{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:E4HUEHICMBGYTZWLPTE4CCG2BA","short_pith_number":"pith:E4HUEHIC","schema_version":"1.0","canonical_sha256":"270f421d02604d89e6cb7cc9c108da0810ebb92afa647750a7f9df34f66de333","source":{"kind":"arxiv","id":"2411.18447","version":1},"attestation_state":"computed","paper":{"title":"Continuous Autoregressive Models with Noise Augmentation Avoid Error Accumulation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.SD","eess.AS"],"primary_cat":"cs.LG","authors_text":"George Fazekas, Javier Nistal, Marco Pasini, Stefan Lattner","submitted_at":"2024-11-27T15:38:20Z","abstract_excerpt":"Autoregressive models are typically applied to sequences of discrete tokens, but recent research indicates that generating sequences of continuous embeddings in an autoregressive manner is also feasible. However, such Continuous Autoregressive Models (CAMs) can suffer from a decline in generation quality over extended sequences due to error accumulation during inference. We introduce a novel method to address this issue by injecting random noise into the input embeddings during training. This procedure makes the model robust against varying error levels at inference. We further reduce error ac"},"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":"2411.18447","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-27T15:38:20Z","cross_cats_sorted":["cs.AI","cs.SD","eess.AS"],"title_canon_sha256":"8daae5ef90894d8cd86261fcb9a5030cf444f2c40640356708bc4ad86af1d5c9","abstract_canon_sha256":"654928c0e4ef1e254eddfa726a5b2f4c24efcded28194cafcacd28bf98a08dd3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:41:16.608679Z","signature_b64":"1elfqnfKPqrBoGQL/vuWxQUdTnN7xDLivmEjpXuXNYcV3MFr/URGKYPFHpVz9304T8BpvsyjGa350E79THz8Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"270f421d02604d89e6cb7cc9c108da0810ebb92afa647750a7f9df34f66de333","last_reissued_at":"2026-07-05T09:41:16.608200Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:41:16.608200Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Continuous Autoregressive Models with Noise Augmentation Avoid Error Accumulation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.SD","eess.AS"],"primary_cat":"cs.LG","authors_text":"George Fazekas, Javier Nistal, Marco Pasini, Stefan Lattner","submitted_at":"2024-11-27T15:38:20Z","abstract_excerpt":"Autoregressive models are typically applied to sequences of discrete tokens, but recent research indicates that generating sequences of continuous embeddings in an autoregressive manner is also feasible. However, such Continuous Autoregressive Models (CAMs) can suffer from a decline in generation quality over extended sequences due to error accumulation during inference. We introduce a novel method to address this issue by injecting random noise into the input embeddings during training. This procedure makes the model robust against varying error levels at inference. We further reduce error ac"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.18447","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/2411.18447/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":"2411.18447","created_at":"2026-07-05T09:41:16.608257+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.18447v1","created_at":"2026-07-05T09:41:16.608257+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.18447","created_at":"2026-07-05T09:41:16.608257+00:00"},{"alias_kind":"pith_short_12","alias_value":"E4HUEHICMBGY","created_at":"2026-07-05T09:41:16.608257+00:00"},{"alias_kind":"pith_short_16","alias_value":"E4HUEHICMBGYTZWL","created_at":"2026-07-05T09:41:16.608257+00:00"},{"alias_kind":"pith_short_8","alias_value":"E4HUEHIC","created_at":"2026-07-05T09:41:16.608257+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03803","citing_title":"LiveBand: Live Accompaniment Generation in the Audio Domain","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24843","citing_title":"Adversarial Error Correction for Visual Autoregressive Generation","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22717","citing_title":"Live Music Diffusion Models: Efficient Fine-Tuning and Post-Training of Interactive Diffusion Music Generators","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16964","citing_title":"SemaVoice: Semantic-Aware Continuous Autoregressive Speech Synthesis","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2602.12370","citing_title":"LLaMo: Scaling Pretrained Language Models for Unified Motion Understanding and Generation with Continuous Autoregressive Tokens","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07955","citing_title":"TimeLesSeg: Unified Contrast-Agnostic Cross-Sectional and Longitudinal MS Lesion Segmentation via a Stochastic Generative Model","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA","json":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA.json","graph_json":"https://pith.science/api/pith-number/E4HUEHICMBGYTZWLPTE4CCG2BA/graph.json","events_json":"https://pith.science/api/pith-number/E4HUEHICMBGYTZWLPTE4CCG2BA/events.json","paper":"https://pith.science/paper/E4HUEHIC"},"agent_actions":{"view_html":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA","download_json":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA.json","view_paper":"https://pith.science/paper/E4HUEHIC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.18447&json=true","fetch_graph":"https://pith.science/api/pith-number/E4HUEHICMBGYTZWLPTE4CCG2BA/graph.json","fetch_events":"https://pith.science/api/pith-number/E4HUEHICMBGYTZWLPTE4CCG2BA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA/action/storage_attestation","attest_author":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA/action/author_attestation","sign_citation":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA/action/citation_signature","submit_replication":"https://pith.science/pith/E4HUEHICMBGYTZWLPTE4CCG2BA/action/replication_record"}},"created_at":"2026-07-05T09:41:16.608257+00:00","updated_at":"2026-07-05T09:41:16.608257+00:00"}