{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KG6I3YXWVCHYHOCPILYWAE3GZQ","short_pith_number":"pith:KG6I3YXW","schema_version":"1.0","canonical_sha256":"51bc8de2f6a88f83b84f42f1601366cc153b967ec03f38c3f01f42b573ea4cb8","source":{"kind":"arxiv","id":"2606.15920","version":2},"attestation_state":"computed","paper":{"title":"OmniOPSD: Rationale-Privileged On-Policy Self-Distillation for Affective Computing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Boxue Yang, Fei Ma, Jingyi Chen, Laizhong Cui, Qi Tian, Shuimu Chen, Xiaojiang Peng, Yuanshen Guan, Zebang Cheng, Zheng Lian","submitted_at":"2026-06-14T16:51:22Z","abstract_excerpt":"Reinforcement learning for multimodal large language models (MLLMs) is often hindered by severe reward sparsity in complex reasoning tasks. This challenge is particularly pronounced in human-centered scenarios involving states, emotions, intentions, and behaviors, where heterogeneous multimodal signals and subjective human factors make high-quality chain-of-thought (CoT) annotations expensive and difficult to obtain. Although many multimodal datasets provide expert-annotated ground-truth labels, directly using these labels for supervised fine-tuning may encourage shortcut learning in multimoda"},"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":"2606.15920","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-14T16:51:22Z","cross_cats_sorted":[],"title_canon_sha256":"242b2447b05575fcaa76971519de7693a6401ef9073e5c3073cd843c02108c37","abstract_canon_sha256":"d425f646b9fbed94d7c01ec145cfa18704f7a0694990075305dfbd5e2fdc8f1b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-10T01:19:45.005063Z","signature_b64":"Rh/Q8HQmHXtEEhYACfaE4zm0AIGMY8c+J/w3irGntMecgoqvbmq80UrklpepmJ6VH75FHl0DzSfnUjIZQQ2OBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"51bc8de2f6a88f83b84f42f1601366cc153b967ec03f38c3f01f42b573ea4cb8","last_reissued_at":"2026-07-10T01:19:45.004682Z","signature_status":"signed_v1","first_computed_at":"2026-07-10T01:19:45.004682Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OmniOPSD: Rationale-Privileged On-Policy Self-Distillation for Affective Computing","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Boxue Yang, Fei Ma, Jingyi Chen, Laizhong Cui, Qi Tian, Shuimu Chen, Xiaojiang Peng, Yuanshen Guan, Zebang Cheng, Zheng Lian","submitted_at":"2026-06-14T16:51:22Z","abstract_excerpt":"Reinforcement learning for multimodal large language models (MLLMs) is often hindered by severe reward sparsity in complex reasoning tasks. This challenge is particularly pronounced in human-centered scenarios involving states, emotions, intentions, and behaviors, where heterogeneous multimodal signals and subjective human factors make high-quality chain-of-thought (CoT) annotations expensive and difficult to obtain. Although many multimodal datasets provide expert-annotated ground-truth labels, directly using these labels for supervised fine-tuning may encourage shortcut learning in multimoda"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.15920","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/2606.15920/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":"2606.15920","created_at":"2026-07-10T01:19:45.004740+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.15920v2","created_at":"2026-07-10T01:19:45.004740+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.15920","created_at":"2026-07-10T01:19:45.004740+00:00"},{"alias_kind":"pith_short_12","alias_value":"KG6I3YXWVCHY","created_at":"2026-07-10T01:19:45.004740+00:00"},{"alias_kind":"pith_short_16","alias_value":"KG6I3YXWVCHYHOCP","created_at":"2026-07-10T01:19:45.004740+00:00"},{"alias_kind":"pith_short_8","alias_value":"KG6I3YXW","created_at":"2026-07-10T01:19:45.004740+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.16322","citing_title":"GMoT: Gated Motion-Aware Tokenization for Fine-Grained Micro-Gesture Video Reasoning with Multimodal LLMs","ref_index":14,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ","json":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ.json","graph_json":"https://pith.science/api/pith-number/KG6I3YXWVCHYHOCPILYWAE3GZQ/graph.json","events_json":"https://pith.science/api/pith-number/KG6I3YXWVCHYHOCPILYWAE3GZQ/events.json","paper":"https://pith.science/paper/KG6I3YXW"},"agent_actions":{"view_html":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ","download_json":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ.json","view_paper":"https://pith.science/paper/KG6I3YXW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.15920&json=true","fetch_graph":"https://pith.science/api/pith-number/KG6I3YXWVCHYHOCPILYWAE3GZQ/graph.json","fetch_events":"https://pith.science/api/pith-number/KG6I3YXWVCHYHOCPILYWAE3GZQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ/action/storage_attestation","attest_author":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ/action/author_attestation","sign_citation":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ/action/citation_signature","submit_replication":"https://pith.science/pith/KG6I3YXWVCHYHOCPILYWAE3GZQ/action/replication_record"}},"created_at":"2026-07-10T01:19:45.004740+00:00","updated_at":"2026-07-10T01:19:45.004740+00:00"}