{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:VAE6TVGPA7YT5VUO6Y25YMANLW","short_pith_number":"pith:VAE6TVGP","canonical_record":{"source":{"id":"2606.02221","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-01T13:20:58Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"2413f530025bb17cd8fc8febfbd55daf9b55df12502e123a0bdcc20da40f0530","abstract_canon_sha256":"462d8cb899aa7f7581376de8ea4dd9d0e0572ec8f05ef814fca9d5bffae91071"},"schema_version":"1.0"},"canonical_sha256":"a809e9d4cf07f13ed68ef635dc300d5d84f9340e97ff2824bd1524eb784fecca","source":{"kind":"arxiv","id":"2606.02221","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.02221","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"arxiv_version","alias_value":"2606.02221v1","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02221","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"pith_short_12","alias_value":"VAE6TVGPA7YT","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"pith_short_16","alias_value":"VAE6TVGPA7YT5VUO","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"pith_short_8","alias_value":"VAE6TVGP","created_at":"2026-06-02T03:04:53Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:VAE6TVGPA7YT5VUO6Y25YMANLW","target":"record","payload":{"canonical_record":{"source":{"id":"2606.02221","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-01T13:20:58Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"2413f530025bb17cd8fc8febfbd55daf9b55df12502e123a0bdcc20da40f0530","abstract_canon_sha256":"462d8cb899aa7f7581376de8ea4dd9d0e0572ec8f05ef814fca9d5bffae91071"},"schema_version":"1.0"},"canonical_sha256":"a809e9d4cf07f13ed68ef635dc300d5d84f9340e97ff2824bd1524eb784fecca","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T03:04:53.515241Z","signature_b64":"MGP9tZj+Qt1SNufj833K3NzPccRL/FHF9RvfFg7SRXCK2TfZP3NCprx3TVZaHZhwspC9hXstxrWJgaiOIK9fCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a809e9d4cf07f13ed68ef635dc300d5d84f9340e97ff2824bd1524eb784fecca","last_reissued_at":"2026-06-02T03:04:53.514829Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T03:04:53.514829Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.02221","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":"2026-06-02T03:04:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nekrmD1gjl8/Sux7qaZtB8gFZwakLZ8mFYWTXJI6LbtNEyhQ9vrDS1gJ/QZnueG4Xmu3mKqtThEk67k6uMUaAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T22:09:24.767306Z"},"content_sha256":"d74ca0d3b5560ad637c4a0a2c7fa7a369b0bf42850685dc634db7dbbcd025ebd","schema_version":"1.0","event_id":"sha256:d74ca0d3b5560ad637c4a0a2c7fa7a369b0bf42850685dc634db7dbbcd025ebd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:VAE6TVGPA7YT5VUO6Y25YMANLW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"CORE-MTL: Rethinking Gradient Balancing via Causal Orthogonal Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Chengfeng Wu, Jingge Wang, Tao Zou, Yanru Wu","submitted_at":"2026-06-01T13:20:58Z","abstract_excerpt":"Multi-task learning (MTL) aims to construct a joint model for multiple tasks by sharing a common representation across domains. To achieve this goal, existing optimization-centric methods either balance task gradients or modify the shared architecture. However, as these approaches remain agnostic to the content of the shared representation, they fail to disentangle task-relevant structure from spurious context, leading to negative transfer and poor generalization. To overcome this limitation, we propose Causal Orthogonal Representations for Multi-Task Learning (CORE-MTL), a causally motivated "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02221","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/2606.02221/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":"2026-06-02T03:04:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"fFoL+zcHn+i1YjHpNfDlioTxXfpW6KkawL3LjKerB9f7tm1IxyyKqWDLQNw7Uux/ZvSJKZLjtqRqvca3Az35Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-22T22:09:24.767854Z"},"content_sha256":"f3dd9e6432ff762a2011ce581908838f38dcde9822e73565654e2cd4d394b051","schema_version":"1.0","event_id":"sha256:f3dd9e6432ff762a2011ce581908838f38dcde9822e73565654e2cd4d394b051"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/VAE6TVGPA7YT5VUO6Y25YMANLW/bundle.json","state_url":"https://pith.science