{"as_of":"2026-08-14T07:36:00Z","caps":{"database_statements":6,"inbound":100,"outbound":100},"context_digest":"sha256:728b8f1a70b1866573b28aba1118cc29e91460dfabc480a6fbea306359e9fa13","coverage":[{"denominator":29,"lane":"reference_resolution","note":"Typed states for the displayed outbound observations.","records_observed":29,"source":"paper_references, paper_reference_links","source_observed_at":"2026-08-09T19:55:16.923519Z","state":"measured"},{"denominator":29,"lane":"standing_notices","note":"One-hop event checks from named stored sources.","records_observed":29,"source":"scholarly_work_events, retraction_status_cache","source_observed_at":"2026-08-14T06:32:32.682623+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/2502.00217/citation-record","integrity":"/paper/2502.00217/integrity","json":"/paper/2502.00217/citation-record.json","paper":"/paper/2502.00217"},"outbound":[{"citation":{"cited_paper":null,"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":"raw_reference","pith_arxiv_id":null,"snapshot_observed_at":"2026-08-09T19:55:17.273030Z","title":"Just pick a sign: Optimizing deep multitask models with gradient sign dropout","venue":null,"work_id":"62a4ccc3-a463-438e-9f10-4d8a2d5a2fbb","year":2020},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":1,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.803979Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:b9c8d07054c3521a7ccf115bc94cd5800b016d406fdccd2ae85e721bb7bcdc94","observation_id":"8e0b8900-7fa1-4245-b316-e4ebb2f84e9f","resolution":{"observed_at":"2026-08-09T19:55:17.277223Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.808923Z","title":"The cityscapes dataset for semantic urban scene understanding","venue":null,"work_id":null,"year":2016},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":2,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.808923Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:6eb1b66b48f24812a3c729f49b6c593bbfe99b5d500d332494252d9064709f24","observation_id":"aa1b066a-010a-4027-90b1-2e894a11b176","resolution":{"observed_at":"2026-08-09T19:55:16.808923Z","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":"2026-08-09T19:55:17.253357Z","title":"Multiple-gradient descent algorithm (mgda) for multiobjective optimization","venue":null,"work_id":"f3e4f49f-c38a-46c6-bb86-bba1a5689328","year":2012},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":3,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.813365Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:d5757f82e7e9b9cabea6f9db97869b13c7dfece2e6faa3f0abdc4b58368ba1cd","observation_id":"4ca1b901-070d-4f1c-9fb0-e340c78397db","resolution":{"observed_at":"2026-08-09T19:55:17.257871Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.817738Z","title":"Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":4,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.817738Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:57163da018e0cc55fe05db725936b4aa0cc4df127d515d3921c95b6feaf49985","observation_id":"d5708693-411c-468b-9259-c7ba5db4bc02","resolution":{"observed_at":"2026-08-09T19:55:16.817738Z","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":"2026-08-09T19:55:16.822134Z","title":"Mask r-cnn","venue":null,"work_id":null,"year":2017},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":5,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.822134Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:83a7738620029479a990b42f5cefe13fedb27aec4f07c57ae12ecd38ee390071","observation_id":"b90c6786-79de-4e48-849a-7a618c375faf","resolution":{"observed_at":"2026-08-09T19:55:16.822134Z","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":"2026-08-09T19:55:17.228569Z","title":null,"venue":null,"work_id":"9d582c0b-85f7-4676-bfd1-8530bf907c62","year":2002},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":6,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.826472Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:1a5d24d4486fb9e868f28716e2d1419430641166498e91c5550791476fd51b75","observation_id":"1505de63-0160-4cd0-a5cd-556af4cabe48","resolution":{"observed_at":"2026-08-09T19:55:17.231960Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.830862Z","title":"Multi-task learning using uncertainty to weigh losses for scene geometry and semantics","venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":7,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.830862Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:4a0ce9105cdc9fc007e7cbf5b3fa6a1e5f4bb5bfa61a463f40b101d2a79c543f","observation_id":"8c2c076f-e279-4b9e-82bb-806563a4433c","resolution":{"observed_at":"2026-08-09T19:55:16.830862Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"1412.6980","last_updated":"2017-01-30T01:27:54Z","snapshot_observed_at":"2026-07-06T02:11:23.670680Z","submitted_at":"2014-12-22T13:54:29Z","title":"Adam: