{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:26RQW4MN64KEJCYOYSTO5IWFM2","short_pith_number":"pith:26RQW4MN","schema_version":"1.0","canonical_sha256":"d7a30b718df714448b0ec4a6eea2c566a9c1a1d02ed87244166ae4e202b389eb","source":{"kind":"arxiv","id":"2205.07547","version":2},"attestation_state":"computed","paper":{"title":"SQ-VAE: Variational Bayes on Discrete Representation with Self-annealed Stochastic Quantization","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Chieh-Hsin Lai, Junki Ohmura, Naoki Murata, Shusuke Takahashi, Takashi Shibuya, Toshimitsu Uesaka, Toshiyuki Kumakura, WeiHsiang Liao, Yuhta Takida, Yuki Mitsufuji","submitted_at":"2022-05-16T09:49:37Z","abstract_excerpt":"One noted issue of vector-quantized variational autoencoder (VQ-VAE) is that the learned discrete representation uses only a fraction of the full capacity of the codebook, also known as codebook collapse. We hypothesize that the training scheme of VQ-VAE, which involves some carefully designed heuristics, underlies this issue. In this paper, we propose a new training scheme that extends the standard VAE via novel stochastic dequantization and quantization, called stochastically quantized variational autoencoder (SQ-VAE). In SQ-VAE, we observe a trend that the quantization is stochastic at 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":"2205.07547","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2022-05-16T09:49:37Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"70ab0bc06101d7b5c9ac50339dac923b679a6dc68ddba4c93b4d6a7f28a49853","abstract_canon_sha256":"f197c3de13bf5b427028b62e36a966f5fc0e78ca85325abd5eca6cba896908ee"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:30:33.890398Z","signature_b64":"DC/dGF03wAaZcV/nkdNkWsrkDE5cbBLUhiWzJcFOFgq2ch4u3dQ9mSBgzr6r0BfK5V/pML7zsRyYgl2Tnth9BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7a30b718df714448b0ec4a6eea2c566a9c1a1d02ed87244166ae4e202b389eb","last_reissued_at":"2026-07-05T04:30:33.889904Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:30:33.889904Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SQ-VAE: Variational Bayes on Discrete Representation with Self-annealed Stochastic Quantization","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Chieh-Hsin Lai, Junki Ohmura, Naoki Murata, Shusuke Takahashi, Takashi Shibuya, Toshimitsu Uesaka, Toshiyuki Kumakura, WeiHsiang Liao, Yuhta Takida, Yuki Mitsufuji","submitted_at":"2022-05-16T09:49:37Z","abstract_excerpt":"One noted issue of vector-quantized variational autoencoder (VQ-VAE) is that the learned discrete representation uses only a fraction of the full capacity of the codebook, also known as codebook collapse. We hypothesize that the training scheme of VQ-VAE, which involves some carefully designed heuristics, underlies this issue. In this paper, we propose a new training scheme that extends the standard VAE via novel stochastic dequantization and quantization, called stochastically quantized variational autoencoder (SQ-VAE). In SQ-VAE, we observe a trend that the quantization is stochastic at the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.07547","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/2205.07547/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":"2205.07547","created_at":"2026-07-05T04:30:33.889962+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.07547v2","created_at":"2026-07-05T04:30:33.889962+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.07547","created_at":"2026-07-05T04:30:33.889962+00:00"},{"alias_kind":"pith_short_12","alias_value":"26RQW4MN64KE","created_at":"2026-07-05T04:30:33.889962+00:00"},{"alias_kind":"pith_short_16","alias_value":"26RQW4MN64KEJCYO","created_at":"2026-07-05T04:30:33.889962+00:00"},{"alias_kind":"pith_short_8","alias_value":"26RQW4MN","created_at":"2026-07-05T04:30:33.889962+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.13517","citing_title":"ArcVQ-VAE: A Spherical Vector Quantization Framework with ArcCosine Additive Margin","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2509.13482","citing_title":"Improving 3D Gaussian Splatting Compression by Scene-Adaptive Lattice Vector Quantization","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2512.01537","citing_title":"Two-Dimensional Quantization for Geometry-Aware Audio Coding","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2309.15505","citing_title":"Finite Scalar Quantization: VQ-VAE Made Simple","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2601.19711","citing_title":"Differentiable Semantic ID for Generative Recommendation","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13517","citing_title":"ArcVQ-VAE: A Spherical Vector Quantization Framework with ArcCosine Additive Margin","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06548","citing_title":"Continuous Latent Diffusion Language Model","ref_index":90,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2","json":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2.json","graph_json":"https://pith.science/api/pith-number/26RQW4MN64KEJCYOYSTO5IWFM2/graph.json","events_json":"https://pith.science/api/pith-number/26RQW4MN64KEJCYOYSTO5IWFM2/events.json","paper":"https://pith.science/paper/26RQW4MN"},"agent_actions":{"view_html":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2","download_json":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2.json","view_paper":"https://pith.science/paper/26RQW4MN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.07547&json=true","fetch_graph":"https://pith.science/api/pith-number/26RQW4MN64KEJCYOYSTO5IWFM2/graph.json","fetch_events":"https://pith.science/api/pith-number/26RQW4MN64KEJCYOYSTO5IWFM2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2/action/storage_attestation","attest_author":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2/action/author_attestation","sign_citation":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2/action/citation_signature","submit_replication":"https://pith.science/pith/26RQW4MN64KEJCYOYSTO5IWFM2/action/replication_record"}},"created_at":"2026-07-05T04:30:33.889962+00:00","updated_at":"2026-07-05T04:30:33.889962+00:00"}