{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EORAJE6JRTD7TJBJTZBWDVGG3W","short_pith_number":"pith:EORAJE6J","schema_version":"1.0","canonical_sha256":"23a20493c98cc7f9a4299e4361d4c6ddb4ffd277dac81f50879b9bf620ff0860","source":{"kind":"arxiv","id":"2507.02591","version":3},"attestation_state":"computed","paper":{"title":"AuroraLong: Bringing RNNs Back to Efficient Open-Ended Video Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Enxin Song, Gaoang Wang, Tian Ye, Weili Xu, Wenhao Chai, Xuexiang Wen","submitted_at":"2025-07-03T12:55:16Z","abstract_excerpt":"The challenge of long video understanding lies in its high computational complexity and prohibitive memory cost, since the memory and computation required by transformer-based LLMs scale quadratically with input sequence length. We propose AuroraLong to address this challenge by replacing the LLM component in MLLMs with a linear RNN language model that handles input sequence of arbitrary length with constant-size hidden states. To further increase throughput and efficiency, we combine visual token merge with linear RNN models by reordering the visual tokens by their sizes in ascending order. D"},"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":"2507.02591","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-07-03T12:55:16Z","cross_cats_sorted":[],"title_canon_sha256":"f92d0846a218dac390eb69314ad86d37181c53dbab9651ebc00a0409e02732e7","abstract_canon_sha256":"040fd464f9933c791afcd503504d91465cd7be27f7db6f124b77d38faeb2d4d9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:41:43.739758Z","signature_b64":"wUMgNzv3AHqV9WZxTyjXnjFPyWT4AHfsOQ04cuBuL9a6SIjha9+v9VcIv8OVNk2gm2YjbUQMW1kt569C9WWzAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"23a20493c98cc7f9a4299e4361d4c6ddb4ffd277dac81f50879b9bf620ff0860","last_reissued_at":"2026-07-05T11:41:43.739295Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:41:43.739295Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AuroraLong: Bringing RNNs Back to Efficient Open-Ended Video Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Enxin Song, Gaoang Wang, Tian Ye, Weili Xu, Wenhao Chai, Xuexiang Wen","submitted_at":"2025-07-03T12:55:16Z","abstract_excerpt":"The challenge of long video understanding lies in its high computational complexity and prohibitive memory cost, since the memory and computation required by transformer-based LLMs scale quadratically with input sequence length. We propose AuroraLong to address this challenge by replacing the LLM component in MLLMs with a linear RNN language model that handles input sequence of arbitrary length with constant-size hidden states. To further increase throughput and efficiency, we combine visual token merge with linear RNN models by reordering the visual tokens by their sizes in ascending order. D"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.02591","kind":"arxiv","version":3},"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/2507.02591/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":"2507.02591","created_at":"2026-07-05T11:41:43.739359+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.02591v3","created_at":"2026-07-05T11:41:43.739359+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.02591","created_at":"2026-07-05T11:41:43.739359+00:00"},{"alias_kind":"pith_short_12","alias_value":"EORAJE6JRTD7","created_at":"2026-07-05T11:41:43.739359+00:00"},{"alias_kind":"pith_short_16","alias_value":"EORAJE6JRTD7TJBJ","created_at":"2026-07-05T11:41:43.739359+00:00"},{"alias_kind":"pith_short_8","alias_value":"EORAJE6J","created_at":"2026-07-05T11:41:43.739359+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W","json":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W.json","graph_json":"https://pith.science/api/pith-number/EORAJE6JRTD7TJBJTZBWDVGG3W/graph.json","events_json":"https://pith.science/api/pith-number/EORAJE6JRTD7TJBJTZBWDVGG3W/events.json","paper":"https://pith.science/paper/EORAJE6J"},"agent_actions":{"view_html":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W","download_json":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W.json","view_paper":"https://pith.science/paper/EORAJE6J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.02591&json=true","fetch_graph":"https://pith.science/api/pith-number/EORAJE6JRTD7TJBJTZBWDVGG3W/graph.json","fetch_events":"https://pith.science/api/pith-number/EORAJE6JRTD7TJBJTZBWDVGG3W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W/action/storage_attestation","attest_author":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W/action/author_attestation","sign_citation":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W/action/citation_signature","submit_replication":"https://pith.science/pith/EORAJE6JRTD7TJBJTZBWDVGG3W/action/replication_record"}},"created_at":"2026-07-05T11:41:43.739359+00:00","updated_at":"2026-07-05T11:41:43.739359+00:00"}