{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:U3R6IKNY24TAQPMQ6YAPUOWM36","short_pith_number":"pith:U3R6IKNY","canonical_record":{"source":{"id":"1808.05240","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-08-15T18:13:12Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"26f0c9e58765854c1f987ec944530b96e49336663c7a93e3faf7ffeee3a1a170","abstract_canon_sha256":"8d36e63dfc7a9137c9f242a377f7234e90607f0fd63c0855e2468cda7c00c56b"},"schema_version":"1.0"},"canonical_sha256":"a6e3e429b8d726083d90f600fa3accdf9aa2d815d03e3725185fb5a820cb897b","source":{"kind":"arxiv","id":"1808.05240","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1808.05240","created_at":"2026-05-17T23:56:53Z"},{"alias_kind":"arxiv_version","alias_value":"1808.05240v4","created_at":"2026-05-17T23:56:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1808.05240","created_at":"2026-05-17T23:56:53Z"},{"alias_kind":"pith_short_12","alias_value":"U3R6IKNY24TA","created_at":"2026-05-18T12:32:56Z"},{"alias_kind":"pith_short_16","alias_value":"U3R6IKNY24TAQPMQ","created_at":"2026-05-18T12:32:56Z"},{"alias_kind":"pith_short_8","alias_value":"U3R6IKNY","created_at":"2026-05-18T12:32:56Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:U3R6IKNY24TAQPMQ6YAPUOWM36","target":"record","payload":{"canonical_record":{"source":{"id":"1808.05240","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-08-15T18:13:12Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"26f0c9e58765854c1f987ec944530b96e49336663c7a93e3faf7ffeee3a1a170","abstract_canon_sha256":"8d36e63dfc7a9137c9f242a377f7234e90607f0fd63c0855e2468cda7c00c56b"},"schema_version":"1.0"},"canonical_sha256":"a6e3e429b8d726083d90f600fa3accdf9aa2d815d03e3725185fb5a820cb897b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:56:53.037841Z","signature_b64":"roVwO9cRj3d63vLXRATl5mqPsUmoKzRfeNdtDwttg2D7ayXa1NnqEYRgnI9e82OfVG/c/N3GHgBt1UnqcGlfAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a6e3e429b8d726083d90f600fa3accdf9aa2d815d03e3725185fb5a820cb897b","last_reissued_at":"2026-05-17T23:56:53.037473Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:56:53.037473Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1808.05240","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:56:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PJmoIX1LHPZM7UirjARR/ULlfYEZYFvStOqUqFHhOVdfPKXgPj2qFK7fFEvDIJejKVJPInG0QDbfCabXSM6MBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T13:04:19.070009Z"},"content_sha256":"2238b461887b5d04487f0f2ba0b191f9a1c35ee129953eee09cdc0315010667b","schema_version":"1.0","event_id":"sha256:2238b461887b5d04487f0f2ba0b191f9a1c35ee129953eee09cdc0315010667b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:U3R6IKNY24TAQPMQ6YAPUOWM36","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Blended Coarse Gradient Descent for Full Quantization of Deep Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Jack Xin, Jiancheng Lyu, Penghang Yin, Shuai Zhang, Stanley Osher, Yingyong Qi","submitted_at":"2018-08-15T18:13:12Z","abstract_excerpt":"Quantized deep neural networks (QDNNs) are attractive due to their much lower memory storage and faster inference speed than their regular full precision counterparts. To maintain the same performance level especially at low bit-widths, QDNNs must be retrained. Their training involves piecewise constant activation functions and discrete weights, hence mathematical challenges arise. We introduce the notion of coarse gradient and propose the blended coarse gradient descent (BCGD) algorithm, for training fully quantized neural networks. Coarse gradient is generally not a gradient of any function "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1808.05240","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":""},"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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:56:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"k5zB2Qt310VZns/WWMzHYyouoW4XASoCtz4OZ2tFmT1ylYgU60Sa/Gn3z8Um4NoObl4FKSncaVK5WdSyJRHRBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T13:04:19.070369Z"},"content_sha256":"732c67d4ef7c85e8a12c2bc16461fa1eb14ee49a21ffc1bfff947b71a1659475","schema_version":"1.0","event_id":"sha256:732c67d4ef7c85e8a12c2bc16461fa1eb14ee49a21ffc1bfff947b71a1659475"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/U3R6IKNY24TAQPMQ6YAPUOWM36/bundle.json","state_url":"https://pith.science