{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KZVYIV4JKM7P3U3G54I54VDODX","short_pith_number":"pith:KZVYIV4J","schema_version":"1.0","canonical_sha256":"566b845789533efdd366ef11de546e1df299c2be1d0e4545def5cf0e72de3f8a","source":{"kind":"arxiv","id":"2503.08685","version":2},"attestation_state":"computed","paper":{"title":"\"Principal Components\" Enable A New Language of Images","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bingchen Zhao, Ismail Elezi, Jiankang Deng, Xiaojuan Qi, Xin Wen","submitted_at":"2025-03-11T17:59:41Z","abstract_excerpt":"We introduce a novel visual tokenization framework that embeds a provable PCA-like structure into the latent token space. While existing visual tokenizers primarily optimize for reconstruction fidelity, they often neglect the structural properties of the latent space--a critical factor for both interpretability and downstream tasks. Our method generates a 1D causal token sequence for images, where each successive token contributes non-overlapping information with mathematically guaranteed decreasing explained variance, analogous to principal component analysis. This structural constraint ensur"},"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":"2503.08685","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-11T17:59:41Z","cross_cats_sorted":[],"title_canon_sha256":"fe09180cf399bddf457b0d649cb91b2c6e2bde1cee9ea9ead351a9c1389abe08","abstract_canon_sha256":"6a98c44e63631a82db9b24d575516dceb5bcb860e086dabbf3a02fdacdd0dee9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:44:02.466943Z","signature_b64":"b7AgJFZeCwKyYLgWw0sZK8/LXJkuUZOQUAO1d5FV7bckg/SznckUnpXCm5PKFrAu+toFMABgyikWrc5BVyB4Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"566b845789533efdd366ef11de546e1df299c2be1d0e4545def5cf0e72de3f8a","last_reissued_at":"2026-07-05T11:44:02.466514Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:44:02.466514Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"\"Principal Components\" Enable A New Language of Images","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bingchen Zhao, Ismail Elezi, Jiankang Deng, Xiaojuan Qi, Xin Wen","submitted_at":"2025-03-11T17:59:41Z","abstract_excerpt":"We introduce a novel visual tokenization framework that embeds a provable PCA-like structure into the latent token space. While existing visual tokenizers primarily optimize for reconstruction fidelity, they often neglect the structural properties of the latent space--a critical factor for both interpretability and downstream tasks. Our method generates a 1D causal token sequence for images, where each successive token contributes non-overlapping information with mathematically guaranteed decreasing explained variance, analogous to principal component analysis. This structural constraint ensur"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.08685","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/2503.08685/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":"2503.08685","created_at":"2026-07-05T11:44:02.466572+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.08685v2","created_at":"2026-07-05T11:44:02.466572+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.08685","created_at":"2026-07-05T11:44:02.466572+00:00"},{"alias_kind":"pith_short_12","alias_value":"KZVYIV4JKM7P","created_at":"2026-07-05T11:44:02.466572+00:00"},{"alias_kind":"pith_short_16","alias_value":"KZVYIV4JKM7P3U3G","created_at":"2026-07-05T11:44:02.466572+00:00"},{"alias_kind":"pith_short_8","alias_value":"KZVYIV4J","created_at":"2026-07-05T11:44:02.466572+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.27696","citing_title":"Structure over Pixels: Learning Variable-Length Visual Programs","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18390","citing_title":"Vision Foundation Models as Generalist Tokenizers for Image Generation","ref_index":83,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12026","citing_title":"Spectral Vision Transformer for Efficient Tokenization with Limited Data","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00503","citing_title":"End-to-End Autoregressive Image Generation with 1D Semantic Tokenizer","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX","json":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX.json","graph_json":"https://pith.science/api/pith-number/KZVYIV4JKM7P3U3G54I54VDODX/graph.json","events_json":"https://pith.science/api/pith-number/KZVYIV4JKM7P3U3G54I54VDODX/events.json","paper":"https://pith.science/paper/KZVYIV4J"},"agent_actions":{"view_html":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX","download_json":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX.json","view_paper":"https://pith.science/paper/KZVYIV4J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.08685&json=true","fetch_graph":"https://pith.science/api/pith-number/KZVYIV4JKM7P3U3G54I54VDODX/graph.json","fetch_events":"https://pith.science/api/pith-number/KZVYIV4JKM7P3U3G54I54VDODX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX/action/storage_attestation","attest_author":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX/action/author_attestation","sign_citation":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX/action/citation_signature","submit_replication":"https://pith.science/pith/KZVYIV4JKM7P3U3G54I54VDODX/action/replication_record"}},"created_at":"2026-07-05T11:44:02.466572+00:00","updated_at":"2026-07-05T11:44:02.466572+00:00"}