{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:T5256D2NMZZMLVEITHGEEJB6OQ","short_pith_number":"pith:T5256D2N","canonical_record":{"source":{"id":"2204.05698","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2022-04-12T10:52:28Z","cross_cats_sorted":[],"title_canon_sha256":"37812c824c77351bd9fd5fa93567df108ae667e05b3841fa2466839c46fd0bc4","abstract_canon_sha256":"096ba48fbdbf11d5ae43df7ab2fd2ecba2d078a179efd648d46c262f345d89dd"},"schema_version":"1.0"},"canonical_sha256":"9f75df0f4d6672c5d48899cc42243e742954caa34b7bb4d4e122baf22b5647a2","source":{"kind":"arxiv","id":"2204.05698","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.05698","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"arxiv_version","alias_value":"2204.05698v1","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.05698","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"pith_short_12","alias_value":"T5256D2NMZZM","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"pith_short_16","alias_value":"T5256D2NMZZMLVEI","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"pith_short_8","alias_value":"T5256D2N","created_at":"2026-07-05T04:14:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:T5256D2NMZZMLVEITHGEEJB6OQ","target":"record","payload":{"canonical_record":{"source":{"id":"2204.05698","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2022-04-12T10:52:28Z","cross_cats_sorted":[],"title_canon_sha256":"37812c824c77351bd9fd5fa93567df108ae667e05b3841fa2466839c46fd0bc4","abstract_canon_sha256":"096ba48fbdbf11d5ae43df7ab2fd2ecba2d078a179efd648d46c262f345d89dd"},"schema_version":"1.0"},"canonical_sha256":"9f75df0f4d6672c5d48899cc42243e742954caa34b7bb4d4e122baf22b5647a2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:14:01.306222Z","signature_b64":"Erui/fX6FgRfOxM5WkXaeNrsgDBvCPmz2IBnX7Fx/r9rmXEkTJJiwPqd3eGlTBGtXWp8xKddmch27EjZ97o3Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9f75df0f4d6672c5d48899cc42243e742954caa34b7bb4d4e122baf22b5647a2","last_reissued_at":"2026-07-05T04:14:01.305754Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:14:01.305754Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2204.05698","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T04:14:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xx1VwEzWK2xzV8Mc6vHDv1YGff52VsFb0Yl5ZRR8ja/JA2PqEdhYuH9SckeR9VRFaGOEdYMZWAEsRVrWpcE5Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T05:47:23.244688Z"},"content_sha256":"5ece9f08e1f196f9b27308fe0d5b7b36b985db1d6c57159b1d6c5cdab8111c44","schema_version":"1.0","event_id":"sha256:5ece9f08e1f196f9b27308fe0d5b7b36b985db1d6c57159b1d6c5cdab8111c44"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:T5256D2NMZZMLVEITHGEEJB6OQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Medusa: Universal Feature Learning via Attentional Multitasking","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jaime Spencer, Richard Bowden, Simon Hadfield","submitted_at":"2022-04-12T10:52:28Z","abstract_excerpt":"Recent approaches to multi-task learning (MTL) have focused on modelling connections between tasks at the decoder level. This leads to a tight coupling between tasks, which need retraining if a new task is inserted or removed. We argue that MTL is a stepping stone towards universal feature learning (UFL), which is the ability to learn generic features that can be applied to new tasks without retraining.\n  We propose Medusa to realize this goal, designing task heads with dual attention mechanisms. The shared feature attention masks relevant backbone features for each task, allowing it to learn "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.05698","kind":"arxiv","version":1},"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/2204.05698/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":"2026-07-05T04:14:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jpQkdNFFO18jlNassvV7gbv90VHkPkYBGEw7xEPt215PHyn5DGTkwH15XggFMtaQu/Mep3cAicgLdGiwBgvxDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T05:47:23.245327Z"},"content_sha256":"4f2843e90a88db9b557b61f5534b07c4f6aba36e4bb564dba25cff977cff480a","schema_version":"1.0","event_id":"sha256:4f2843e90a88db9b557b61f5534b07c4f6aba36e4bb564dba25cff977cff480a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/T5256D2NMZZMLVEITHGEEJB6OQ/bundle.json","state_url":"https://pith.science/pith/T5256D2