{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:WVN4EI5SO32O2DG6R4SJKZSJD2","short_pith_number":"pith:WVN4EI5S","schema_version":"1.0","canonical_sha256":"b55bc223b276f4ed0cde8f249566491e85647edf643a040525748a8537a018b8","source":{"kind":"arxiv","id":"1609.07061","version":1},"attestation_state":"computed","paper":{"title":"Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.NE","authors_text":"Daniel Soudry, Itay Hubara, Matthieu Courbariaux, Ran El-Yaniv, Yoshua Bengio","submitted_at":"2016-09-22T16:48:03Z","abstract_excerpt":"We introduce a method to train Quantized Neural Networks (QNNs) --- neural networks with extremely low precision (e.g., 1-bit) weights and activations, at run-time. At train-time the quantized weights and activations are used for computing the parameter gradients. During the forward pass, QNNs drastically reduce memory size and accesses, and replace most arithmetic operations with bit-wise operations. As a result, power consumption is expected to be drastically reduced. We trained QNNs over the MNIST, CIFAR-10, SVHN and ImageNet datasets. The resulting QNNs achieve prediction accuracy comparab"},"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":"1609.07061","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.NE","submitted_at":"2016-09-22T16:48:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"d5a809ba6c2980ac1cf7ab823db86b8fc8010fb0b838a863e12cf69aec9703c1","abstract_canon_sha256":"1785c3d9401a8a0afa3cf071e9618cf8d1967ef04138b1b080674e77310415d1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T01:04:04.533482Z","signature_b64":"53so3PfxddAvA+aeWxKTqm2Xv9I/MUUu7bDvBkY5AbmX8hIYkACt41/NSJG36mtqlDtIgljAfD9Qg+GMdIh7AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b55bc223b276f4ed0cde8f249566491e85647edf643a040525748a8537a018b8","last_reissued_at":"2026-05-18T01:04:04.532919Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T01:04:04.532919Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.NE","authors_text":"Daniel Soudry, Itay Hubara, Matthieu Courbariaux, Ran El-Yaniv, Yoshua Bengio","submitted_at":"2016-09-22T16:48:03Z","abstract_excerpt":"We introduce a method to train Quantized Neural Networks (QNNs) --- neural networks with extremely low precision (e.g., 1-bit) weights and activations, at run-time. At train-time the quantized weights and activations are used for computing the parameter gradients. During the forward pass, QNNs drastically reduce memory size and accesses, and replace most arithmetic operations with bit-wise operations. As a result, power consumption is expected to be drastically reduced. We trained QNNs over the MNIST, CIFAR-10, SVHN and ImageNet datasets. The resulting QNNs achieve prediction accuracy comparab"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1609.07061","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":""},"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":"1609.07061","created_at":"2026-05-18T01:04:04.533009+00:00"},{"alias_kind":"arxiv_version","alias_value":"1609.07061v1","created_at":"2026-05-18T01:04:04.533009+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1609.07061","created_at":"2026-05-18T01:04:04.533009+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2211.17192","citing_title":"Fast Inference from Transformers via Speculative Decoding","ref_index":53,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10886","citing_title":"LoKA: Low-precision Kernel Applications for Recommendation Models At Scale","ref_index":38,"is_internal_anchor":true},{"citing_arxiv_id":"1710.03740","citing_title":"Mixed Precision Training","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10886","citing_title":"LoKA: Low-precision Kernel Applications for Recommendation Models At Scale","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"1704.04861","citing_title":"MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2","json":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2.json","graph_json":"https://pith.science/api/pith-number/WVN4EI5SO32O2DG6R4SJKZSJD2/graph.json","events_json":"https://pith.science/api/pith-number/WVN4EI5SO32O2DG6R4SJKZSJD2/events.json","paper":"https://pith.science/paper/WVN4EI5S"},"agent_actions":{"view_html":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2","download_json":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2.json","view_paper":"https://pith.science/paper/WVN4EI5S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1609.07061&json=true","fetch_graph":"https://pith.science/api/pith-number/WVN4EI5SO32O2DG6R4SJKZSJD2/graph.json","fetch_events":"https://pith.science/api/pith-number/WVN4EI5SO32O2DG6R4SJKZSJD2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2/action/storage_attestation","attest_author":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2/action/author_attestation","sign_citation":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2/action/citation_signature","submit_replication":"https://pith.science/pith/WVN4EI5SO32O2DG6R4SJKZSJD2/action/replication_record"}},"created_at":"2026-05-18T01:04:04.533009+00:00","updated_at":"2026-05-18T01:04:04.533009+00:00"}