{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7EQON5KGFPBI55VE7DYQ7L64Y4","short_pith_number":"pith:7EQON5KG","schema_version":"1.0","canonical_sha256":"f920e6f5462bc28ef6a4f8f10fafdcc712d25a7d16c6e9f8ab13e515b5c87ec0","source":{"kind":"arxiv","id":"2505.03798","version":1},"attestation_state":"computed","paper":{"title":"Position: Foundation Models Need Digital Twin Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hao Ding, Lalithkumar Seenivasan, Mathias Unberath, Tianmin Shu, Yiqing Shen","submitted_at":"2025-05-01T22:17:41Z","abstract_excerpt":"Current foundation models (FMs) rely on token representations that directly fragment continuous real-world multimodal data into discrete tokens. They limit FMs to learning real-world knowledge and relationships purely through statistical correlation rather than leveraging explicit domain knowledge. Consequently, current FMs struggle with maintaining semantic coherence across modalities, capturing fine-grained spatial-temporal dynamics, and performing causal reasoning. These limitations cannot be overcome by simply scaling up model size or expanding datasets. This position paper argues that the"},"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":"2505.03798","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-01T22:17:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"84ca7a30e72c903d8f0809a723dd48b603eae9f8792f63de0f2dcc89e9eab6d1","abstract_canon_sha256":"72150bcc67f68570eea14b8a6aa6bb80804e47d0ba2be913030d745a54cd0b1f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:59:31.692887Z","signature_b64":"no+a9FM1CTVlSBKQ+Eq+vBVdgHAheC+k5g8qf2/6jpFLP9Ww/E+RBEhuzQ0yow+YuzFiNEi0/ldMbcZ+ewIfDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f920e6f5462bc28ef6a4f8f10fafdcc712d25a7d16c6e9f8ab13e515b5c87ec0","last_reissued_at":"2026-07-05T10:59:31.692370Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:59:31.692370Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Position: Foundation Models Need Digital Twin Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hao Ding, Lalithkumar Seenivasan, Mathias Unberath, Tianmin Shu, Yiqing Shen","submitted_at":"2025-05-01T22:17:41Z","abstract_excerpt":"Current foundation models (FMs) rely on token representations that directly fragment continuous real-world multimodal data into discrete tokens. They limit FMs to learning real-world knowledge and relationships purely through statistical correlation rather than leveraging explicit domain knowledge. Consequently, current FMs struggle with maintaining semantic coherence across modalities, capturing fine-grained spatial-temporal dynamics, and performing causal reasoning. These limitations cannot be overcome by simply scaling up model size or expanding datasets. This position paper argues that the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.03798","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/2505.03798/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":"2505.03798","created_at":"2026-07-05T10:59:31.692430+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.03798v1","created_at":"2026-07-05T10:59:31.692430+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.03798","created_at":"2026-07-05T10:59:31.692430+00:00"},{"alias_kind":"pith_short_12","alias_value":"7EQON5KGFPBI","created_at":"2026-07-05T10:59:31.692430+00:00"},{"alias_kind":"pith_short_16","alias_value":"7EQON5KGFPBI55VE","created_at":"2026-07-05T10:59:31.692430+00:00"},{"alias_kind":"pith_short_8","alias_value":"7EQON5KG","created_at":"2026-07-05T10:59:31.692430+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17279","citing_title":"Training LLMs with Reinforcement Learning over Digital Twin Representations for Reasoning-Intensive Surgical VideoQA","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4","json":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4.json","graph_json":"https://pith.science/api/pith-number/7EQON5KGFPBI55VE7DYQ7L64Y4/graph.json","events_json":"https://pith.science/api/pith-number/7EQON5KGFPBI55VE7DYQ7L64Y4/events.json","paper":"https://pith.science/paper/7EQON5KG"},"agent_actions":{"view_html":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4","download_json":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4.json","view_paper":"https://pith.science/paper/7EQON5KG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.03798&json=true","fetch_graph":"https://pith.science/api/pith-number/7EQON5KGFPBI55VE7DYQ7L64Y4/graph.json","fetch_events":"https://pith.science/api/pith-number/7EQON5KGFPBI55VE7DYQ7L64Y4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4/action/storage_attestation","attest_author":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4/action/author_attestation","sign_citation":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4/action/citation_signature","submit_replication":"https://pith.science/pith/7EQON5KGFPBI55VE7DYQ7L64Y4/action/replication_record"}},"created_at":"2026-07-05T10:59:31.692430+00:00","updated_at":"2026-07-05T10:59:31.692430+00:00"}