{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QCSSJS5F7PPMK65YG2TUQGQ4CK","short_pith_number":"pith:QCSSJS5F","schema_version":"1.0","canonical_sha256":"80a524cba5fbdec57bb836a7481a1c12811469216095956baf0dd22dc92fe771","source":{"kind":"arxiv","id":"2503.11197","version":4},"attestation_state":"computed","paper":{"title":"Reinforcement Learning Outperforms Supervised Fine-Tuning: A Case Study on Audio Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Gang Li, Heinrich Dinkel, Jian Luan, Jizhong Liu, Junbo Zhang, Yadong Niu","submitted_at":"2025-03-14T08:43:53Z","abstract_excerpt":"Recently, reinforcement learning (RL) has been shown to greatly enhance the reasoning capabilities of large language models (LLMs), and RL-based approaches have been progressively applied to visual multimodal tasks. However, the audio modality has largely been overlooked in these developments. Thus, we conduct a series of RL explorations in audio understanding and reasoning, specifically focusing on the audio question answering (AQA) task. We leverage the group relative policy optimization (GRPO) algorithm to Qwen2-Audio-7B-Instruct, and our experiments demonstrated state-of-the-art performanc"},"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":"2503.11197","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SD","submitted_at":"2025-03-14T08:43:53Z","cross_cats_sorted":["cs.AI","cs.CL","eess.AS"],"title_canon_sha256":"a0d80d85c8a7488cbfccf28b280f9533197906f65c6867a009f3f7c413718f3d","abstract_canon_sha256":"b2a397ac2c1e27870196bde1b1d1f376926cb3489de83246d479ccaa9a5a3faf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:02:42.644013Z","signature_b64":"/ef8Nvt2mmhGZhEtbQx2Kn3jBDZuQ/l1gaOMfvo4seNaH+NSnRuJ9WxNLRAfDoj/eOmmmA+4/YD3+YdAhkMsBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"80a524cba5fbdec57bb836a7481a1c12811469216095956baf0dd22dc92fe771","last_reissued_at":"2026-07-05T11:02:42.643458Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:02:42.643458Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reinforcement Learning Outperforms Supervised Fine-Tuning: A Case Study on Audio Question Answering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","eess.AS"],"primary_cat":"cs.SD","authors_text":"Gang Li, Heinrich Dinkel, Jian Luan, Jizhong Liu, Junbo Zhang, Yadong Niu","submitted_at":"2025-03-14T08:43:53Z","abstract_excerpt":"Recently, reinforcement learning (RL) has been shown to greatly enhance the reasoning capabilities of large language models (LLMs), and RL-based approaches have been progressively applied to visual multimodal tasks. However, the audio modality has largely been overlooked in these developments. Thus, we conduct a series of RL explorations in audio understanding and reasoning, specifically focusing on the audio question answering (AQA) task. We leverage the group relative policy optimization (GRPO) algorithm to Qwen2-Audio-7B-Instruct, and our experiments demonstrated state-of-the-art performanc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.11197","kind":"arxiv","version":4},"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/2503.11197/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":"2503.11197","created_at":"2026-07-05T11:02:42.643526+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.11197v4","created_at":"2026-07-05T11:02:42.643526+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.11197","created_at":"2026-07-05T11:02:42.643526+00:00"},{"alias_kind":"pith_short_12","alias_value":"QCSSJS5F7PPM","created_at":"2026-07-05T11:02:42.643526+00:00"},{"alias_kind":"pith_short_16","alias_value":"QCSSJS5F7PPMK65Y","created_at":"2026-07-05T11:02:42.643526+00:00"},{"alias_kind":"pith_short_8","alias_value":"QCSSJS5F","created_at":"2026-07-05T11:02:42.643526+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":15,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07264","citing_title":"VISA: A Visual Information Strengthened Audio-Reasoning System for the Interspeech 2026 ARC Agent Track","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02615","citing_title":"FSA-GRPO: Teaching Auditory LLMs to Use Few-shot Demonstrations","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27190","citing_title":"Learning When to Think While Listening in Large Audio-Language Models","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00507","citing_title":"LaSR: Context-Aware Speech Recognition via Latent Reasoning","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2511.20785","citing_title":"LongVT: Incentivizing \"Thinking with Long Videos\" via Native Tool Calling","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21008","citing_title":"A Survey of Audio Reasoning in Multimodal Foundation Models","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15984","citing_title":"Beyond Content: A Comprehensive Speech Toxicity Dataset and Detection Framework Incorporating Paralinguistic Cues","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2509.02547","citing_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","ref_index":262,"is_internal_anchor":false},{"citing_arxiv_id":"2510.09592","citing_title":"Mind-Paced Speaking: A Dual-Brain Approach to Real-Time Reasoning in Spoken Language Models","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2507.08128","citing_title":"Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models","ref_index":73,"is_internal_anchor":false},{"citing_arxiv_id":"2502.17419","citing_title":"From System 1 to System 2: A Survey of Reasoning Large Language Models","ref_index":292,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09996","citing_title":"Omni-Persona: Systematic Benchmarking and Improving Omnimodal Personalization","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18187","citing_title":"Audio-DeepThinker: Progressive Reasoning-Aware Reinforcement Learning for High-Quality Chain-of-Thought Emergence in Audio Language Models","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12145","citing_title":"Why Your Tokenizer Fails in Information Fusion: A Timing-Aware Pre-Quantization Fusion for Video-Enhanced Audio Tokenization","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15849","citing_title":"TinyMU: A Compact Audio-Language Model for Music Understanding","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK","json":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK.json","graph_json":"https://pith.science/api/pith-number/QCSSJS5F7PPMK65YG2TUQGQ4CK/graph.json","events_json":"https://pith.science/api/pith-number/QCSSJS5F7PPMK65YG2TUQGQ4CK/events.json","paper":"https://pith.science/paper/QCSSJS5F"},"agent_actions":{"view_html":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK","download_json":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK.json","view_paper":"https://pith.science/paper/QCSSJS5F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.11197&json=true","fetch_graph":"https://pith.science/api/pith-number/QCSSJS5F7PPMK65YG2TUQGQ4CK/graph.json","fetch_events":"https://pith.science/api/pith-number/QCSSJS5F7PPMK65YG2TUQGQ4CK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK/action/storage_attestation","attest_author":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK/action/author_attestation","sign_citation":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK/action/citation_signature","submit_replication":"https://pith.science/pith/QCSSJS5F7PPMK65YG2TUQGQ4CK/action/replication_record"}},"created_at":"2026-07-05T11:02:42.643526+00:00","updated_at":"2026-07-05T11:02:42.643526+00:00"}