{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:GKR777IYPSZ65FHN5JBSCENPII","short_pith_number":"pith:GKR777IY","canonical_record":{"source":{"id":"2110.11773","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-22T13:25:01Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"b8b7a9e580ffa5e27c17e48ed55be65981b11ffe8557548d763ec28590358fe3","abstract_canon_sha256":"694407e50dc15c614497f77dc3ebf8793ec44f618a9114995572254761bb7b15"},"schema_version":"1.0"},"canonical_sha256":"32a3fffd187cb3ee94edea432111af420e5a8522f0ecefe3b2baf1e84a9fb543","source":{"kind":"arxiv","id":"2110.11773","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2110.11773","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"arxiv_version","alias_value":"2110.11773v2","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.11773","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"pith_short_12","alias_value":"GKR777IYPSZ6","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"pith_short_16","alias_value":"GKR777IYPSZ65FHN","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"pith_short_8","alias_value":"GKR777IY","created_at":"2026-07-05T03:50:52Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:GKR777IYPSZ65FHN5JBSCENPII","target":"record","payload":{"canonical_record":{"source":{"id":"2110.11773","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-22T13:25:01Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"b8b7a9e580ffa5e27c17e48ed55be65981b11ffe8557548d763ec28590358fe3","abstract_canon_sha256":"694407e50dc15c614497f77dc3ebf8793ec44f618a9114995572254761bb7b15"},"schema_version":"1.0"},"canonical_sha256":"32a3fffd187cb3ee94edea432111af420e5a8522f0ecefe3b2baf1e84a9fb543","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:50:52.304832Z","signature_b64":"0jo2StQwE4xF6P8NNPEE6pieWxvqkZBa+750j4fwYAEpacg1P84WVAzJ4nNS/6dc861BYpRiExawpVut9y5xAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"32a3fffd187cb3ee94edea432111af420e5a8522f0ecefe3b2baf1e84a9fb543","last_reissued_at":"2026-07-05T03:50:52.304346Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:50:52.304346Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2110.11773","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T03:50:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gtkSxt2EaOyBZJuQjTy8CYxKbx9qLbpaDsvbQ8mi7ryLHJn9uHe1RDA24ix/uNJoU3QN08SrpOvfPso+JcuEBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T20:47:06.102247Z"},"content_sha256":"f8cf1db99627447178fb619c0e0e4ee57bf98127b58fd45c8e44e018ad062449","schema_version":"1.0","event_id":"sha256:f8cf1db99627447178fb619c0e0e4ee57bf98127b58fd45c8e44e018ad062449"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:GKR777IYPSZ65FHN5JBSCENPII","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Sinkformers: Transformers with Doubly Stochastic Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Gabriel Peyr\\'e, Mathieu Blondel, Michael E. Sander, Pierre Ablin","submitted_at":"2021-10-22T13:25:01Z","abstract_excerpt":"Attention based models such as Transformers involve pairwise interactions between data points, modeled with a learnable attention matrix. Importantly, this attention matrix is normalized with the SoftMax operator, which makes it row-wise stochastic. In this paper, we propose instead to use Sinkhorn's algorithm to make attention matrices doubly stochastic. We call the resulting model a Sinkformer. We show that the row-wise stochastic attention matrices in classical Transformers get close to doubly stochastic matrices as the number of epochs increases, justifying the use of Sinkhorn normalizatio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.11773","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2110.11773/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":"2026-07-05T03:50:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8cicSPH2RfAh8qzePqAfljdTvH0PcmJtYoNIc64wgKSo0MQSZZjAG0UR2LyeqxgX6+Kia6nKs+Eq5W7PRTEAAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T20:47:06.103197Z"},"content_sha256":"024670a3aab1b4cbb6b73a200ca4730fadd4cb3287d75d9359487ee38fc4fa50","schema_version":"1.0","event_id":"sha256:024670a3aab1b4cbb6b73a200ca4730fadd4cb3287d75d9359487ee38fc4fa50"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GKR777IYPSZ65FHN5JBSCENPII/bundle.json","state_url":"https://pith.science/pith/GKR777IYPSZ65FHN5JBSCENPII/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GKR777IYPSZ65FHN5JBSCENPII/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":"2026-08-15T20:47:06Z","links":{"resolver":"https://pith.science/pith/GKR777IYPSZ65FHN5JBSCENPII","bundle":"https://pith.science/pith/GKR777IYPSZ65FHN5JBSCENPII/bundle.json","state":"https://pith.science/pith/GKR777IYPSZ65FHN5JBSCENPII/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GKR777IYPSZ65FHN5JBSCENPII/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:GKR777IYPSZ65FHN5JBSCENPII","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":"694407e50dc15c614497f77dc3ebf8793ec44f618a9114995572254761bb7b15","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-22T13:25:01Z","title_canon_sha256":"b8b7a9e580ffa5e27c17e48ed55be65981b11ffe8557548d763ec28590358fe3"},"schema_version":"1.0","source":{"id":"2110.11773","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2110.11773","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"arxiv_version","alias_value":"2110.11773v2","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.11773","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"pith_short_12","alias_value":"GKR777IYPSZ6","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"pith_short_16","alias_value":"GKR777IYPSZ65FHN","created_at":"2026-07-05T03:50:52Z"},{"alias_kind":"pith_short_8","alias_value":"GKR777IY","created_at":"2026-07-05T03:50:52Z"}],"graph_snapshots":[{"event_id":"sha256:024670a3aab1b4cbb6b73a200ca4730fadd4cb3287d75d9359487ee38fc4fa50","target":"graph","created_at":"2026-07-05T03:50:52Z","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/2110.11773/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Attention based models such as Transformers involve pairwise interactions between data points, modeled with a learnable attention matrix. Importantly, this attention matrix is normalized with the SoftMax operator, which makes it row-wise stochastic. In this paper, we propose instead to use Sinkhorn's algorithm to make attention matrices doubly stochastic. We call the resulting model a Sinkformer. We show that the row-wise stochastic attention matrices in classical Transformers get close to doubly stochastic matrices as the number of epochs increases, justifying the use of Sinkhorn normalizatio","authors_text":"Gabriel Peyr\\'e, Mathieu Blondel, Michael E. Sander, Pierre Ablin","cross_cats":["stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-22T13:25:01Z","title":"Sinkformers: Transformers with Doubly Stochastic Attention"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.11773","kind":"arxiv","version":2},"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:f8cf1db99627447178fb619c0e0e4ee57bf98127b58fd45c8e44e018ad062449","target":"record","created_at":"2026-07-05T03:50:52Z","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":"694407e50dc15c614497f77dc3ebf8793ec44f618a9114995572254761bb7b15","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-22T13:25:01Z","title_canon_sha256":"b8b7a9e580ffa5e27c17e48ed55be65981b11ffe8557548d763ec28590358fe3"},"schema_version":"1.0","source":{"id":"2110.11773","kind":"arxiv","version":2}},"canonical_sha256":"32a3fffd187cb3ee94edea432111af420e5a8522f0ecefe3b2baf1e84a9fb543","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"32a3fffd187cb3ee94edea432111af420e5a8522f0ecefe3b2baf1e84a9fb543","first_computed_at":"2026-07-05T03:50:52.304346Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T03:50:52.304346Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"0jo2StQwE4xF6P8NNPEE6pieWxvqkZBa+750j4fwYAEpacg1P84WVAzJ4nNS/6dc861BYpRiExawpVut9y5xAA==","signature_status":"signed_v1","signed_at":"2026-07-05T03:50:52.304832Z","signed_message":"canonical_sha256_bytes"},"source_id":"2110.11773","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f8cf1db99627447178fb619c0e0e4ee57bf98127b58fd45c8e44e018ad062449","sha256:024670a3aab1b4cbb6b73a200ca4730fadd4cb3287d75d9359487ee38fc4fa50"],"state_sha256":"9f09b16fcf9a1c26f714030b93a2120494367b8bdc8c14f98046d9d3c7dd6a23"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"giODqdgDYi26CWb74QRiObZw7Furb5KLlRFnXqiyNLYmzRStOOqokLnuT+mR9zsFLxCH7lBgb2UJ1EIskpWbDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-15T20:47:06.111325Z","bundle_sha256":"dce05ab49e996e1b06831cb76e28a4d4c59f55b0233d7fb89edbca6799f0301d"}}