/pith/VAE6TVGPA7YT5VUO6Y25YMANLW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/VAE6TVGPA7YT5VUO6Y25YMANLW/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":"2026-08-22T22:09:24Z","links":{"resolver":"https://pith.science/pith/VAE6TVGPA7YT5VUO6Y25YMANLW","bundle":"https://pith.science/pith/VAE6TVGPA7YT5VUO6Y25YMANLW/bundle.json","state":"https://pith.science/pith/VAE6TVGPA7YT5VUO6Y25YMANLW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/VAE6TVGPA7YT5VUO6Y25YMANLW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:VAE6TVGPA7YT5VUO6Y25YMANLW","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":"462d8cb899aa7f7581376de8ea4dd9d0e0572ec8f05ef814fca9d5bffae91071","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-01T13:20:58Z","title_canon_sha256":"2413f530025bb17cd8fc8febfbd55daf9b55df12502e123a0bdcc20da40f0530"},"schema_version":"1.0","source":{"id":"2606.02221","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.02221","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"arxiv_version","alias_value":"2606.02221v1","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02221","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"pith_short_12","alias_value":"VAE6TVGPA7YT","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"pith_short_16","alias_value":"VAE6TVGPA7YT5VUO","created_at":"2026-06-02T03:04:53Z"},{"alias_kind":"pith_short_8","alias_value":"VAE6TVGP","created_at":"2026-06-02T03:04:53Z"}],"graph_snapshots":[{"event_id":"sha256:f3dd9e6432ff762a2011ce581908838f38dcde9822e73565654e2cd4d394b051","target":"graph","created_at":"2026-06-02T03:04:53Z","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/2606.02221/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Multi-task learning (MTL) aims to construct a joint model for multiple tasks by sharing a common representation across domains. To achieve this goal, existing optimization-centric methods either balance task gradients or modify the shared architecture. However, as these approaches remain agnostic to the content of the shared representation, they fail to disentangle task-relevant structure from spurious context, leading to negative transfer and poor generalization. To overcome this limitation, we propose Causal Orthogonal Representations for Multi-Task Learning (CORE-MTL), a causally motivated ","authors_text":"Chengfeng Wu, Jingge Wang, Tao Zou, Yanru Wu","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-01T13:20:58Z","title":"CORE-MTL: Rethinking Gradient Balancing via Causal Orthogonal Representations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02221","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:d74ca0d3b5560ad637c4a0a2c7fa7a369b0bf42850685dc634db7dbbcd025ebd","target":"record","created_at":"2026-06-02T03:04:53Z","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":"462d8cb899aa7f7581376de8ea4dd9d0e0572ec8f05ef814fca9d5bffae91071","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-01T13:20:58Z","title_canon_sha256":"2413f530025bb17cd8fc8febfbd55daf9b55df12502e123a0bdcc20da40f0530"},"schema_version":"1.0","source":{"id":"2606.02221","kind":"arxiv","version":1}},"canonical_sha256":"a809e9d4cf07f13ed68ef635dc300d5d84f9340e97ff2824bd1524eb784fecca","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a809e9d4cf07f13ed68ef635dc300d5d84f9340e97ff2824bd1524eb784fecca","first_computed_at":"2026-06-02T03:04:53.514829Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-02T03:04:53.514829Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MGP9tZj+Qt1SNufj833K3NzPccRL/FHF9RvfFg7SRXCK2TfZP3NCprx3TVZaHZhwspC9hXstxrWJgaiOIK9fCw==","signature_status":"signed_v1","signed_at":"2026-06-02T03:04:53.515241Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.02221","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d74ca0d3b5560ad637c4a0a2c7fa7a369b0bf42850685dc634db7dbbcd025ebd","sha256:f3dd9e6432ff762a2011ce581908838f38dcde9822e73565654e2cd4d394b051"],"state_sha256":"845fcd4efbd9c67c264744619e74621c8aa6c41693aef1ab351609787b37391a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1SFgV2tmyMiPIPgXlyyC0tn23gD9/NU3V5jq3cL6LxGUrGV7VSINiCToHbcmINzakKgqcQMJ85tXmBeW9IJvBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-22T22:09:24.773446Z","bundle_sha256":"17bef7dfea3086c25469cbc9d7d2859473a3b845ece211dc10f760629bd36c2d"}}