A Method for Stochastic Optimization","version":9},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"1412.6980","snapshot_observed_at":"2026-08-09T19:55:16.835027Z","title":null,"venue":null,"work_id":null,"year":2014},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":8,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.835027Z"},"links":{"cited_paper":"/paper/1412.6980","citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:5e5eff858511f10fd13eee3a490009c83fadf05474fab93a1dc67733b18bcfad","observation_id":"e9792c14-6ec6-41da-b084-e878caf1aaeb","resolution":{"observed_at":"2026-08-09T19:55:16.835027Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2111.10603","last_updated":"2022-07-27T15:44:33Z","snapshot_observed_at":"2026-08-13T17:30:22.706063Z","submitted_at":"2021-11-20T14:28:32Z","title":"Reasonable Effectiveness of Random Weighting: A Litmus Test for Multi-Task Learning","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2111.10603","snapshot_observed_at":"2026-08-09T19:55:16.839556Z","title":null,"venue":null,"work_id":null,"year":2021},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":9,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.839556Z"},"links":{"cited_paper":"/paper/2111.10603","citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:b199215422c9433e8919240faafadb0c36560144c444402e85bbef8700740078","observation_id":"47f94d36-0331-4981-92f6-2dcb37f410a9","resolution":{"observed_at":"2026-08-09T19:55:16.839556Z","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":"2026-08-09T19:55:17.207605Z","title":"Mtmamba: Enhancing multi-task dense scene understanding by mamba-based decoders","venue":null,"work_id":"af7191b1-b3bc-4525-a342-c46a4cd7f57b","year":2025},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":10,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.844038Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:3787815caceb8c6c1a953b2683de68cca1c220847e2a16c9f8431a8eed97f5c8","observation_id":"6c3ad359-f39b-4ff8-add5-7e184a58554d","resolution":{"observed_at":"2026-08-09T19:55:17.212282Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.196877Z","title":"Conflict-averse gradient descent for multi-task learning","venue":null,"work_id":"8fac0bad-b815-4bb4-97d8-9ec3193110e6","year":2021},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":11,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.848224Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:bf57c9ba5b5daf640921bda4936fdba1fa6d2ded9cdf7d338d3f032c7a161da6","observation_id":"ef1f616c-d3b6-4c67-8c30-ab2b5da828f1","resolution":{"observed_at":"2026-08-09T19:55:17.200355Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.184892Z","title":"Famo: Fast adaptive multitask optimization","venue":null,"work_id":"1bd5db74-3765-4504-9885-da8912de9c8a","year":2023},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":12,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.852368Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:8cdc48479e2ccb47334efa0e4344abf7837daba8b1612ecb0df1fd903cb376f6","observation_id":"26e2906a-3a78-4b7d-8375-fd2b60d3ec8d","resolution":{"observed_at":"2026-08-09T19:55:17.189283Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.171916Z","title":"Towards impartial multi-task learning","venue":null,"work_id":"0920793e-c021-4d16-80f5-cb3c99db7411","year":2021},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":13,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.856745Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:c635374039bc3204f48ea8a47a7e663d297287fb4987909a313224c8af7fef86","observation_id":"72484afc-7ef1-4069-9261-cec4ebe5b2bc","resolution":{"observed_at":"2026-08-09T19:55:17.176281Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.159205Z","title":null,"venue":null,"work_id":"68248733-3bef-4c1c-9031-7ace481a7b25","year":2019},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":14,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.860780Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:9f246fd595f3a93e7d0633ce880d9a7dc5dc6d9917fd03532bdf87a43870e3ad","observation_id":"e79d41f0-9054-456a-8a66-18bfbc4eb3e5","resolution":{"observed_at":"2026-08-09T19:55:17.163332Z","resolver_source":"raw_fallback","status":"unresolved"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.864848Z","title":"Deep learning face attributes in the wild","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":15,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.864848Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:dc6a56507a71be64d4792d2714ee3ea4abe3027c5a01153865cb91a01e78d163","observation_id":"9dd34d52-2bc1-4eb0-95db-4746f7ffd1b5","resolution":{"observed_at":"2026-08-09T19:55:16.864848Z","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":"2026-08-09T19:55:16.868664Z","title":"Fully