{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:SQQ5L33NF4PZFFHW4ZEXWUWEEV","short_pith_number":"pith:SQQ5L33N","canonical_record":{"source":{"id":"2306.03076","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-05T17:52:44Z","cross_cats_sorted":["cs.AR"],"title_canon_sha256":"2acb796ecda59ab0e1190306fc719187fa1aad77c9c2bb8a93839d706495df54","abstract_canon_sha256":"8bee8b58ab401a5ea87e427495c58e79a445c93d2cfe3cc994e0a0b67b99cb8d"},"schema_version":"1.0"},"canonical_sha256":"9421d5ef6d2f1f9294f6e6497b52c425471dad5449f88c2465422bae612c94cc","source":{"kind":"arxiv","id":"2306.03076","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2306.03076","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"arxiv_version","alias_value":"2306.03076v1","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.03076","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"pith_short_12","alias_value":"SQQ5L33NF4PZ","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"pith_short_16","alias_value":"SQQ5L33NF4PZFFHW","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"pith_short_8","alias_value":"SQQ5L33N","created_at":"2026-07-05T06:17:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:SQQ5L33NF4PZFFHW4ZEXWUWEEV","target":"record","payload":{"canonical_record":{"source":{"id":"2306.03076","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-05T17:52:44Z","cross_cats_sorted":["cs.AR"],"title_canon_sha256":"2acb796ecda59ab0e1190306fc719187fa1aad77c9c2bb8a93839d706495df54","abstract_canon_sha256":"8bee8b58ab401a5ea87e427495c58e79a445c93d2cfe3cc994e0a0b67b99cb8d"},"schema_version":"1.0"},"canonical_sha256":"9421d5ef6d2f1f9294f6e6497b52c425471dad5449f88c2465422bae612c94cc","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:17:34.589961Z","signature_b64":"nfjqJne5GKGTtIJpB0Km1cwhs5+4WqAulWpxsdBjuw5Y456JtE3jt5cClvloMX5EOtz76aFcF4p1Zhlp4SGqDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9421d5ef6d2f1f9294f6e6497b52c425471dad5449f88c2465422bae612c94cc","last_reissued_at":"2026-07-05T06:17:34.589463Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:17:34.589463Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2306.03076","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T06:17:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"q58lYO0GWfZ3DkMk9Bj/jI/d93JuQVG8wFD0/FSp4ZVbVh8pY+1EQQBWw+vuL/4MY/m2fteIYr6kKPjXMgJIAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T16:55:08.717893Z"},"content_sha256":"3e5d524e474fd9967a6715402f733c313c16121510fea174710085e8b2b54f2c","schema_version":"1.0","event_id":"sha256:3e5d524e474fd9967a6715402f733c313c16121510fea174710085e8b2b54f2c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:SQQ5L33NF4PZFFHW4ZEXWUWEEV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Sensitivity-Aware Finetuning for Accuracy Recovery on Deep Learning Hardware","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AR"],"primary_cat":"cs.LG","authors_text":"Darius Bunandar, Lakshmi Nair","submitted_at":"2023-06-05T17:52:44Z","abstract_excerpt":"Existing methods to recover model accuracy on analog-digital hardware in the presence of quantization and analog noise include noise-injection training. However, it can be slow in practice, incurring high computational costs, even when starting from pretrained models. We introduce the Sensitivity-Aware Finetuning (SAFT) approach that identifies noise sensitive layers in a model, and uses the information to freeze specific layers for noise-injection training. Our results show that SAFT achieves comparable accuracy to noise-injection training and is 2x to 8x faster."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.03076","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2306.03076/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T06:17:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2lQLfodLG4xllMXZHd+6DLKN9kIhpDcVt96vdfep3aMX0HjNLa8hpTjfbQeSlhl8U7XJUtyCBtn3laeSjJU2BA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T16:55:08.718276Z"},"content_sha256":"c75b176d1be8fdefcb986c07ab6556e6cb0370ff93c0f67634c629d6352985b7","schema_version":"1.0","event_id":"sha256:c75b176d1be8fdefcb986c07ab6556e6cb0370ff93c0f67634c629d6352985b7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV/bundle.json","state_url":"https://pith.science/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV/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":"2026-08-03T16:55:08Z","links":{"resolver":"https://pith.science/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV","bundle":"https://pith.science/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV/bundle.json","state":"https://pith.science/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SQQ5L33NF4PZFFHW4ZEXWUWEEV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:SQQ5L33NF4PZFFHW4ZEXWUWEEV","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":"8bee8b58ab401a5ea87e427495c58e79a445c93d2cfe3cc994e0a0b67b99cb8d","cross_cats_sorted":["cs.AR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-05T17:52:44Z","title_canon_sha256":"2acb796ecda59ab0e1190306fc719187fa1aad77c9c2bb8a93839d706495df54"},"schema_version":"1.0","source":{"id":"2306.03076","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2306.03076","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"arxiv_version","alias_value":"2306.03076v1","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2306.03076","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"pith_short_12","alias_value":"SQQ5L33NF4PZ","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"pith_short_16","alias_value":"SQQ5L33NF4PZFFHW","created_at":"2026-07-05T06:17:34Z"},{"alias_kind":"pith_short_8","alias_value":"SQQ5L33N","created_at":"2026-07-05T06:17:34Z"}],"graph_snapshots":[{"event_id":"sha256:c75b176d1be8fdefcb986c07ab6556e6cb0370ff93c0f67634c629d6352985b7","target":"graph","created_at":"2026-07-05T06:17:34Z","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"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2306.03076/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Existing methods to recover model accuracy on analog-digital hardware in the presence of quantization and analog noise include noise-injection training. However, it can be slow in practice, incurring high computational costs, even when starting from pretrained models. We introduce the Sensitivity-Aware Finetuning (SAFT) approach that identifies noise sensitive layers in a model, and uses the information to freeze specific layers for noise-injection training. Our results show that SAFT achieves comparable accuracy to noise-injection training and is 2x to 8x faster.","authors_text":"Darius Bunandar, Lakshmi Nair","cross_cats":["cs.AR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-05T17:52:44Z","title":"Sensitivity-Aware Finetuning for Accuracy Recovery on Deep Learning Hardware"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2306.03076","kind":"arxiv","version":1},"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":"sha256:3e5d524e474fd9967a6715402f733c313c16121510fea174710085e8b2b54f2c","target":"record","created_at":"2026-07-05T06:17:34Z","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":"8bee8b58ab401a5ea87e427495c58e79a445c93d2cfe3cc994e0a0b67b99cb8d","cross_cats_sorted":["cs.AR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-06-05T17:52:44Z","title_canon_sha256":"2acb796ecda59ab0e1190306fc719187fa1aad77c9c2bb8a93839d706495df54"},"schema_version":"1.0","source":{"id":"2306.03076","kind":"arxiv","version":1}},"canonical_sha256":"9421d5ef6d2f1f9294f6e6497b52c425471dad5449f88c2465422bae612c94cc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9421d5ef6d2f1f9294f6e6497b52c425471dad5449f88c2465422bae612c94cc","first_computed_at":"2026-07-05T06:17:34.589463Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:17:34.589463Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nfjqJne5GKGTtIJpB0Km1cwhs5+4WqAulWpxsdBjuw5Y456JtE3jt5cClvloMX5EOtz76aFcF4p1Zhlp4SGqDA==","signature_status":"signed_v1","signed_at":"2026-07-05T06:17:34.589961Z","signed_message":"canonical_sha256_bytes"},"source_id":"2306.03076","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3e5d524e474fd9967a6715402f733c313c16121510fea174710085e8b2b54f2c","sha256:c75b176d1be8fdefcb986c07ab6556e6cb0370ff93c0f67634c629d6352985b7"],"state_sha256":"f108642613c6eda31a8ff2ae8b4bb3dd8aac656f03a1abaae375d64fcc8a4091"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qJyGZnxtOCQWp0TrA5NKM2kSv6YJ6K9+GAYPsmd7JGPk0GwXnkh+c6NPfem0eO5a7WkhQN6gK28oDTytwMXiBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T16:55:08.720983Z","bundle_sha256":"b0ad44f9e77566a7b3ed5cdceb8f12694dcf4401b7cce17c17642346193d55f0"}}