{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TRDYRXUQMLQYLETZSQKOGYTX24","short_pith_number":"pith:TRDYRXUQ","schema_version":"1.0","canonical_sha256":"9c4788de9062e18592799414e36277d70bf44e13c437c500b145ca593f2a29ad","source":{"kind":"arxiv","id":"2505.04623","version":1},"attestation_state":"computed","paper":{"title":"EchoInk-R1: Exploring Audio-Visual Reasoning in Multimodal LLMs via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.CV","authors_text":"Chi-Wing Fu, Jifeng Dai, Pheng-Ann Heng, Wenhai Wang, Xiaowei Hu, Zhenghao Xing","submitted_at":"2025-05-07T17:59:49Z","abstract_excerpt":"Multimodal large language models (MLLMs) have advanced perception across text, vision, and audio, yet they often struggle with structured cross-modal reasoning, particularly when integrating audio and visual signals. We introduce EchoInk-R1, a reinforcement learning framework that enhances such reasoning in MLLMs. Built upon the Qwen2.5-Omni-7B foundation and optimized with Group Relative Policy Optimization (GRPO), EchoInk-R1 tackles multiple-choice question answering over synchronized audio-image pairs. To enable this, we curate AVQA-R1-6K, a dataset pairing such audio-image inputs with mult"},"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":"2505.04623","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-05-07T17:59:49Z","cross_cats_sorted":["eess.AS"],"title_canon_sha256":"f56607595f4f18c047898dfe176d43d76c11c8f6915652fdeea4f0c032b61436","abstract_canon_sha256":"ff8f037bcb4a441fc2e2d451c269e1b4214b53984df7c55952df89c236df61f0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:48.584682Z","signature_b64":"eOB87bJfYLeLXNfJPGVqSfgoTizW3WdijkRXbQjTiOnIpQio1tgkkU1NBF5vSrIzd21++WHar5dGJP9SoczlBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9c4788de9062e18592799414e36277d70bf44e13c437c500b145ca593f2a29ad","last_reissued_at":"2026-07-05T11:23:48.584186Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:48.584186Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EchoInk-R1: Exploring Audio-Visual Reasoning in Multimodal LLMs via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["eess.AS"],"primary_cat":"cs.CV","authors_text":"Chi-Wing Fu, Jifeng Dai, Pheng-Ann Heng, Wenhai Wang, Xiaowei Hu, Zhenghao Xing","submitted_at":"2025-05-07T17:59:49Z","abstract_excerpt":"Multimodal large language models (MLLMs) have advanced perception across text, vision, and audio, yet they often struggle with structured cross-modal reasoning, particularly when integrating audio and visual signals. We introduce EchoInk-R1, a reinforcement learning framework that enhances such reasoning in MLLMs. Built upon the Qwen2.5-Omni-7B foundation and optimized with Group Relative Policy Optimization (GRPO), EchoInk-R1 tackles multiple-choice question answering over synchronized audio-image pairs. To enable this, we curate AVQA-R1-6K, a dataset pairing such audio-image inputs with mult"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.04623","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/2505.04623/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":"2505.04623","created_at":"2026-07-05T11:23:48.584245+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.04623v1","created_at":"2026-07-05T11:23:48.584245+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.04623","created_at":"2026-07-05T11:23:48.584245+00:00"},{"alias_kind":"pith_short_12","alias_value":"TRDYRXUQMLQY","created_at":"2026-07-05T11:23:48.584245+00:00"},{"alias_kind":"pith_short_16","alias_value":"TRDYRXUQMLQYLETZ","created_at":"2026-07-05T11:23:48.584245+00:00"},{"alias_kind":"pith_short_8","alias_value":"TRDYRXUQ","created_at":"2026-07-05T11:23:48.584245+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.22012","citing_title":"LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2509.02547","citing_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","ref_index":264,"is_internal_anchor":false},{"citing_arxiv_id":"2510.15148","citing_title":"XModBench: Benchmarking Cross-Modal Capabilities and Consistency in Omni-Language Models","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11244","citing_title":"Script-a-Video: Deep Structured Audio-visual Captions via Factorized Streams and Relational Grounding","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11554","citing_title":"Relax: An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05522","citing_title":"Cross-Modal Coreference Alignment: Enabling Reliable Information Transfer in Omni-LLMs","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14520","citing_title":"Chain of Modality: From Static Fusion to Dynamic Orchestration in Omni-MLLMs","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16617","citing_title":"AVRT: Audio-Visual Reasoning Transfer through Single-Modality Teachers","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24","json":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24.json","graph_json":"https://pith.science/api/pith-number/TRDYRXUQMLQYLETZSQKOGYTX24/graph.json","events_json":"https://pith.science/api/pith-number/TRDYRXUQMLQYLETZSQKOGYTX24/events.json","paper":"https://pith.science/paper/TRDYRXUQ"},"agent_actions":{"view_html":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24","download_json":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24.json","view_paper":"https://pith.science/paper/TRDYRXUQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.04623&json=true","fetch_graph":"https://pith.science/api/pith-number/TRDYRXUQMLQYLETZSQKOGYTX24/graph.json","fetch_events":"https://pith.science/api/pith-number/TRDYRXUQMLQYLETZSQKOGYTX24/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24/action/storage_attestation","attest_author":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24/action/author_attestation","sign_citation":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24/action/citation_signature","submit_replication":"https://pith.science/pith/TRDYRXUQMLQYLETZSQKOGYTX24/action/replication_record"}},"created_at":"2026-07-05T11:23:48.584245+00:00","updated_at":"2026-07-05T11:23:48.584245+00:00"}