convolutional networks for semantic segmentation","venue":null,"work_id":null,"year":2015},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":16,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.868664Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:603947901d43cf8a7798049a806b550b194606454026404d2dc660907afca0e1","observation_id":"667faedd-712a-4be5-8017-ac133c33bb0d","resolution":{"observed_at":"2026-08-09T19:55:16.868664Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}},{"citation":{"cited_paper":{"arxiv_id":"2202.01017","last_updated":"2022-07-08T12:00:51Z","snapshot_observed_at":"2026-08-13T16:48:33.513314Z","submitted_at":"2022-02-02T13:21:53Z","title":"Multi-Task Learning as a Bargaining Game","version":2},"cited_work":{"arxiv_id":null,"doi":null,"metadata_source":null,"pith_arxiv_id":"2202.01017","snapshot_observed_at":"2026-08-09T19:55:16.872799Z","title":"Multi-task learning as a bargaining game","venue":null,"work_id":null,"year":2022},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":17,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.872799Z"},"links":{"cited_paper":"/paper/2202.01017","citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:7ecfc82b96d167cc812dde07a14b77b17297633eee147ef5065a896ad4d24f78","observation_id":"b7ab6734-944b-49c7-b94a-00b79bdac299","resolution":{"observed_at":"2026-08-09T19:55:16.872799Z","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":"2026-08-09T19:55:17.132552Z","title":"and Koltun, V","venue":null,"work_id":"ecfa2ba7-e3e2-4c8f-9839-eb1376e0ed61","year":2018},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":18,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.877294Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:dfacbc9786e2130b90ad68a9156df42accb7cd2e5224c4682c7216f4e974f0ce","observation_id":"ba82d28e-938b-4c85-9415-75eba713f1c8","resolution":{"observed_at":"2026-08-09T19:55:17.136615Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.118839Z","title":"Indoor segmentation and support inference from rgbd images","venue":null,"work_id":"90c47c40-b190-4151-a120-60affd85ad26","year":2012},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":19,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.881298Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:c556d0b7cc38fa3a16c557511ad9bdebb0acc720858670ef68884e632c1e3172","observation_id":"b9f4d012-1fcf-49b8-8203-5bfe2013b990","resolution":{"observed_at":"2026-08-09T19:55:17.123472Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.106691Z","title":"and Zhang, A","venue":null,"work_id":"efbab530-631c-4a0b-a223-1f8f5f35c3cb","year":2021},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":20,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.885396Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:08479aa43a3e01d06aae542825293a58b5d3eb6a1042c087acb65b3dbf49f099","observation_id":"3ce16ca8-26ab-4f2d-8d31-974b5d0fc98d","resolution":{"observed_at":"2026-08-09T19:55:17.110574Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.094237Z","title":"ptflops: a flops counting tool for neural networks in pytorch framework, 2018-2024","venue":null,"work_id":"9a8dd1f3-33f3-4785-bb92-b052a8eaddef","year":2018},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":21,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.889718Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:b860ae3de5844774ed9336b93371424f81ec24f088d343fc78732902845f2a99","observation_id":"d005711a-f927-400e-9af3-391b54f137b1","resolution":{"observed_at":"2026-08-09T19:55:17.098343Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.893918Z","title":null,"venue":null,"work_id":null,"year":2018},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":22,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.893918Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:fb846dc4bb0d31a9014c8bc93689d88054173b4c2d593860dbef9d6339cf4a8e","observation_id":"2bb83cff-fc82-4450-825b-8d9a0c3dbca2","resolution":{"observed_at":"2026-08-09T19:55:16.893918Z","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":"2026-08-09T19:55:17.073239Z","title":"Multi-task learning for dense prediction tasks: A survey","venue":null,"work_id":"54d374ff-78dc-445f-af13-0391f689e9ca","year":2021},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":23,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.898048Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:8d7ae60b4d88c34c0deda391cfc1481d478f5f8dd340ef3a3dfccb031e332ded","observation_id":"f678d07e-588b-4400-85bd-6a5cc821dec3","resolution":{"observed_at":"2026-08-09T19:55:17.078029Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.059063Z","title":"Direction-oriented