links":{"cited_paper":"/paper/2406.04660","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:e562a1d87134cf994122bf565c27481d94b2f42c9823d4b4309e64e8ef448b28","observation_id":"5bc3bf02-a5de-4c8d-adfa-f947ccf833b0","resolution":{"observed_at":"2026-08-07T13:42:40.758830Z","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":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:44.615702Z","title":"Icassp 2024 speech signal improvement challenge,","venue":null,"work_id":"c7972776-064d-4940-ba20-d22c39bb85e4","year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":26,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.915443Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:bfd7e50fb33689db0bf96540913a736575df2ae99d74b5138585b64d1fafaf3c","observation_id":"54651185-5f5a-464d-9925-227b7caeefb0","resolution":{"observed_at":"2026-08-07T13:42:44.658530Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2410.06675","last_updated":"2025-01-15T09:50:10Z","snapshot_observed_at":"2026-07-06T19:30:17.275798Z","submitted_at":"2024-10-09T08:30:59Z","title":"SCOREQ: Speech Quality Assessment with Contrastive Regression","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2410.06675","snapshot_observed_at":"2026-08-07T13:42:41.419789Z","title":"Scoreq: Speech quality assessment with contrastive regression,","venue":null,"work_id":null,"year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":29,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.419789Z"},"links":{"cited_paper":"/paper/2410.06675","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:0607f1d51dd8649415102d9dde886b12e198042dfdaeae20708160ed0f27ceb0","observation_id":"4f43fbe4-1f5f-4c6c-becd-d29978df1be6","resolution":{"observed_at":"2026-08-07T13:42:41.419789Z","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":"2026-08-07T13:42:41.587230Z","title":"Do- main adaptation and autoencoder-based unsupervised speech enhancement,","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":30,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.587230Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:919282dd9ed3d2f29baa8ab0dbf30d10913bab5a9b86342a174069882f210dd2","observation_id":"6b0ccf30-8112-4dd6-81d9-42553d8ac460","resolution":{"observed_at":"2026-08-07T13:42:41.587230Z","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":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:44.426843Z","title":"Medical image denoising using convolutional de- noising autoencoders,","venue":null,"work_id":"cd904713-2541-47f4-867a-1a5b2fdc6f6c","year":2016},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":31,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.711774Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:1814279d640a1b7ae9c007ccd36863f49a8ed2c5762fc139f8f52f641c45e202","observation_id":"f11b3ce9-39c7-4833-b654-1f33f4406337","resolution":{"observed_at":"2026-08-07T13:42:44.499649Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:44.064066Z","title":"Neural cascade architecture for multi- channel acoustic echo suppression,","venue":null,"work_id":"0b7f8f5c-f9db-4fdb-8f43-362df5963738","year":2022},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":33,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.914328Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:698ee03925e43b6eb1e8cddbeb172bbe9cd9462c2ab928b595f87fbee68483b1","observation_id":"9543ab20-8336-4368-9896-72f919dc0d86","resolution":{"observed_at":"2026-08-07T13:42:44.162551Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:43.883297Z","title":"Task splitting for dnn-based acous- tic echo and noise removal,","venue":null,"work_id":"51d14fdb-8220-4dfc-bdbb-02c881d528c4","year":2022},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":34,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:42.025664Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:47d381623bf1ec301a926904ca1d9790c572ee90a02b394354b6805b6dfe02a4","observation_id":"ecab8680-b463-473e-a1b2-1bb8089e6e9e","resolution":{"observed_at":"2026-08-07T13:42:43.979802Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:43.665925Z","title":"Deep learning on image denoising: An overview,","venue":null,"work_id":"fbd4f2cb-8e17-4e84-8f4e-26c019a4df64","year":2020},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":35,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:42.104499Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:d1ea0b09311eea3f9b0f333d419b88139182e6c8b95a387dc6bdfe611ce420c5","observation_id":"48154d5e-cad9-406d-b0e0-735c677de124","resolution":{"observed_at":"2026-08-07T13:42:43.771611Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-07T13:42:43.505162Z","title":"A