{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SAQS6LFQIOOEFXLSPU4VCXIWHG","short_pith_number":"pith:SAQS6LFQ","schema_version":"1.0","canonical_sha256":"90212f2cb0439c42dd727d39515d163988862a6c8eaaca417bd6c59738bda01b","source":{"kind":"arxiv","id":"2603.06331","version":2},"attestation_state":"computed","paper":{"title":"WorldCache: Accelerating World Models for Free via Heterogeneous Token Caching","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chuanguang Yang, Dingrui Wang, Guoxin Fan, Haotong Qin, Libo Huang, Longlong Liao, Michele Magno, Mingqiang Wu, Weilun Feng, Xiangqi Li, Yongjun Xu, Yuqi Li, Zhulin An","submitted_at":"2026-03-06T14:43:47Z","abstract_excerpt":"Diffusion-based world models have shown strong potential for unified world simulation, but the iterative denoising remains too costly for interactive use and long-horizon rollouts. While feature caching can accelerate inference without training, we find that policies designed for single-modal diffusion transfer poorly to world models due to two world-model-specific obstacles: \\emph{token heterogeneity} from multi-modal coupling and spatial variation, and \\emph{non-uniform temporal dynamics} where a small set of hard tokens drives error growth, making uniform skipping either unstable or overly "},"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":"2603.06331","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-03-06T14:43:47Z","cross_cats_sorted":[],"title_canon_sha256":"b232867695b0ac036b16c3e6ec22218b75f3c55f576c9e06ecb27f2e4ba79165","abstract_canon_sha256":"6909369b0969cddf4411306e00d5327e8c04dc71c639f045bc6a2ca4e77a5f7d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:52.151448Z","signature_b64":"hcQsu9Oz5eQ5VjWf2C+doNXBgCGUoBMeUa5dEa0Q+RFiAwAhbj99AxFF0JttHbvd7za3qgmhJDMEztMI+ye7Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"90212f2cb0439c42dd727d39515d163988862a6c8eaaca417bd6c59738bda01b","last_reissued_at":"2026-06-02T02:04:52.150975Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:52.150975Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"WorldCache: Accelerating World Models for Free via Heterogeneous Token Caching","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chuanguang Yang, Dingrui Wang, Guoxin Fan, Haotong Qin, Libo Huang, Longlong Liao, Michele Magno, Mingqiang Wu, Weilun Feng, Xiangqi Li, Yongjun Xu, Yuqi Li, Zhulin An","submitted_at":"2026-03-06T14:43:47Z","abstract_excerpt":"Diffusion-based world models have shown strong potential for unified world simulation, but the iterative denoising remains too costly for interactive use and long-horizon rollouts. While feature caching can accelerate inference without training, we find that policies designed for single-modal diffusion transfer poorly to world models due to two world-model-specific obstacles: \\emph{token heterogeneity} from multi-modal coupling and spatial variation, and \\emph{non-uniform temporal dynamics} where a small set of hard tokens drives error growth, making uniform skipping either unstable or overly "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.06331","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/2603.06331/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":"2603.06331","created_at":"2026-06-02T02:04:52.151039+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.06331v2","created_at":"2026-06-02T02:04:52.151039+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.06331","created_at":"2026-06-02T02:04:52.151039+00:00"},{"alias_kind":"pith_short_12","alias_value":"SAQS6LFQIOOE","created_at":"2026-06-02T02:04:52.151039+00:00"},{"alias_kind":"pith_short_16","alias_value":"SAQS6LFQIOOEFXLS","created_at":"2026-06-02T02:04:52.151039+00:00"},{"alias_kind":"pith_short_8","alias_value":"SAQS6LFQ","created_at":"2026-06-02T02:04:52.151039+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2603.28489","citing_title":"Video Generation Models as World Models: Efficient Paradigms, Architectures and Algorithms","ref_index":149,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG","json":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG.json","graph_json":"https://pith.science/api/pith-number/SAQS6LFQIOOEFXLSPU4VCXIWHG/graph.json","events_json":"https://pith.science/api/pith-number/SAQS6LFQIOOEFXLSPU4VCXIWHG/events.json","paper":"https://pith.science/paper/SAQS6LFQ"},"agent_actions":{"view_html":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG","download_json":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG.json","view_paper":"https://pith.science/paper/SAQS6LFQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.06331&json=true","fetch_graph":"https://pith.science/api/pith-number/SAQS6LFQIOOEFXLSPU4VCXIWHG/graph.json","fetch_events":"https://pith.science/api/pith-number/SAQS6LFQIOOEFXLSPU4VCXIWHG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG/action/storage_attestation","attest_author":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG/action/author_attestation","sign_citation":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG/action/citation_signature","submit_replication":"https://pith.science/pith/SAQS6LFQIOOEFXLSPU4VCXIWHG/action/replication_record"}},"created_at":"2026-06-02T02:04:52.151039+00:00","updated_at":"2026-06-02T02:04:52.151039+00:00"}