multi-objective learning: Simple and provable stochastic algorithms","venue":null,"work_id":"070dba84-39fb-4a72-a765-ced71889614d","year":2024},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":24,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.902728Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:e8fd4b4898c959aeacb7a9b955cf6ce33794e86f33cfc44c8e8b869d12e27e0c","observation_id":"f1584432-83ff-456d-b7bd-201882151b3c","resolution":{"observed_at":"2026-08-09T19:55:17.063997Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.907016Z","title":"Multi-task reinforcement learning with soft modularization","venue":null,"work_id":null,"year":2020},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":25,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.907016Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:50cbdf70cd6479f90af8e1506a9f3a4118c21765c15f5710160cd13091632eaf","observation_id":"7d3317a5-a1e0-4691-8dd2-e609ee0d7b6e","resolution":{"observed_at":"2026-08-09T19:55:16.907016Z","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":"2026-08-09T19:55:17.035592Z","title":"Gradient surgery for multi-task learning","venue":null,"work_id":"3bcdf9d1-8223-49b0-ba5c-a4b1cc9ea9ba","year":2020},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":26,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.911053Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:24a0d44fad6c435ba6d5c3a2af9c477b81eeaedf8d6069571bcc29a26757a872","observation_id":"18544085-0580-446f-90ed-db3fd4097c65","resolution":{"observed_at":"2026-08-09T19:55:17.039601Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.020881Z","title":"Meta-world: A benchmark and evaluation for multi-task and meta reinforcement learning","venue":null,"work_id":"5ba63a2c-b493-44d0-88d2-ec4e1c62d745","year":2020},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":27,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.915114Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:07544bb45fb8a727cacdfde850e1b591fd432aab7d2461738acfa86e071c23b6","observation_id":"bb91fc5b-fc1c-4c1c-b429-1c02da8054ab","resolution":{"observed_at":"2026-08-09T19:55:17.025168Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:17.005822Z","title":"Sgw-based multi-task learning in vision tasks","venue":null,"work_id":"25ee3198-58d9-46a9-b427-4ee7cc84625f","year":2025},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":28,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.919227Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:aa4389329ba0d0725185df55027345c38f18e62aabbf0506ae2364982981fc59","observation_id":"6596d8de-40bd-4a02-a0d6-3e5d52c45e7a","resolution":{"observed_at":"2026-08-09T19:55:17.011422Z","resolver_source":"raw_fallback","status":"verified_fuzzy"},"standing_notice":{"events":[],"observation":"No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.","reason":null,"source_receipts":[{"observed_at":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+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":"2026-08-09T19:55:16.923519Z","title":"write newline","venue":null,"work_id":null,"year":null},"citing_paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone","version":1},"reference_index":29,"source":"arxiv_source","source_observed_at":"2026-08-09T19:55:16.923519Z"},"links":{"citing_paper":"/paper/2502.00217"},"observation_digest":"sha256:9cee7d4d9b816675a5d2cc15b20240682e1beb4c68deaf269512c768a396b672","observation_id":"ce8b91ec-be6b-42f6-b3bc-57591737b531","resolution":{"observed_at":"2026-08-09T19:55:16.923519Z","resolver_source":null,"status":"unresolved"},"standing_notice":{"events":[],"reason":"canonical_work_link_unavailable","source_receipts":[],"state":"unavailable"}}],"paper":{"arxiv_id":"2502.00217","last_updated":"2025-01-31T23:11:12Z","latest_version":1,"primary_category":"cs.LG","snapshot_observed_at":"2026-08-10T13:09:49.272157Z","submitted_at":"2025-01-31T23:11:12Z","title":"Fantastic Multi-Task Gradient Updates and How to Find Them In a Cone"},"reference_resolution":{"displayed":29,"state_counts":{"malformed_identifier":0,"metadata_mismatch":0,"parse_uncertain":0,"unresolved":14,"verified_exact":0,"verified_fuzzy":15},"total_outbound_references":29},"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":"2026-08-14T06:32:32.682623+00:00","source":"crossref"},{"observed_at":"2026-08-14T06:32:18.44784+00:00","source":"retraction_watch"}],"thesis":"As of 14 August 2026, this Paper Citation Record lists 29 of 29 outbound references and 0 inbound Pith citation observations for arXiv:2502.00217."}