{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KKB3BH4YXSSFOXH7FM3A6YQRCW","short_pith_number":"pith:KKB3BH4Y","schema_version":"1.0","canonical_sha256":"5283b09f98bca4575cff2b360f621115a936d1eb4c0d844337269ecd2b1abb47","source":{"kind":"arxiv","id":"2505.18164","version":1},"attestation_state":"computed","paper":{"title":"Model-Distributed Inference for Large Language Models at the Edge","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Davide Macario, Erdem Koyuncu, Hulya Seferoglu","submitted_at":"2025-05-13T12:07:37Z","abstract_excerpt":"We introduce Model-Distributed Inference for Large-Language Models (MDI-LLM), a novel framework designed to facilitate the deployment of state-of-the-art large-language models (LLMs) across low-power devices at the edge. This is accomplished by dividing the model into multiple partitions, which are then assigned to different devices/nodes within the network. These nodes exchange intermediate activation vectors via device-to-device links, enabling collaborative computation. To enhance the efficiency of this process, we propose the \"recurrent pipeline parallelism\" technique, which reduces idle t"},"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":"2505.18164","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-13T12:07:37Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"dca0e231eb1e0bdf97958d4ec10126c464db907a920cf02060fc5ed34dff4728","abstract_canon_sha256":"094d948d6a1f4a5352c0cb604a2f5b3878947889683d9908060e138ea798d752"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:08:42.217680Z","signature_b64":"44KI008p22Zplv0TRJFJUuIuzwNRlnEf+6QEy4imt3oROeuGqqK/RetATuhXgJvgvnRuM9ylUX9+Lx3QIuYfDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5283b09f98bca4575cff2b360f621115a936d1eb4c0d844337269ecd2b1abb47","last_reissued_at":"2026-07-05T11:08:42.217269Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:08:42.217269Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Model-Distributed Inference for Large Language Models at the Edge","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Davide Macario, Erdem Koyuncu, Hulya Seferoglu","submitted_at":"2025-05-13T12:07:37Z","abstract_excerpt":"We introduce Model-Distributed Inference for Large-Language Models (MDI-LLM), a novel framework designed to facilitate the deployment of state-of-the-art large-language models (LLMs) across low-power devices at the edge. This is accomplished by dividing the model into multiple partitions, which are then assigned to different devices/nodes within the network. These nodes exchange intermediate activation vectors via device-to-device links, enabling collaborative computation. To enhance the efficiency of this process, we propose the \"recurrent pipeline parallelism\" technique, which reduces idle t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18164","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/2505.18164/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":"2505.18164","created_at":"2026-07-05T11:08:42.217328+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.18164v1","created_at":"2026-07-05T11:08:42.217328+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18164","created_at":"2026-07-05T11:08:42.217328+00:00"},{"alias_kind":"pith_short_12","alias_value":"KKB3BH4YXSSF","created_at":"2026-07-05T11:08:42.217328+00:00"},{"alias_kind":"pith_short_16","alias_value":"KKB3BH4YXSSFOXH7","created_at":"2026-07-05T11:08:42.217328+00:00"},{"alias_kind":"pith_short_8","alias_value":"KKB3BH4Y","created_at":"2026-07-05T11:08:42.217328+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.00672","citing_title":"Toward Edge General Intelligence with Multiple-Large Language Model (Multi-LLM): Architecture, Trust, and Orchestration","ref_index":127,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW","json":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW.json","graph_json":"https://pith.science/api/pith-number/KKB3BH4YXSSFOXH7FM3A6YQRCW/graph.json","events_json":"https://pith.science/api/pith-number/KKB3BH4YXSSFOXH7FM3A6YQRCW/events.json","paper":"https://pith.science/paper/KKB3BH4Y"},"agent_actions":{"view_html":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW","download_json":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW.json","view_paper":"https://pith.science/paper/KKB3BH4Y","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.18164&json=true","fetch_graph":"https://pith.science/api/pith-number/KKB3BH4YXSSFOXH7FM3A6YQRCW/graph.json","fetch_events":"https://pith.science/api/pith-number/KKB3BH4YXSSFOXH7FM3A6YQRCW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW/action/storage_attestation","attest_author":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW/action/author_attestation","sign_citation":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW/action/citation_signature","submit_replication":"https://pith.science/pith/KKB3BH4YXSSFOXH7FM3A6YQRCW/action/replication_record"}},"created_at":"2026-07-05T11:08:42.217328+00:00","updated_at":"2026-07-05T11:08:42.217328+00:00"}