{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:OR4HZNQUPUOVNEFV3LOFBKIGCR","short_pith_number":"pith:OR4HZNQU","schema_version":"1.0","canonical_sha256":"74787cb6147d1d5690b5dadc50a906147364e8930fa2ac69ef00385e23873a8b","source":{"kind":"arxiv","id":"2210.14793","version":1},"attestation_state":"computed","paper":{"title":"M$^3$ViT: Mixture-of-Experts Vision Transformer for Efficient Multi-task Learning with Model-Accelerator Co-design","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cong Hao, Hanxue Liang, Kai Zou, Rishov Sarkar, Tianlong Chen, Yu Cheng, Zhangyang Wang, Zhiwen Fan, Ziyu Jiang","submitted_at":"2022-10-26T15:40:24Z","abstract_excerpt":"Multi-task learning (MTL) encapsulates multiple learned tasks in a single model and often lets those tasks learn better jointly. However, when deploying MTL onto those real-world systems that are often resource-constrained or latency-sensitive, two prominent challenges arise: (i) during training, simultaneously optimizing all tasks is often difficult due to gradient conflicts across tasks; (ii) at inference, current MTL regimes have to activate nearly the entire model even to just execute a single task. Yet most real systems demand only one or two tasks at each moment, and switch between tasks"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2210.14793","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-10-26T15:40:24Z","cross_cats_sorted":[],"title_canon_sha256":"fa46b7d7329cdb338b5afff53fe03a268e5ba3285cd78a473d110fa228e15b85","abstract_canon_sha256":"f448af31b3a38fec30a1f6121c05cdf86f4b4c4efc57000d14db59fc91e22386"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:10:49.787201Z","signature_b64":"hKwAS6SnTmY2V3NwI1dJAl2GLhEFXZ8/SfRmrKgIEHjDQVjrN5/ncsd0/1Tb4A7K2vVwi8X5JMteCD8YyyIaCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"74787cb6147d1d5690b5dadc50a906147364e8930fa2ac69ef00385e23873a8b","last_reissued_at":"2026-07-05T05:10:49.786785Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:10:49.786785Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"M$^3$ViT: Mixture-of-Experts Vision Transformer for Efficient Multi-task Learning with Model-Accelerator Co-design","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cong Hao, Hanxue Liang, Kai Zou, Rishov Sarkar, Tianlong Chen, Yu Cheng, Zhangyang Wang, Zhiwen Fan, Ziyu Jiang","submitted_at":"2022-10-26T15:40:24Z","abstract_excerpt":"Multi-task learning (MTL) encapsulates multiple learned tasks in a single model and often lets those tasks learn better jointly. However, when deploying MTL onto those real-world systems that are often resource-constrained or latency-sensitive, two prominent challenges arise: (i) during training, simultaneously optimizing all tasks is often difficult due to gradient conflicts across tasks; (ii) at inference, current MTL regimes have to activate nearly the entire model even to just execute a single task. Yet most real systems demand only one or two tasks at each moment, and switch between tasks"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.14793","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/2210.14793/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2210.14793","created_at":"2026-07-05T05:10:49.786841+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.14793v1","created_at":"2026-07-05T05:10:49.786841+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.14793","created_at":"2026-07-05T05:10:49.786841+00:00"},{"alias_kind":"pith_short_12","alias_value":"OR4HZNQUPUOV","created_at":"2026-07-05T05:10:49.786841+00:00"},{"alias_kind":"pith_short_16","alias_value":"OR4HZNQUPUOVNEFV","created_at":"2026-07-05T05:10:49.786841+00:00"},{"alias_kind":"pith_short_8","alias_value":"OR4HZNQU","created_at":"2026-07-05T05:10:49.786841+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR","json":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR.json","graph_json":"https://pith.science/api/pith-number/OR4HZNQUPUOVNEFV3LOFBKIGCR/graph.json","events_json":"https://pith.science/api/pith-number/OR4HZNQUPUOVNEFV3LOFBKIGCR/events.json","paper":"https://pith.science/paper/OR4HZNQU"},"agent_actions":{"view_html":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR","download_json":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR.json","view_paper":"https://pith.science/paper/OR4HZNQU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.14793&json=true","fetch_graph":"https://pith.science/api/pith-number/OR4HZNQUPUOVNEFV3LOFBKIGCR/graph.json","fetch_events":"https://pith.science/api/pith-number/OR4HZNQUPUOVNEFV3LOFBKIGCR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR/action/storage_attestation","attest_author":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR/action/author_attestation","sign_citation":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR/action/citation_signature","submit_replication":"https://pith.science/pith/OR4HZNQUPUOVNEFV3LOFBKIGCR/action/replication_record"}},"created_at":"2026-07-05T05:10:49.786841+00:00","updated_at":"2026-07-05T05:10:49.786841+00:00"}