{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:C5TUS6HMELPFDBW7XSO3XS3OSO","short_pith_number":"pith:C5TUS6HM","canonical_record":{"source":{"id":"1812.02224","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-12-05T21:00:44Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"fd007adb59d6a987b9d74f17bb5fa345931aefd0f6b217e31639fcf9f995eddd","abstract_canon_sha256":"b8b9f17b63868b170481a76be2c85f73bdf2b7028499ca0cf2e005b819cf19f8"},"schema_version":"1.0"},"canonical_sha256":"17674978ec22de5186dfbc9dbbcb6e93ad283122c955b18ad0ebfa7335951096","source":{"kind":"arxiv","id":"1812.02224","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1812.02224","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"arxiv_version","alias_value":"1812.02224v2","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1812.02224","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"pith_short_12","alias_value":"C5TUS6HMELPF","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"pith_short_16","alias_value":"C5TUS6HMELPFDBW7","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"pith_short_8","alias_value":"C5TUS6HM","created_at":"2026-07-05T01:54:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:C5TUS6HMELPFDBW7XSO3XS3OSO","target":"record","payload":{"canonical_record":{"source":{"id":"1812.02224","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-12-05T21:00:44Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"fd007adb59d6a987b9d74f17bb5fa345931aefd0f6b217e31639fcf9f995eddd","abstract_canon_sha256":"b8b9f17b63868b170481a76be2c85f73bdf2b7028499ca0cf2e005b819cf19f8"},"schema_version":"1.0"},"canonical_sha256":"17674978ec22de5186dfbc9dbbcb6e93ad283122c955b18ad0ebfa7335951096","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:54:39.568310Z","signature_b64":"9hVWvvccN9qTBoWSWzUxbw8EJayQ//VfSOul7D91aTVSjC53CYRUYoBSZcFY13KXaIDMUMt6RXCyVBrf4N8FCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"17674978ec22de5186dfbc9dbbcb6e93ad283122c955b18ad0ebfa7335951096","last_reissued_at":"2026-07-05T01:54:39.567848Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:54:39.567848Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1812.02224","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T01:54:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y/Jm3RmpryVwgN0mFPUR5tkpuQRTDE6lJ3oRNgOKR2pdIwIzSyAm7jUspikAdeNPHwhUTSELLNQfPH4L3fJRAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-25T00:40:09.513548Z"},"content_sha256":"79fafd366252a713b5084fc8c193d81bda51840272234c68da63b2908dd6048d","schema_version":"1.0","event_id":"sha256:79fafd366252a713b5084fc8c193d81bda51840272234c68da63b2908dd6048d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:C5TUS6HMELPFDBW7XSO3XS3OSO","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Adapting Auxiliary Losses Using Gradient Similarity","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Balaji Lakshminarayanan, Mehrdad Farajtabar, Razvan Pascanu, Siddhant M. Jayakumar, Wojciech M. Czarnecki, Yunshu Du","submitted_at":"2018-12-05T21:00:44Z","abstract_excerpt":"One approach to deal with the statistical inefficiency of neural networks is to rely on auxiliary losses that help to build useful representations. However, it is not always trivial to know if an auxiliary task will be helpful for the main task and when it could start hurting. We propose to use the cosine similarity between gradients of tasks as an adaptive weight to detect when an auxiliary loss is helpful to the main loss. We show that our approach is guaranteed to converge to critical points of the main task and demonstrate the practical usefulness of the proposed algorithm in a few domains"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1812.02224","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/1812.02224/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T01:54:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"U5kN3MR+LnXGT8blhL63DUeMdmlwNJI8Fnw98OG4GiirZyT087/OrgVesv8971vqYOovPg3fK6zOImbI2fTtBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-25T00:40:09.513917Z"},"content_sha256":"f7b4d3cb7482e3e9e0e2413aa7bef8f66e670199706bfb8ac9b8b15d23ced210","schema_version":"1.0","event_id":"sha256:f7b4d3cb7482e3e9e0e2413aa7bef8f66e670199706bfb8ac9b8b15d23ced210"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/C5TUS6HMELPFDBW7XSO3XS3OSO/bundle.json","state_url":"https://pith.science