{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:HGRCIWDROIJFWHF36FTR3JLEAI","short_pith_number":"pith:HGRCIWDR","schema_version":"1.0","canonical_sha256":"39a224587172125b1cbbf1671da564020fdff51bcccd66383ff1d231a0ae55b0","source":{"kind":"arxiv","id":"2402.10450","version":3},"attestation_state":"computed","paper":{"title":"PRISE: LLM-Style Sequence Compression for Learning Temporal Action Abstractions in Control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrey Kolobov, Ching-An Cheng, Furong Huang, Hal Daum\\'e III, Ruijie Zheng","submitted_at":"2024-02-16T04:55:09Z","abstract_excerpt":"Temporal action abstractions, along with belief state representations, are a powerful knowledge sharing mechanism for sequential decision making. In this work, we propose a novel view that treats inducing temporal action abstractions as a sequence compression problem. To do so, we bring a subtle but critical component of LLM training pipelines -- input tokenization via byte pair encoding (BPE) -- to the seemingly distant task of learning skills of variable time span in continuous control domains. We introduce an approach called Primitive Sequence Encoding (PRISE) that combines continuous actio"},"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":"2402.10450","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-16T04:55:09Z","cross_cats_sorted":[],"title_canon_sha256":"0d1997fafc43017ab384da6a4cf0893b1ecffbe853094201948ae0e1c5756a8b","abstract_canon_sha256":"c5f9226d732bf91cae13e984ca0be402a8788d0f20f7e312d10958bd3a564bd6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:06.040602Z","signature_b64":"/0IoZk0rO5m+UymANBRLMI3FNWY1Q7/C/ySoFSW+DFc6Arx7oC8lmfhIHq8XFHq3/Ndomrk8lGqpkQVNzkfnAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"39a224587172125b1cbbf1671da564020fdff51bcccd66383ff1d231a0ae55b0","last_reissued_at":"2026-07-05T08:28:06.040195Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:06.040195Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PRISE: LLM-Style Sequence Compression for Learning Temporal Action Abstractions in Control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Andrey Kolobov, Ching-An Cheng, Furong Huang, Hal Daum\\'e III, Ruijie Zheng","submitted_at":"2024-02-16T04:55:09Z","abstract_excerpt":"Temporal action abstractions, along with belief state representations, are a powerful knowledge sharing mechanism for sequential decision making. In this work, we propose a novel view that treats inducing temporal action abstractions as a sequence compression problem. To do so, we bring a subtle but critical component of LLM training pipelines -- input tokenization via byte pair encoding (BPE) -- to the seemingly distant task of learning skills of variable time span in continuous control domains. We introduce an approach called Primitive Sequence Encoding (PRISE) that combines continuous actio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.10450","kind":"arxiv","version":3},"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/2402.10450/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":"2402.10450","created_at":"2026-07-05T08:28:06.040243+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.10450v3","created_at":"2026-07-05T08:28:06.040243+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.10450","created_at":"2026-07-05T08:28:06.040243+00:00"},{"alias_kind":"pith_short_12","alias_value":"HGRCIWDROIJF","created_at":"2026-07-05T08:28:06.040243+00:00"},{"alias_kind":"pith_short_16","alias_value":"HGRCIWDROIJFWHF3","created_at":"2026-07-05T08:28:06.040243+00:00"},{"alias_kind":"pith_short_8","alias_value":"HGRCIWDR","created_at":"2026-07-05T08:28:06.040243+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22480","citing_title":"ARP: Enhancing Quantized Skill Abstractions via Visual Alignment and Iterative Refinement for Robotic Manipulation","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18597","citing_title":"Latent Action Reparameterization for Efficient Agent Inference","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI","json":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI.json","graph_json":"https://pith.science/api/pith-number/HGRCIWDROIJFWHF36FTR3JLEAI/graph.json","events_json":"https://pith.science/api/pith-number/HGRCIWDROIJFWHF36FTR3JLEAI/events.json","paper":"https://pith.science/paper/HGRCIWDR"},"agent_actions":{"view_html":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI","download_json":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI.json","view_paper":"https://pith.science/paper/HGRCIWDR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.10450&json=true","fetch_graph":"https://pith.science/api/pith-number/HGRCIWDROIJFWHF36FTR3JLEAI/graph.json","fetch_events":"https://pith.science/api/pith-number/HGRCIWDROIJFWHF36FTR3JLEAI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI/action/storage_attestation","attest_author":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI/action/author_attestation","sign_citation":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI/action/citation_signature","submit_replication":"https://pith.science/pith/HGRCIWDROIJFWHF36FTR3JLEAI/action/replication_record"}},"created_at":"2026-07-05T08:28:06.040243+00:00","updated_at":"2026-07-05T08:28:06.040243+00:00"}