{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5Z7OJXXAHFMXAYCU4NXZHJT4YU","short_pith_number":"pith:5Z7OJXXA","schema_version":"1.0","canonical_sha256":"ee7ee4dee03959706054e36f93a67cc50460f3de14931996e55a63c176263ada","source":{"kind":"arxiv","id":"2606.20537","version":1},"attestation_state":"computed","paper":{"title":"Execution-State Capsules: Graph-Bound Execution-State Checkpoint and Restore for Low-Latency, Small-Batch, On-Device Physical-AI Serving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Liang Su","submitted_at":"2026-06-18T17:49:36Z","abstract_excerpt":"Mainstream LLM serving systems reuse prefix work mainly through paged or radix key-value (KV) caches. This is highly effective for high-throughput, high-concurrency serving, but it manages only one positional fragment of execution state: the KV cache. We study the opposite regime: low-latency, small-batch, on-device physical-AI serving, where interactive LLM agents, speech systems, and robot policies repeatedly branch, reset, interrupt, and re-enter under tight responsiveness budgets. We introduce execution-state capsules, a graph-bound checkpoint and restore mechanism for the complete restora"},"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":"2606.20537","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-18T17:49:36Z","cross_cats_sorted":["cs.DC"],"title_canon_sha256":"1d7eecd035c2f028c541151f22de7f7246f2287ab26687bb3b8c3f7727a8af5d","abstract_canon_sha256":"ef93567bb5fd63db439d118e6e39781104342d46c85f62611bb33461e7d2ba07"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:13:14.769306Z","signature_b64":"Ht0+PZ6oYjEpAzMODNbpD+BbQbBruaiowRIZ7mX3Wf+XsvrT93FNJ1Q+H1cVL9Nc1IZDxc7HdqRKXBjGvco9CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ee7ee4dee03959706054e36f93a67cc50460f3de14931996e55a63c176263ada","last_reissued_at":"2026-06-19T16:13:14.768697Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:13:14.768697Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Execution-State Capsules: Graph-Bound Execution-State Checkpoint and Restore for Low-Latency, Small-Batch, On-Device Physical-AI Serving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.DC"],"primary_cat":"cs.LG","authors_text":"Liang Su","submitted_at":"2026-06-18T17:49:36Z","abstract_excerpt":"Mainstream LLM serving systems reuse prefix work mainly through paged or radix key-value (KV) caches. This is highly effective for high-throughput, high-concurrency serving, but it manages only one positional fragment of execution state: the KV cache. We study the opposite regime: low-latency, small-batch, on-device physical-AI serving, where interactive LLM agents, speech systems, and robot policies repeatedly branch, reset, interrupt, and re-enter under tight responsiveness budgets. We introduce execution-state capsules, a graph-bound checkpoint and restore mechanism for the complete restora"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.20537","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/2606.20537/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":"2606.20537","created_at":"2026-06-19T16:13:14.768753+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.20537v1","created_at":"2026-06-19T16:13:14.768753+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.20537","created_at":"2026-06-19T16:13:14.768753+00:00"},{"alias_kind":"pith_short_12","alias_value":"5Z7OJXXAHFMX","created_at":"2026-06-19T16:13:14.768753+00:00"},{"alias_kind":"pith_short_16","alias_value":"5Z7OJXXAHFMXAYCU","created_at":"2026-06-19T16:13:14.768753+00:00"},{"alias_kind":"pith_short_8","alias_value":"5Z7OJXXA","created_at":"2026-06-19T16:13:14.768753+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.02501","citing_title":"Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots","ref_index":46,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU","json":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU.json","graph_json":"https://pith.science/api/pith-number/5Z7OJXXAHFMXAYCU4NXZHJT4YU/graph.json","events_json":"https://pith.science/api/pith-number/5Z7OJXXAHFMXAYCU4NXZHJT4YU/events.json","paper":"https://pith.science/paper/5Z7OJXXA"},"agent_actions":{"view_html":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU","download_json":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU.json","view_paper":"https://pith.science/paper/5Z7OJXXA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.20537&json=true","fetch_graph":"https://pith.science/api/pith-number/5Z7OJXXAHFMXAYCU4NXZHJT4YU/graph.json","fetch_events":"https://pith.science/api/pith-number/5Z7OJXXAHFMXAYCU4NXZHJT4YU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU/action/storage_attestation","attest_author":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU/action/author_attestation","sign_citation":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU/action/citation_signature","submit_replication":"https://pith.science/pith/5Z7OJXXAHFMXAYCU4NXZHJT4YU/action/replication_record"}},"created_at":"2026-06-19T16:13:14.768753+00:00","updated_at":"2026-06-19T16:13:14.768753+00:00"}