NMZZMLVEITHGEEJB6OQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/T5256D2NMZZMLVEITHGEEJB6OQ/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":"2026-07-07T05:47:23Z","links":{"resolver":"https://pith.science/pith/T5256D2NMZZMLVEITHGEEJB6OQ","bundle":"https://pith.science/pith/T5256D2NMZZMLVEITHGEEJB6OQ/bundle.json","state":"https://pith.science/pith/T5256D2NMZZMLVEITHGEEJB6OQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/T5256D2NMZZMLVEITHGEEJB6OQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:T5256D2NMZZMLVEITHGEEJB6OQ","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":"096ba48fbdbf11d5ae43df7ab2fd2ecba2d078a179efd648d46c262f345d89dd","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2022-04-12T10:52:28Z","title_canon_sha256":"37812c824c77351bd9fd5fa93567df108ae667e05b3841fa2466839c46fd0bc4"},"schema_version":"1.0","source":{"id":"2204.05698","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2204.05698","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"arxiv_version","alias_value":"2204.05698v1","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.05698","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"pith_short_12","alias_value":"T5256D2NMZZM","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"pith_short_16","alias_value":"T5256D2NMZZMLVEI","created_at":"2026-07-05T04:14:01Z"},{"alias_kind":"pith_short_8","alias_value":"T5256D2N","created_at":"2026-07-05T04:14:01Z"}],"graph_snapshots":[{"event_id":"sha256:4f2843e90a88db9b557b61f5534b07c4f6aba36e4bb564dba25cff977cff480a","target":"graph","created_at":"2026-07-05T04:14:01Z","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/2204.05698/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent approaches to multi-task learning (MTL) have focused on modelling connections between tasks at the decoder level. This leads to a tight coupling between tasks, which need retraining if a new task is inserted or removed. We argue that MTL is a stepping stone towards universal feature learning (UFL), which is the ability to learn generic features that can be applied to new tasks without retraining.\n  We propose Medusa to realize this goal, designing task heads with dual attention mechanisms. The shared feature attention masks relevant backbone features for each task, allowing it to learn ","authors_text":"Jaime Spencer, Richard Bowden, Simon Hadfield","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2022-04-12T10:52:28Z","title":"Medusa: Universal Feature Learning via Attentional Multitasking"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.05698","kind":"arxiv","version":1},"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":"sha256:5ece9f08e1f196f9b27308fe0d5b7b36b985db1d6c57159b1d6c5cdab8111c44","target":"record","created_at":"2026-07-05T04:14:01Z","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":"096ba48fbdbf11d5ae43df7ab2fd2ecba2d078a179efd648d46c262f345d89dd","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2022-04-12T10:52:28Z","title_canon_sha256":"37812c824c77351bd9fd5fa93567df108ae667e05b3841fa2466839c46fd0bc4"},"schema_version":"1.0","source":{"id":"2204.05698","kind":"arxiv","version":1}},"canonical_sha256":"9f75df0f4d6672c5d48899cc42243e742954caa34b7bb4d4e122baf22b5647a2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9f75df0f4d6672c5d48899cc42243e742954caa34b7bb4d4e122baf22b5647a2","first_computed_at":"2026-07-05T04:14:01.305754Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T04:14:01.305754Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Erui/fX6FgRfOxM5WkXaeNrsgDBvCPmz2IBnX7Fx/r9rmXEkTJJiwPqd3eGlTBGtXWp8xKddmch27EjZ97o3Bg==","signature_status":"signed_v1","signed_at":"2026-07-05T04:14:01.306222Z","signed_message":"canonical_sha256_bytes"},"source_id":"2204.05698","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5ece9f08e1f196f9b27308fe0d5b7b36b985db1d6c57159b1d6c5cdab8111c44","sha256:4f2843e90a88db9b557b61f5534b07c4f6aba36e4bb564dba25cff977cff480a"],"state_sha256":"3ef9ffe5849212c936b988c040c3827bdbae6f9a1bce51aa0debc2ea7affbdd1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LEgT/UU14kJWgvGwlPd+DNkpIK/7aSoVW51NyerFjB2joIQWTqQX9Qc/A5tFGLflPX2tccdm555y6G+4z0irDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T05:47:23.248745Z","bundle_sha256":"10faf45d58fe9a181136b2dcbbeaea3e63776ddbd6d847031c3f4b7e36c09427"}}