{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FR2JQEAOBYF656QP6RFG2E7ZD2","short_pith_number":"pith:FR2JQEAO","schema_version":"1.0","canonical_sha256":"2c7498100e0e0beefa0ff44a6d13f91eb3f6c9c56c845ee83010dc5a77a34f70","source":{"kind":"arxiv","id":"2506.17826","version":1},"attestation_state":"computed","paper":{"title":"Actionable Interpretability via Causal Hypergraphs: Unravelling Batch Size Effects in Deep Learning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anoushka Harit, Pietro Lio, Zhongtian Sun","submitted_at":"2025-06-21T21:38:43Z","abstract_excerpt":"While the impact of batch size on generalisation is well studied in vision tasks, its causal mechanisms remain underexplored in graph and text domains. We introduce a hypergraph-based causal framework, HGCNet, that leverages deep structural causal models (DSCMs) to uncover how batch size influences generalisation via gradient noise, minima sharpness, and model complexity. Unlike prior approaches based on static pairwise dependencies, HGCNet employs hypergraphs to capture higher-order interactions across training dynamics. Using do-calculus, we quantify direct and mediated effects of batch size"},"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":"2506.17826","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-21T21:38:43Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"934713a16a5e70b7514795d10a407d9355fb0c49558582d7fa160c42aa2cc72f","abstract_canon_sha256":"3c0e582ad17c1361b7ea786650e44ad4707de50f124d1bb8d02ffda41bb1d4f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:23.846508Z","signature_b64":"MyuWNy5qyJqrwtoj2HFqAuFnNad+2u5zdWYriMEbW/a9yaEWfF/X1weVBV5tCjW9LAInr5Twelpk58sVgW7KCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2c7498100e0e0beefa0ff44a6d13f91eb3f6c9c56c845ee83010dc5a77a34f70","last_reissued_at":"2026-07-05T11:25:23.845969Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:23.845969Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Actionable Interpretability via Causal Hypergraphs: Unravelling Batch Size Effects in Deep Learning","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anoushka Harit, Pietro Lio, Zhongtian Sun","submitted_at":"2025-06-21T21:38:43Z","abstract_excerpt":"While the impact of batch size on generalisation is well studied in vision tasks, its causal mechanisms remain underexplored in graph and text domains. We introduce a hypergraph-based causal framework, HGCNet, that leverages deep structural causal models (DSCMs) to uncover how batch size influences generalisation via gradient noise, minima sharpness, and model complexity. Unlike prior approaches based on static pairwise dependencies, HGCNet employs hypergraphs to capture higher-order interactions across training dynamics. Using do-calculus, we quantify direct and mediated effects of batch size"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.17826","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/2506.17826/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":"2506.17826","created_at":"2026-07-05T11:25:23.846040+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.17826v1","created_at":"2026-07-05T11:25:23.846040+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.17826","created_at":"2026-07-05T11:25:23.846040+00:00"},{"alias_kind":"pith_short_12","alias_value":"FR2JQEAOBYF6","created_at":"2026-07-05T11:25:23.846040+00:00"},{"alias_kind":"pith_short_16","alias_value":"FR2JQEAOBYF656QP","created_at":"2026-07-05T11:25:23.846040+00:00"},{"alias_kind":"pith_short_8","alias_value":"FR2JQEAO","created_at":"2026-07-05T11:25:23.846040+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.09334","citing_title":"RicciFlowRec: A Geometric Root Cause Recommender Using Ricci Curvature on Financial Graphs","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2","json":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2.json","graph_json":"https://pith.science/api/pith-number/FR2JQEAOBYF656QP6RFG2E7ZD2/graph.json","events_json":"https://pith.science/api/pith-number/FR2JQEAOBYF656QP6RFG2E7ZD2/events.json","paper":"https://pith.science/paper/FR2JQEAO"},"agent_actions":{"view_html":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2","download_json":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2.json","view_paper":"https://pith.science/paper/FR2JQEAO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.17826&json=true","fetch_graph":"https://pith.science/api/pith-number/FR2JQEAOBYF656QP6RFG2E7ZD2/graph.json","fetch_events":"https://pith.science/api/pith-number/FR2JQEAOBYF656QP6RFG2E7ZD2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2/action/storage_attestation","attest_author":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2/action/author_attestation","sign_citation":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2/action/citation_signature","submit_replication":"https://pith.science/pith/FR2JQEAOBYF656QP6RFG2E7ZD2/action/replication_record"}},"created_at":"2026-07-05T11:25:23.846040+00:00","updated_at":"2026-07-05T11:25:23.846040+00:00"}