/pith/C5TUS6HMELPFDBW7XSO3XS3OSO/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/C5TUS6HMELPFDBW7XSO3XS3OSO/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-25T00:40:09Z","links":{"resolver":"https://pith.science/pith/C5TUS6HMELPFDBW7XSO3XS3OSO","bundle":"https://pith.science/pith/C5TUS6HMELPFDBW7XSO3XS3OSO/bundle.json","state":"https://pith.science/pith/C5TUS6HMELPFDBW7XSO3XS3OSO/state.json","well_known_bundle":"https://pith.science/.well-known/pith/C5TUS6HMELPFDBW7XSO3XS3OSO/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:C5TUS6HMELPFDBW7XSO3XS3OSO","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b8b9f17b63868b170481a76be2c85f73bdf2b7028499ca0cf2e005b819cf19f8","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-12-05T21:00:44Z","title_canon_sha256":"fd007adb59d6a987b9d74f17bb5fa345931aefd0f6b217e31639fcf9f995eddd"},"schema_version":"1.0","source":{"id":"1812.02224","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1812.02224","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"arxiv_version","alias_value":"1812.02224v2","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1812.02224","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"pith_short_12","alias_value":"C5TUS6HMELPF","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"pith_short_16","alias_value":"C5TUS6HMELPFDBW7","created_at":"2026-07-05T01:54:39Z"},{"alias_kind":"pith_short_8","alias_value":"C5TUS6HM","created_at":"2026-07-05T01:54:39Z"}],"graph_snapshots":[{"event_id":"sha256:f7b4d3cb7482e3e9e0e2413aa7bef8f66e670199706bfb8ac9b8b15d23ced210","target":"graph","created_at":"2026-07-05T01:54:39Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/1812.02224/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"One approach to deal with the statistical inefficiency of neural networks is to rely on auxiliary losses that help to build useful representations. However, it is not always trivial to know if an auxiliary task will be helpful for the main task and when it could start hurting. We propose to use the cosine similarity between gradients of tasks as an adaptive weight to detect when an auxiliary loss is helpful to the main loss. We show that our approach is guaranteed to converge to critical points of the main task and demonstrate the practical usefulness of the proposed algorithm in a few domains","authors_text":"Balaji Lakshminarayanan, Mehrdad Farajtabar, Razvan Pascanu, Siddhant M. Jayakumar, Wojciech M. Czarnecki, Yunshu Du","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-12-05T21:00:44Z","title":"Adapting Auxiliary Losses Using Gradient Similarity"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1812.02224","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:79fafd366252a713b5084fc8c193d81bda51840272234c68da63b2908dd6048d","target":"record","created_at":"2026-07-05T01:54:39Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b8b9f17b63868b170481a76be2c85f73bdf2b7028499ca0cf2e005b819cf19f8","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-12-05T21:00:44Z","title_canon_sha256":"fd007adb59d6a987b9d74f17bb5fa345931aefd0f6b217e31639fcf9f995eddd"},"schema_version":"1.0","source":{"id":"1812.02224","kind":"arxiv","version":2}},"canonical_sha256":"17674978ec22de5186dfbc9dbbcb6e93ad283122c955b18ad0ebfa7335951096","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"17674978ec22de5186dfbc9dbbcb6e93ad283122c955b18ad0ebfa7335951096","first_computed_at":"2026-07-05T01:54:39.567848Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T01:54:39.567848Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9hVWvvccN9qTBoWSWzUxbw8EJayQ//VfSOul7D91aTVSjC53CYRUYoBSZcFY13KXaIDMUMt6RXCyVBrf4N8FCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T01:54:39.568310Z","signed_message":"canonical_sha256_bytes"},"source_id":"1812.02224","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:79fafd366252a713b5084fc8c193d81bda51840272234c68da63b2908dd6048d","sha256:f7b4d3cb7482e3e9e0e2413aa7bef8f66e670199706bfb8ac9b8b15d23ced210"],"state_sha256":"ed55f55b59175e2997f77cf58e01888d174264addbf992d086e6a6e06d73cb20"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CGYwUqitEB4SNyMFqijGfBUB3EPn3+8iTVhSpvZq9BrYTXn+Vf5SPQLQREsIcFlSaZHArvXsEDOinDoCVwERCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-25T00:40:09.515969Z","bundle_sha256":"c7c96d6f6b640cd141c3204425783d44d2de47e9d389976aa91dc956b04c555d"}}