/pith/U3R6IKNY24TAQPMQ6YAPUOWM36/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/U3R6IKNY24TAQPMQ6YAPUOWM36/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-04T13:04:19Z","links":{"resolver":"https://pith.science/pith/U3R6IKNY24TAQPMQ6YAPUOWM36","bundle":"https://pith.science/pith/U3R6IKNY24TAQPMQ6YAPUOWM36/bundle.json","state":"https://pith.science/pith/U3R6IKNY24TAQPMQ6YAPUOWM36/state.json","well_known_bundle":"https://pith.science/.well-known/pith/U3R6IKNY24TAQPMQ6YAPUOWM36/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:U3R6IKNY24TAQPMQ6YAPUOWM36","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"8d36e63dfc7a9137c9f242a377f7234e90607f0fd63c0855e2468cda7c00c56b","cross_cats_sorted":["cs.CV","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-08-15T18:13:12Z","title_canon_sha256":"26f0c9e58765854c1f987ec944530b96e49336663c7a93e3faf7ffeee3a1a170"},"schema_version":"1.0","source":{"id":"1808.05240","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1808.05240","created_at":"2026-05-17T23:56:53Z"},{"alias_kind":"arxiv_version","alias_value":"1808.05240v4","created_at":"2026-05-17T23:56:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1808.05240","created_at":"2026-05-17T23:56:53Z"},{"alias_kind":"pith_short_12","alias_value":"U3R6IKNY24TA","created_at":"2026-05-18T12:32:56Z"},{"alias_kind":"pith_short_16","alias_value":"U3R6IKNY24TAQPMQ","created_at":"2026-05-18T12:32:56Z"},{"alias_kind":"pith_short_8","alias_value":"U3R6IKNY","created_at":"2026-05-18T12:32:56Z"}],"graph_snapshots":[{"event_id":"sha256:732c67d4ef7c85e8a12c2bc16461fa1eb14ee49a21ffc1bfff947b71a1659475","target":"graph","created_at":"2026-05-17T23:56:53Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Quantized deep neural networks (QDNNs) are attractive due to their much lower memory storage and faster inference speed than their regular full precision counterparts. To maintain the same performance level especially at low bit-widths, QDNNs must be retrained. Their training involves piecewise constant activation functions and discrete weights, hence mathematical challenges arise. We introduce the notion of coarse gradient and propose the blended coarse gradient descent (BCGD) algorithm, for training fully quantized neural networks. Coarse gradient is generally not a gradient of any function ","authors_text":"Jack Xin, Jiancheng Lyu, Penghang Yin, Shuai Zhang, Stanley Osher, Yingyong Qi","cross_cats":["cs.CV","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-08-15T18:13:12Z","title":"Blended Coarse Gradient Descent for Full Quantization of Deep Neural Networks"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1808.05240","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:2238b461887b5d04487f0f2ba0b191f9a1c35ee129953eee09cdc0315010667b","target":"record","created_at":"2026-05-17T23:56:53Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"8d36e63dfc7a9137c9f242a377f7234e90607f0fd63c0855e2468cda7c00c56b","cross_cats_sorted":["cs.CV","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-08-15T18:13:12Z","title_canon_sha256":"26f0c9e58765854c1f987ec944530b96e49336663c7a93e3faf7ffeee3a1a170"},"schema_version":"1.0","source":{"id":"1808.05240","kind":"arxiv","version":4}},"canonical_sha256":"a6e3e429b8d726083d90f600fa3accdf9aa2d815d03e3725185fb5a820cb897b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a6e3e429b8d726083d90f600fa3accdf9aa2d815d03e3725185fb5a820cb897b","first_computed_at":"2026-05-17T23:56:53.037473Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:56:53.037473Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"roVwO9cRj3d63vLXRATl5mqPsUmoKzRfeNdtDwttg2D7ayXa1NnqEYRgnI9e82OfVG/c/N3GHgBt1UnqcGlfAw==","signature_status":"signed_v1","signed_at":"2026-05-17T23:56:53.037841Z","signed_message":"canonical_sha256_bytes"},"source_id":"1808.05240","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2238b461887b5d04487f0f2ba0b191f9a1c35ee129953eee09cdc0315010667b","sha256:732c67d4ef7c85e8a12c2bc16461fa1eb14ee49a21ffc1bfff947b71a1659475"],"state_sha256":"7421be889e4e7e1338d19b637444556aced63a2717cc0a35cda3f5129f4ae052"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"95A8vZYr6ZR3Gf2BqTN6QY8jcxeezG4r6yYCGunGjoOTckUnfNvCPWHSg2F0RCLOaPkjNzfGpMDMoRPIutMkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-04T13:04:19.072441Z","bundle_sha256":"e9f543e58f3ffd8c053db416c3e05ff8114f7a7b0ba5044d6f89b7c281d83692"}}