review of the deep learning methods for medical images super resolution problems,","venue":null,"work_id":"6303cab6-43d3-42d8-89f1-be7e0bdf9835","year":2021},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":36,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:42.235277Z"},"links":{"citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:56247d5dca02ef5022c7751be27476a8f122b8fd16ef9cf87a1897550738a160","observation_id":"87513493-1128-4794-813d-244aeee0a603","resolution":{"observed_at":"2026-08-07T13:42:43.580443Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2212.04559","last_updated":"2022-12-08T21:00:15Z","snapshot_observed_at":"2026-07-06T14:28:29.365366Z","submitted_at":"2022-12-08T21:00:15Z","title":"SpeechLMScore: Evaluating speech generation using speech language model","version":1},"cited_work":{"arxiv_id":"2212.04559","doi":null,"metadata_source":"pith","pith_arxiv_id":"2212.04559","snapshot_observed_at":"2026-08-07T13:42:42.922449Z","title":"SpeechLMScore: Evaluating speech generation using speech language model","venue":"eess.AS","work_id":"60c24800-d16b-44ae-804d-a1cb1047e843","year":2022},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":2022,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:40.354094Z"},"links":{"cited_paper":"/paper/2212.04559","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:31b4e75d880237046c008a701aa988eaac6c90fc58721470eeaa228d4fa25a64","observation_id":"908ec97b-432e-4761-956d-d96b12707b8a","resolution":{"observed_at":"2026-08-07T13:42:42.972593Z","resolver_source":"local_arxiv","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}},{"citation":{"cited_paper":{"arxiv_id":"2303.11510","last_updated":"2023-05-09T00:26:43Z","snapshot_observed_at":"2026-07-06T15:05:57.311806Z","submitted_at":"2023-03-21T00:11:47Z","title":"ICASSP 2023 Deep Noise Suppression Challenge","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2303.11510","snapshot_observed_at":"2026-08-07T13:42:39.748666Z","title":"Available: https://arxiv.org/abs/2303.11510","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":2023,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:39.748666Z"},"links":{"cited_paper":"/paper/2303.11510","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:38f33f96500c8bcd307427a7d38105f6ad3dbbe8929bcf000b22af5ffaa0e0d0","observation_id":"bd69fe6a-349b-4f5f-be5e-4ca7af14ca70","resolution":{"observed_at":"2026-08-07T13:42:39.748666Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2401.14444","last_updated":"2024-01-25T18:08:00Z","snapshot_observed_at":"2026-07-06T17:20:37.280501Z","submitted_at":"2024-01-25T18:08:00Z","title":"ICASSP 2024 Speech Signal Improvement Challenge","version":1},"cited_work":{"arxiv_id":"2401.14444","doi":null,"metadata_source":"pith","pith_arxiv_id":"2401.14444","snapshot_observed_at":"2026-08-07T13:42:42.609714Z","title":"ICASSP 2024 Speech Signal Improvement Challenge","venue":"cs.SD","work_id":"54c4b591-7215-4a4a-9abf-f782b1622510","year":2024},"citing_paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm","version":1},"reference_index":2024,"source":"pdf_text","source_observed_at":"2026-08-07T13:42:41.085257Z"},"links":{"cited_paper":"/paper/2401.14444","citing_paper":"/paper/2505.21156"},"observation_digest":"sha256:da6f16eb7e5a5fb97dd5f32137c349e4156d2683f44d979d2e5841ff53f0b865","observation_id":"3646aa66-9ea5-4e07-8b1c-6ca5d0b6f0e0","resolution":{"observed_at":"2026-08-07T13:42:42.718972Z","resolver_source":"local_arxiv","status":"metadata_mismatch"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-07T06:34:17.273281+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"state":"measured"}}],"paper":{"arxiv_id":"2505.21156","last_updated":"2025-05-27T13:12:45Z","latest_version":1,"primary_category":"cs.SD","snapshot_observed_at":"2026-08-07T13:31:27.084867Z","submitted_at":"2025-05-27T13:12:45Z","title":"Model as Loss: A Self-Consistent Training Paradigm"},"reference_resolution":{"displayed":36,"state_counts":{"malformed_identifier":0,"metadata_mismatch":3,"parse_uncertain":0,"unresolved":12,"verified_exact":0,"verified_fuzzy":21},"total_outbound_references":36},"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":"2026-08-07T06:34:17.273281+00:00","source":"crossref"},{"observed_at":"2026-08-07T06:34:11.927384+00:00","source":"retraction_watch"}],"thesis":"As of 8 August 2026, this Paper Citation Record lists 36 of 36 outbound references and 1 inbound Pith citation observation for arXiv:2505.21156."}