{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:PT6NKGMBL7SZ6COSSYK6OL6A7H","short_pith_number":"pith:PT6NKGMB","schema_version":"1.0","canonical_sha256":"7cfcd519815fe59f09d29615e72fc0f9dfd6ca6d84f3f3bc202a3c936c17a524","source":{"kind":"arxiv","id":"1603.01025","version":2},"attestation_state":"computed","paper":{"title":"Convolutional Neural Networks using Logarithmic Data Representation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.NE","authors_text":"Boris Murmann, Daisuke Miyashita, Edward H. Lee","submitted_at":"2016-03-03T08:51:52Z","abstract_excerpt":"Recent advances in convolutional neural networks have considered model complexity and hardware efficiency to enable deployment onto embedded systems and mobile devices. For example, it is now well-known that the arithmetic operations of deep networks can be encoded down to 8-bit fixed-point without significant deterioration in performance. However, further reduction in precision down to as low as 3-bit fixed-point results in significant losses in performance. In this paper we propose a new data representation that enables state-of-the-art networks to be encoded to 3 bits with negligible loss i"},"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":"1603.01025","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.NE","submitted_at":"2016-03-03T08:51:52Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3e3eb9cbda3b51d46822ca2277b996fad92420964f9b58938da2eba158a9315b","abstract_canon_sha256":"8d698538479543c7995e781d5dcb23eac6724349aefb91d6007208c488569197"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T01:18:56.876619Z","signature_b64":"cbRo291ng7eawSWFtJSz+bnVvEHJVhq+KsdmsvkKL6CAD+TyBuZGmseau/SyM737PGrgUh8fHTvrH23yBw1PBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7cfcd519815fe59f09d29615e72fc0f9dfd6ca6d84f3f3bc202a3c936c17a524","last_reissued_at":"2026-05-18T01:18:56.876056Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T01:18:56.876056Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Convolutional Neural Networks using Logarithmic Data Representation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.NE","authors_text":"Boris Murmann, Daisuke Miyashita, Edward H. Lee","submitted_at":"2016-03-03T08:51:52Z","abstract_excerpt":"Recent advances in convolutional neural networks have considered model complexity and hardware efficiency to enable deployment onto embedded systems and mobile devices. For example, it is now well-known that the arithmetic operations of deep networks can be encoded down to 8-bit fixed-point without significant deterioration in performance. However, further reduction in precision down to as low as 3-bit fixed-point results in significant losses in performance. In this paper we propose a new data representation that enables state-of-the-art networks to be encoded to 3 bits with negligible loss i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1603.01025","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1603.01025","created_at":"2026-05-18T01:18:56.876141+00:00"},{"alias_kind":"arxiv_version","alias_value":"1603.01025v2","created_at":"2026-05-18T01:18:56.876141+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1603.01025","created_at":"2026-05-18T01:18:56.876141+00:00"},{"alias_kind":"pith_short_12","alias_value":"PT6NKGMBL7SZ","created_at":"2026-05-18T12:30:39.010887+00:00"},{"alias_kind":"pith_short_16","alias_value":"PT6NKGMBL7SZ6COS","created_at":"2026-05-18T12:30:39.010887+00:00"},{"alias_kind":"pith_short_8","alias_value":"PT6NKGMB","created_at":"2026-05-18T12:30:39.010887+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"1907.03372","citing_title":"QUOTIENT: Two-Party Secure Neural Network Training and Prediction","ref_index":42,"is_internal_anchor":true},{"citing_arxiv_id":"1910.09876","citing_title":"Neural Network Training with Approximate Logarithmic Computations","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2410.04960","citing_title":"On Efficient Variants of Segment Anything Model: A Survey","ref_index":190,"is_internal_anchor":true},{"citing_arxiv_id":"2209.05433","citing_title":"FP8 Formats for Deep Learning","ref_index":15,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06082","citing_title":"PoTAcc: A Pipeline for End-to-End Acceleration of Power-of-Two Quantized DNNs","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H","json":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H.json","graph_json":"https://pith.science/api/pith-number/PT6NKGMBL7SZ6COSSYK6OL6A7H/graph.json","events_json":"https://pith.science/api/pith-number/PT6NKGMBL7SZ6COSSYK6OL6A7H/events.json","paper":"https://pith.science/paper/PT6NKGMB"},"agent_actions":{"view_html":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H","download_json":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H.json","view_paper":"https://pith.science/paper/PT6NKGMB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1603.01025&json=true","fetch_graph":"https://pith.science/api/pith-number/PT6NKGMBL7SZ6COSSYK6OL6A7H/graph.json","fetch_events":"https://pith.science/api/pith-number/PT6NKGMBL7SZ6COSSYK6OL6A7H/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H/action/storage_attestation","attest_author":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H/action/author_attestation","sign_citation":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H/action/citation_signature","submit_replication":"https://pith.science/pith/PT6NKGMBL7SZ6COSSYK6OL6A7H/action/replication_record"}},"created_at":"2026-05-18T01:18:56.876141+00:00","updated_at":"2026-05-18T01:18:56.876141+00:00"}