{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ADUU6TAJ4JU6LCG3W6CQJ3JWDD","short_pith_number":"pith:ADUU6TAJ","schema_version":"1.0","canonical_sha256":"00e94f4c09e269e588dbb78504ed3618d2c6017afbe8e6f4eb0ddeee6a909ad0","source":{"kind":"arxiv","id":"2601.13563","version":5},"attestation_state":"computed","paper":{"title":"ButterflyMoE: Compression-Scalable Ternary Experts via Structured Butterfly Orbits","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aryan Karmore","submitted_at":"2026-01-20T03:39:33Z","abstract_excerpt":"In current Mixture of Experts (MoE) architectures, linear memory scaling is present, the memory grows as the number of experts increases. $N$ independent expert weight matrices require $\\mathcal{O}(N \\cdot d^2)$ memory which exceeds the memory budget of edge devices. Current compression methods like quantization, pruning, and low-rank factorization reduce constant factors, but the scaling bottleneck is still unresolved. We introduce ButterflyMoE, a method which treats experts not as independent matrices but as geometric reorientations of a shared quantized substrate. Diversity amongst the expe"},"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":"2601.13563","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-20T03:39:33Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d378b5c9ca99ba38f3bbd961898c466b344e1c723e91ba1aab00a8ef53cdaadc","abstract_canon_sha256":"33e80bbc05de205d7c566c4ef2a4d31aba15846f1ca1911399cb00046c33dc07"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-09T01:19:50.963365Z","signature_b64":"vle7AFzGI75oiW8ZCn0qXbyX3sQT7Ycwfaw/lUgsDW9yZcdBmMSOU+SAFKbIKFlOnH9W7FPWqkGizZzonCFYDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"00e94f4c09e269e588dbb78504ed3618d2c6017afbe8e6f4eb0ddeee6a909ad0","last_reissued_at":"2026-07-09T01:19:50.962872Z","signature_status":"signed_v1","first_computed_at":"2026-07-09T01:19:50.962872Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ButterflyMoE: Compression-Scalable Ternary Experts via Structured Butterfly Orbits","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Aryan Karmore","submitted_at":"2026-01-20T03:39:33Z","abstract_excerpt":"In current Mixture of Experts (MoE) architectures, linear memory scaling is present, the memory grows as the number of experts increases. $N$ independent expert weight matrices require $\\mathcal{O}(N \\cdot d^2)$ memory which exceeds the memory budget of edge devices. Current compression methods like quantization, pruning, and low-rank factorization reduce constant factors, but the scaling bottleneck is still unresolved. We introduce ButterflyMoE, a method which treats experts not as independent matrices but as geometric reorientations of a shared quantized substrate. Diversity amongst the expe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.13563","kind":"arxiv","version":5},"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/2601.13563/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":"2601.13563","created_at":"2026-07-09T01:19:50.962931+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.13563v5","created_at":"2026-07-09T01:19:50.962931+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.13563","created_at":"2026-07-09T01:19:50.962931+00:00"},{"alias_kind":"pith_short_12","alias_value":"ADUU6TAJ4JU6","created_at":"2026-07-09T01:19:50.962931+00:00"},{"alias_kind":"pith_short_16","alias_value":"ADUU6TAJ4JU6LCG3","created_at":"2026-07-09T01:19:50.962931+00:00"},{"alias_kind":"pith_short_8","alias_value":"ADUU6TAJ","created_at":"2026-07-09T01:19:50.962931+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/ADUU6TAJ4JU6LCG3W6CQJ3JWDD","json":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD.json","graph_json":"https://pith.science/api/pith-number/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/graph.json","events_json":"https://pith.science/api/pith-number/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/events.json","paper":"https://pith.science/paper/ADUU6TAJ"},"agent_actions":{"view_html":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD","download_json":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD.json","view_paper":"https://pith.science/paper/ADUU6TAJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.13563&json=true","fetch_graph":"https://pith.science/api/pith-number/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/graph.json","fetch_events":"https://pith.science/api/pith-number/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/action/storage_attestation","attest_author":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/action/author_attestation","sign_citation":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/action/citation_signature","submit_replication":"https://pith.science/pith/ADUU6TAJ4JU6LCG3W6CQJ3JWDD/action/replication_record"}},"created_at":"2026-07-09T01:19:50.962931+00:00","updated_at":"2026-07-09T01:19:50.962931+00:00"}