{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:N3BUC6TD32ULFGVUMKN2AII5Q4","short_pith_number":"pith:N3BUC6TD","schema_version":"1.0","canonical_sha256":"6ec3417a63dea8b29ab4629ba0211d870f9f5a104c183206257b3a61d5ed6f6e","source":{"kind":"arxiv","id":"2506.01371","version":1},"attestation_state":"computed","paper":{"title":"SVQA-R1: Reinforcing Spatial Reasoning in MLLMs via View-Consistent Reward Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Haibin Ling, Peiyao Wang","submitted_at":"2025-06-02T06:58:43Z","abstract_excerpt":"Spatial reasoning remains a critical yet underdeveloped capability in existing vision-language models (VLMs), especially for Spatial Visual Question Answering (Spatial VQA) tasks that require understanding relative positions, distances, and object configurations. Inspired by the R1 paradigm introduced in DeepSeek-R1, which enhances reasoning in language models through rule-based reinforcement learning (RL), we propose SVQA-R1, the first framework to extend R1-style training to spatial VQA. In particular, we introduce Spatial-GRPO, a novel group-wise RL strategy that constructs view-consistent "},"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":"2506.01371","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-06-02T06:58:43Z","cross_cats_sorted":[],"title_canon_sha256":"23e7b1089ea125d3a08cfe1d76da5d41b535b7d88136e341740a0e25b1f049d1","abstract_canon_sha256":"a638686f3aab330100ecf1d3aa19896ece60684085977b08cef2683f9649c7db"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:12.335029Z","signature_b64":"9UpfacL3a402Hsd5lb0YDwnqFaqbgMIWOU3YdeNSTY8b9n9KQZ3osZlxQ2DvIt/JvIkzIcdWx2IHEu42F2glCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ec3417a63dea8b29ab4629ba0211d870f9f5a104c183206257b3a61d5ed6f6e","last_reissued_at":"2026-07-05T11:14:12.334510Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:12.334510Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SVQA-R1: Reinforcing Spatial Reasoning in MLLMs via View-Consistent Reward Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Haibin Ling, Peiyao Wang","submitted_at":"2025-06-02T06:58:43Z","abstract_excerpt":"Spatial reasoning remains a critical yet underdeveloped capability in existing vision-language models (VLMs), especially for Spatial Visual Question Answering (Spatial VQA) tasks that require understanding relative positions, distances, and object configurations. Inspired by the R1 paradigm introduced in DeepSeek-R1, which enhances reasoning in language models through rule-based reinforcement learning (RL), we propose SVQA-R1, the first framework to extend R1-style training to spatial VQA. In particular, we introduce Spatial-GRPO, a novel group-wise RL strategy that constructs view-consistent "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.01371","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/2506.01371/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":"2506.01371","created_at":"2026-07-05T11:14:12.334571+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.01371v1","created_at":"2026-07-05T11:14:12.334571+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.01371","created_at":"2026-07-05T11:14:12.334571+00:00"},{"alias_kind":"pith_short_12","alias_value":"N3BUC6TD32UL","created_at":"2026-07-05T11:14:12.334571+00:00"},{"alias_kind":"pith_short_16","alias_value":"N3BUC6TD32ULFGVU","created_at":"2026-07-05T11:14:12.334571+00:00"},{"alias_kind":"pith_short_8","alias_value":"N3BUC6TD","created_at":"2026-07-05T11:14:12.334571+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2505.17012","citing_title":"SpatialScore: Towards Comprehensive Evaluation for Spatial Intelligence","ref_index":84,"is_internal_anchor":false},{"citing_arxiv_id":"2509.02547","citing_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","ref_index":236,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4","json":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4.json","graph_json":"https://pith.science/api/pith-number/N3BUC6TD32ULFGVUMKN2AII5Q4/graph.json","events_json":"https://pith.science/api/pith-number/N3BUC6TD32ULFGVUMKN2AII5Q4/events.json","paper":"https://pith.science/paper/N3BUC6TD"},"agent_actions":{"view_html":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4","download_json":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4.json","view_paper":"https://pith.science/paper/N3BUC6TD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.01371&json=true","fetch_graph":"https://pith.science/api/pith-number/N3BUC6TD32ULFGVUMKN2AII5Q4/graph.json","fetch_events":"https://pith.science/api/pith-number/N3BUC6TD32ULFGVUMKN2AII5Q4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4/action/storage_attestation","attest_author":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4/action/author_attestation","sign_citation":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4/action/citation_signature","submit_replication":"https://pith.science/pith/N3BUC6TD32ULFGVUMKN2AII5Q4/action/replication_record"}},"created_at":"2026-07-05T11:14:12.334571+00:00","updated_at":"2026-07-05T11:14:12.334571+00:00"}