{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BSKHYYNEYCX5Q5ZVMFKMYRUQVT","short_pith_number":"pith:BSKHYYNE","schema_version":"1.0","canonical_sha256":"0c947c61a4c0afd877356154cc4690accd231500e2d9f7ae672e4e68caf4d22d","source":{"kind":"arxiv","id":"2410.06424","version":2},"attestation_state":"computed","paper":{"title":"Restructuring Vector Quantization with the Rotation Trick","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Aniketh Iyengar, Christopher Fifty, Christopher R\\'e, Dennis Duan, Ehsan Amid, Jerry W. Liu, Ronald G. Junkins, Sebastian Thrun","submitted_at":"2024-10-08T23:39:34Z","abstract_excerpt":"Vector Quantized Variational AutoEncoders (VQ-VAEs) are designed to compress a continuous input to a discrete latent space and reconstruct it with minimal distortion. They operate by maintaining a set of vectors -- often referred to as the codebook -- and quantizing each encoder output to the nearest vector in the codebook. However, as vector quantization is non-differentiable, the gradient to the encoder flows around the vector quantization layer rather than through it in a straight-through approximation. This approximation may be undesirable as all information from the vector quantization op"},"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":"2410.06424","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-08T23:39:34Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"b21b14c6f78e420ef6bf1a6874b8c89c2c1a8688730079c14f23f0b4881c6cee","abstract_canon_sha256":"67daca251606a348b5d0cf179690b06d3796b948c8768b42fba5d68c81dbdb00"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:31:48.151703Z","signature_b64":"5rSxtj1u97XL64exPNLgmtK63lCdaORZmMrMsEq82lDAaYz7OEAksKLLhnYIC1OOZ+d3Wj7lFz55R5f82mfXBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0c947c61a4c0afd877356154cc4690accd231500e2d9f7ae672e4e68caf4d22d","last_reissued_at":"2026-07-05T10:31:48.150968Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:31:48.150968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Restructuring Vector Quantization with the Rotation Trick","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Aniketh Iyengar, Christopher Fifty, Christopher R\\'e, Dennis Duan, Ehsan Amid, Jerry W. Liu, Ronald G. Junkins, Sebastian Thrun","submitted_at":"2024-10-08T23:39:34Z","abstract_excerpt":"Vector Quantized Variational AutoEncoders (VQ-VAEs) are designed to compress a continuous input to a discrete latent space and reconstruct it with minimal distortion. They operate by maintaining a set of vectors -- often referred to as the codebook -- and quantizing each encoder output to the nearest vector in the codebook. However, as vector quantization is non-differentiable, the gradient to the encoder flows around the vector quantization layer rather than through it in a straight-through approximation. This approximation may be undesirable as all information from the vector quantization op"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.06424","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/2410.06424/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":"2410.06424","created_at":"2026-07-05T10:31:48.151057+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.06424v2","created_at":"2026-07-05T10:31:48.151057+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.06424","created_at":"2026-07-05T10:31:48.151057+00:00"},{"alias_kind":"pith_short_12","alias_value":"BSKHYYNEYCX5","created_at":"2026-07-05T10:31:48.151057+00:00"},{"alias_kind":"pith_short_16","alias_value":"BSKHYYNEYCX5Q5ZV","created_at":"2026-07-05T10:31:48.151057+00:00"},{"alias_kind":"pith_short_8","alias_value":"BSKHYYNE","created_at":"2026-07-05T10:31:48.151057+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.27696","citing_title":"Structure over Pixels: Learning Variable-Length Visual Programs","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT","json":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT.json","graph_json":"https://pith.science/api/pith-number/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/graph.json","events_json":"https://pith.science/api/pith-number/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/events.json","paper":"https://pith.science/paper/BSKHYYNE"},"agent_actions":{"view_html":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT","download_json":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT.json","view_paper":"https://pith.science/paper/BSKHYYNE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.06424&json=true","fetch_graph":"https://pith.science/api/pith-number/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/graph.json","fetch_events":"https://pith.science/api/pith-number/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/action/storage_attestation","attest_author":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/action/author_attestation","sign_citation":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/action/citation_signature","submit_replication":"https://pith.science/pith/BSKHYYNEYCX5Q5ZVMFKMYRUQVT/action/replication_record"}},"created_at":"2026-07-05T10:31:48.151057+00:00","updated_at":"2026-07-05T10:31:48.151057+00:00"}