{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:Q6GBTZX3I7JU4EZ37UNXGGFPC7","short_pith_number":"pith:Q6GBTZX3","schema_version":"1.0","canonical_sha256":"878c19e6fb47d34e133bfd1b7318af17f87bb3bf208b60910af7c18b4c07046c","source":{"kind":"arxiv","id":"2411.05282","version":4},"attestation_state":"computed","paper":{"title":"MicroScopiQ: Accelerating Foundational Models through Outlier-Aware Microscaling Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.AR","authors_text":"Akshat Ramachandran, Souvik Kundu, Tushar Krishna","submitted_at":"2024-11-08T02:25:45Z","abstract_excerpt":"Quantization of foundational models (FMs) is significantly more challenging than traditional DNNs due to the emergence of large magnitude values called outliers. Existing outlier-aware algorithm-architecture co-design techniques either use mixed-precision, retaining outliers at high precision but compromise hardware efficiency, or quantize inliers and outliers at the same precision, improving hardware efficiency at the cost of accuracy. To address this mutual exclusivity, we propose MicroScopiQ, a novel co-design technique that leverages pruning to complement outlier-aware quantization. MicroS"},"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":"2411.05282","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2024-11-08T02:25:45Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"b13af6f17c535d43fa4b856f265b626145f8ff6bdb77ade208466bd4d6efbede","abstract_canon_sha256":"5c044efb9fcc633b8a780c135d46fe628d6c911652eb620677165ff0d37ccf2b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:55:53.942091Z","signature_b64":"dzN5JdEFY02Af+bXMSyuqBdXtthtNPhL92eMc9K81K9YDiFx9TIrh1Bnje7jKn0uEwD3aViFez6Yk0Fux/1gCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"878c19e6fb47d34e133bfd1b7318af17f87bb3bf208b60910af7c18b4c07046c","last_reissued_at":"2026-07-05T10:55:53.941654Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:55:53.941654Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MicroScopiQ: Accelerating Foundational Models through Outlier-Aware Microscaling Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.AR","authors_text":"Akshat Ramachandran, Souvik Kundu, Tushar Krishna","submitted_at":"2024-11-08T02:25:45Z","abstract_excerpt":"Quantization of foundational models (FMs) is significantly more challenging than traditional DNNs due to the emergence of large magnitude values called outliers. Existing outlier-aware algorithm-architecture co-design techniques either use mixed-precision, retaining outliers at high precision but compromise hardware efficiency, or quantize inliers and outliers at the same precision, improving hardware efficiency at the cost of accuracy. To address this mutual exclusivity, we propose MicroScopiQ, a novel co-design technique that leverages pruning to complement outlier-aware quantization. MicroS"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.05282","kind":"arxiv","version":4},"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/2411.05282/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":"2411.05282","created_at":"2026-07-05T10:55:53.941711+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.05282v4","created_at":"2026-07-05T10:55:53.941711+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.05282","created_at":"2026-07-05T10:55:53.941711+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q6GBTZX3I7JU","created_at":"2026-07-05T10:55:53.941711+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q6GBTZX3I7JU4EZ3","created_at":"2026-07-05T10:55:53.941711+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q6GBTZX3","created_at":"2026-07-05T10:55:53.941711+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.09954","citing_title":"AIRCHITECT v2: Learning the Hardware Accelerator Design Space through Unified Representations","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7","json":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7.json","graph_json":"https://pith.science/api/pith-number/Q6GBTZX3I7JU4EZ37UNXGGFPC7/graph.json","events_json":"https://pith.science/api/pith-number/Q6GBTZX3I7JU4EZ37UNXGGFPC7/events.json","paper":"https://pith.science/paper/Q6GBTZX3"},"agent_actions":{"view_html":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7","download_json":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7.json","view_paper":"https://pith.science/paper/Q6GBTZX3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.05282&json=true","fetch_graph":"https://pith.science/api/pith-number/Q6GBTZX3I7JU4EZ37UNXGGFPC7/graph.json","fetch_events":"https://pith.science/api/pith-number/Q6GBTZX3I7JU4EZ37UNXGGFPC7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7/action/storage_attestation","attest_author":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7/action/author_attestation","sign_citation":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7/action/citation_signature","submit_replication":"https://pith.science/pith/Q6GBTZX3I7JU4EZ37UNXGGFPC7/action/replication_record"}},"created_at":"2026-07-05T10:55:53.941711+00:00","updated_at":"2026-07-05T10:55:53.941711+00:00"}