{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BTYGCXULXORYTNOTLGOOTSLRMH","short_pith_number":"pith:BTYGCXUL","schema_version":"1.0","canonical_sha256":"0cf0615e8bbba389b5d3599ce9c97161e567b553f14d26c6f01dd258589f7c95","source":{"kind":"arxiv","id":"2606.26861","version":1},"attestation_state":"computed","paper":{"title":"Cascaded Multi-Granularity Pruning for On-Device LLM Inference in Industrial IoT","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Gaoliang Peng, Jinghan Wang, Tianchen Liu, Wei Zhang, Xiaotong Huang, Yanjun Chen","submitted_at":"2026-06-25T10:44:48Z","abstract_excerpt":"Deploying large language models (LLMs) on Industrial Internet of Things (IIoT) edge devices demands extreme compression, yet existing structured pruning methods collapse at high compression ratios due to one-shot importance estimation, and their cross-architecture behavior remains unpredictable. This article presents a cascaded multi-granularity pruning framework that removes layers, attention heads, and feed-forward channels in coarse-to-fine order, with lightweight low-rank recovery between stages to re-estimate component importance. An information-theoretic analysis motivates this ordering,"},"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":"2606.26861","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-25T10:44:48Z","cross_cats_sorted":[],"title_canon_sha256":"7732063002ed61818983c1ea05c632cac91d583f5dddac4d85179dae13794e59","abstract_canon_sha256":"3542647ece4a32615a20bbff6bdac118dfed7c85825613d235bf7b3821495b61"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T01:16:02.181991Z","signature_b64":"l+fBD9sLraFbL2qolvzO3tka8aOjfBoM6bwvx3T+E2fBVZVT6+hsJuLvapjoyv0lN5IDIoZCZ6O/YUxW5tyhBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0cf0615e8bbba389b5d3599ce9c97161e567b553f14d26c6f01dd258589f7c95","last_reissued_at":"2026-06-26T01:16:02.181591Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T01:16:02.181591Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Cascaded Multi-Granularity Pruning for On-Device LLM Inference in Industrial IoT","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Gaoliang Peng, Jinghan Wang, Tianchen Liu, Wei Zhang, Xiaotong Huang, Yanjun Chen","submitted_at":"2026-06-25T10:44:48Z","abstract_excerpt":"Deploying large language models (LLMs) on Industrial Internet of Things (IIoT) edge devices demands extreme compression, yet existing structured pruning methods collapse at high compression ratios due to one-shot importance estimation, and their cross-architecture behavior remains unpredictable. This article presents a cascaded multi-granularity pruning framework that removes layers, attention heads, and feed-forward channels in coarse-to-fine order, with lightweight low-rank recovery between stages to re-estimate component importance. An information-theoretic analysis motivates this ordering,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26861","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/2606.26861/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":"2606.26861","created_at":"2026-06-26T01:16:02.181648+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.26861v1","created_at":"2026-06-26T01:16:02.181648+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26861","created_at":"2026-06-26T01:16:02.181648+00:00"},{"alias_kind":"pith_short_12","alias_value":"BTYGCXULXORY","created_at":"2026-06-26T01:16:02.181648+00:00"},{"alias_kind":"pith_short_16","alias_value":"BTYGCXULXORYTNOT","created_at":"2026-06-26T01:16:02.181648+00:00"},{"alias_kind":"pith_short_8","alias_value":"BTYGCXUL","created_at":"2026-06-26T01:16:02.181648+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/BTYGCXULXORYTNOTLGOOTSLRMH","json":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH.json","graph_json":"https://pith.science/api/pith-number/BTYGCXULXORYTNOTLGOOTSLRMH/graph.json","events_json":"https://pith.science/api/pith-number/BTYGCXULXORYTNOTLGOOTSLRMH/events.json","paper":"https://pith.science/paper/BTYGCXUL"},"agent_actions":{"view_html":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH","download_json":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH.json","view_paper":"https://pith.science/paper/BTYGCXUL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.26861&json=true","fetch_graph":"https://pith.science/api/pith-number/BTYGCXULXORYTNOTLGOOTSLRMH/graph.json","fetch_events":"https://pith.science/api/pith-number/BTYGCXULXORYTNOTLGOOTSLRMH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH/action/storage_attestation","attest_author":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH/action/author_attestation","sign_citation":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH/action/citation_signature","submit_replication":"https://pith.science/pith/BTYGCXULXORYTNOTLGOOTSLRMH/action/replication_record"}},"created_at":"2026-06-26T01:16:02.181648+00:00","updated_at":"2026-06-26T01:16:02.181648+00:00"}