{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:333KXOOJSJVZ5CKBPYFCF72FFH","short_pith_number":"pith:333KXOOJ","schema_version":"1.0","canonical_sha256":"def6abb9c9926b9e89417e0a22ff4529e39c961ad8d20be2de7492fc8703ad1a","source":{"kind":"arxiv","id":"2406.09679","version":1},"attestation_state":"computed","paper":{"title":"Exploring Training on Heterogeneous Data with Mixture of Low-rank Adapters","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Haolin Li, Jiangchao Yao, Siyuan Du, Yanfeng Wang, Ya Zhang, Yuhang Zhou, Zihua Zhao","submitted_at":"2024-06-14T03:04:05Z","abstract_excerpt":"Training a unified model to take multiple targets into account is a trend towards artificial general intelligence. However, how to efficiently mitigate the training conflicts among heterogeneous data collected from different domains or tasks remains under-explored. In this study, we explore to leverage Mixture of Low-rank Adapters (MoLA) to mitigate conflicts in heterogeneous data training, which requires to jointly train the multiple low-rank adapters and their shared backbone. Specifically, we introduce two variants of MoLA, namely, MoLA-Grad and MoLA-Router, to respectively handle the targe"},"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":"2406.09679","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-06-14T03:04:05Z","cross_cats_sorted":[],"title_canon_sha256":"5c09bdc425a60ee98562779ef9dd643461b1a019318f7060054b0e5a3bc70b3f","abstract_canon_sha256":"34b58b9e503eb6072aff054c05827ea3f7c5f58be95709005fbb409b2bedd576"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:31:57.982601Z","signature_b64":"yaBeS4pC5g6Z7eVRXT5wCOHBYJt9uzk6/YMGqsRmr5KaVEakLll99BdtIt1rwifasOA0juLIVI4Qmi65bXk8CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"def6abb9c9926b9e89417e0a22ff4529e39c961ad8d20be2de7492fc8703ad1a","last_reissued_at":"2026-07-05T08:31:57.982119Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:31:57.982119Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Exploring Training on Heterogeneous Data with Mixture of Low-rank Adapters","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Haolin Li, Jiangchao Yao, Siyuan Du, Yanfeng Wang, Ya Zhang, Yuhang Zhou, Zihua Zhao","submitted_at":"2024-06-14T03:04:05Z","abstract_excerpt":"Training a unified model to take multiple targets into account is a trend towards artificial general intelligence. However, how to efficiently mitigate the training conflicts among heterogeneous data collected from different domains or tasks remains under-explored. In this study, we explore to leverage Mixture of Low-rank Adapters (MoLA) to mitigate conflicts in heterogeneous data training, which requires to jointly train the multiple low-rank adapters and their shared backbone. Specifically, we introduce two variants of MoLA, namely, MoLA-Grad and MoLA-Router, to respectively handle the targe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.09679","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/2406.09679/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":"2406.09679","created_at":"2026-07-05T08:31:57.982183+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.09679v1","created_at":"2026-07-05T08:31:57.982183+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.09679","created_at":"2026-07-05T08:31:57.982183+00:00"},{"alias_kind":"pith_short_12","alias_value":"333KXOOJSJVZ","created_at":"2026-07-05T08:31:57.982183+00:00"},{"alias_kind":"pith_short_16","alias_value":"333KXOOJSJVZ5CKB","created_at":"2026-07-05T08:31:57.982183+00:00"},{"alias_kind":"pith_short_8","alias_value":"333KXOOJ","created_at":"2026-07-05T08:31:57.982183+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26287","citing_title":"GeMoE: Gating Entropy is All You Need for Uncertainty-aware Adaptive Routing in MoE-based Large Vision-Language Models","ref_index":65,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25131","citing_title":"Towards Unified Multi-task EEG Analysis with Low-Rank Adaptation","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH","json":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH.json","graph_json":"https://pith.science/api/pith-number/333KXOOJSJVZ5CKBPYFCF72FFH/graph.json","events_json":"https://pith.science/api/pith-number/333KXOOJSJVZ5CKBPYFCF72FFH/events.json","paper":"https://pith.science/paper/333KXOOJ"},"agent_actions":{"view_html":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH","download_json":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH.json","view_paper":"https://pith.science/paper/333KXOOJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.09679&json=true","fetch_graph":"https://pith.science/api/pith-number/333KXOOJSJVZ5CKBPYFCF72FFH/graph.json","fetch_events":"https://pith.science/api/pith-number/333KXOOJSJVZ5CKBPYFCF72FFH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH/action/storage_attestation","attest_author":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH/action/author_attestation","sign_citation":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH/action/citation_signature","submit_replication":"https://pith.science/pith/333KXOOJSJVZ5CKBPYFCF72FFH/action/replication_record"}},"created_at":"2026-07-05T08:31:57.982183+00:00","updated_at":"2026-07-05T08:31:57.982183+00:00"}