{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GBDHJFIHV4FXW5DVIOS4WIDE7I","short_pith_number":"pith:GBDHJFIH","schema_version":"1.0","canonical_sha256":"3046749507af0b7b747543a5cb2064fa006c3b6fdfe833c8158f2d13dc5a477d","source":{"kind":"arxiv","id":"2606.00487","version":1},"attestation_state":"computed","paper":{"title":"TAPS: Target-Aware Prefix Tree Selection for Diffusion-Drafted Speculative Decoding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Junnan Huang, Xinyu Chen, Zhuoyu Wang","submitted_at":"2026-05-30T02:39:40Z","abstract_excerpt":"Using a diffusion model for parallel drafting is a promising approach for speculative decoding. By predicting tokens at multiple future positions in a single forward pass, diffusion drafters substantially reduce drafting latency. However, this shifts the bottleneck to verification: verifying a single sequence limits acceptance length, while verifying large draft trees incurs excessive target-model latency. We identify a key mismatch in existing draft-tree methods: existing diffusion-tree methods rank nodes by the marginal probability, ignoring that verification is prefix-conditioned. As a resu"},"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":"2606.00487","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-30T02:39:40Z","cross_cats_sorted":[],"title_canon_sha256":"83111eeb480026c9a25e2392e78d4f4b3827b01da49eefa3f2a0b1ce19115520","abstract_canon_sha256":"3e6363dcaf9cc89187fc5a0ab13a6e65e6a13201567c7042a77083583578e516"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:55.992016Z","signature_b64":"gePUXKVyH3DcJiuH71PnnIq1iI08UyEX4Wpg9XFKNZTmFQIKRUppUmnZe8emmkL1iSJoZTIT+z47gYcdBgWKBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3046749507af0b7b747543a5cb2064fa006c3b6fdfe833c8158f2d13dc5a477d","last_reissued_at":"2026-06-02T01:03:55.991591Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:55.991591Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TAPS: Target-Aware Prefix Tree Selection for Diffusion-Drafted Speculative Decoding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Junnan Huang, Xinyu Chen, Zhuoyu Wang","submitted_at":"2026-05-30T02:39:40Z","abstract_excerpt":"Using a diffusion model for parallel drafting is a promising approach for speculative decoding. By predicting tokens at multiple future positions in a single forward pass, diffusion drafters substantially reduce drafting latency. However, this shifts the bottleneck to verification: verifying a single sequence limits acceptance length, while verifying large draft trees incurs excessive target-model latency. We identify a key mismatch in existing draft-tree methods: existing diffusion-tree methods rank nodes by the marginal probability, ignoring that verification is prefix-conditioned. As a resu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00487","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/2606.00487/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":"2606.00487","created_at":"2026-06-02T01:03:55.991651+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00487v1","created_at":"2026-06-02T01:03:55.991651+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00487","created_at":"2026-06-02T01:03:55.991651+00:00"},{"alias_kind":"pith_short_12","alias_value":"GBDHJFIHV4FX","created_at":"2026-06-02T01:03:55.991651+00:00"},{"alias_kind":"pith_short_16","alias_value":"GBDHJFIHV4FXW5DV","created_at":"2026-06-02T01:03:55.991651+00:00"},{"alias_kind":"pith_short_8","alias_value":"GBDHJFIH","created_at":"2026-06-02T01:03:55.991651+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I","json":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I.json","graph_json":"https://pith.science/api/pith-number/GBDHJFIHV4FXW5DVIOS4WIDE7I/graph.json","events_json":"https://pith.science/api/pith-number/GBDHJFIHV4FXW5DVIOS4WIDE7I/events.json","paper":"https://pith.science/paper/GBDHJFIH"},"agent_actions":{"view_html":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I","download_json":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I.json","view_paper":"https://pith.science/paper/GBDHJFIH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00487&json=true","fetch_graph":"https://pith.science/api/pith-number/GBDHJFIHV4FXW5DVIOS4WIDE7I/graph.json","fetch_events":"https://pith.science/api/pith-number/GBDHJFIHV4FXW5DVIOS4WIDE7I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I/action/storage_attestation","attest_author":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I/action/author_attestation","sign_citation":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I/action/citation_signature","submit_replication":"https://pith.science/pith/GBDHJFIHV4FXW5DVIOS4WIDE7I/action/replication_record"}},"created_at":"2026-06-02T01:03:55.991651+00:00","updated_at":"2026-06-02T01:03:55.991651+00:00"}