{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2017:BWYZ5PAJSARALRMON6YFVTQVFD","short_pith_number":"pith:BWYZ5PAJ","canonical_record":{"source":{"id":"1707.05612","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-07-18T13:51:32Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"f6aafd43e8d5afb01ef53afc75c809df67dd98c8f119ec9bb3133aea7fa087da","abstract_canon_sha256":"9c5c7c56ce1b9b3416611767ded9ac8b8bcea192d6b3f87fc6f3af7e3b75d59d"},"schema_version":"1.0"},"canonical_sha256":"0db19ebc09902205c58e6fb05ace1528e458bac2c4ee722a55e0eadb6e4143e3","source":{"kind":"arxiv","id":"1707.05612","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1707.05612","created_at":"2026-05-18T00:09:36Z"},{"alias_kind":"arxiv_version","alias_value":"1707.05612v4","created_at":"2026-05-18T00:09:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1707.05612","created_at":"2026-05-18T00:09:36Z"},{"alias_kind":"pith_short_12","alias_value":"BWYZ5PAJSARA","created_at":"2026-05-18T12:31:08Z"},{"alias_kind":"pith_short_16","alias_value":"BWYZ5PAJSARALRMO","created_at":"2026-05-18T12:31:08Z"},{"alias_kind":"pith_short_8","alias_value":"BWYZ5PAJ","created_at":"2026-05-18T12:31:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2017:BWYZ5PAJSARALRMON6YFVTQVFD","target":"record","payload":{"canonical_record":{"source":{"id":"1707.05612","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-07-18T13:51:32Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"f6aafd43e8d5afb01ef53afc75c809df67dd98c8f119ec9bb3133aea7fa087da","abstract_canon_sha256":"9c5c7c56ce1b9b3416611767ded9ac8b8bcea192d6b3f87fc6f3af7e3b75d59d"},"schema_version":"1.0"},"canonical_sha256":"0db19ebc09902205c58e6fb05ace1528e458bac2c4ee722a55e0eadb6e4143e3","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:09:36.556546Z","signature_b64":"sGbUbpplgZnkwHisfhB8r06h5thRZxl2WXZhvqRtE8AQOgtmjnJnFVUNQNrf4ROflMDrqcfY2wnGYbvGS/5VAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0db19ebc09902205c58e6fb05ace1528e458bac2c4ee722a55e0eadb6e4143e3","last_reissued_at":"2026-05-18T00:09:36.556030Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:09:36.556030Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1707.05612","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":"2026-05-18T00:09:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cLyemmZxoSW3e7q7lved96h46yV1AXk67HH9sGz3+Ihpmkkm+lar6U3XDy42YqVpzsrYpoYmJgtiGZ6mNCkfBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T15:16:52.219839Z"},"content_sha256":"9540e8b1811a1093d22b2381d7d0d72160ec4609b42fb81d5480b0e3e5f305b3","schema_version":"1.0","event_id":"sha256:9540e8b1811a1093d22b2381d7d0d72160ec4609b42fb81d5480b0e3e5f305b3"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2017:BWYZ5PAJSARALRMON6YFVTQVFD","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"VSE++: Improving Visual-Semantic Embeddings with Hard Negatives","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"David J. Fleet, Fartash Faghri, Jamie Ryan Kiros, Sanja Fidler","submitted_at":"2017-07-18T13:51:32Z","abstract_excerpt":"We present a new technique for learning visual-semantic embeddings for cross-modal retrieval. Inspired by hard negative mining, the use of hard negatives in structured prediction, and ranking loss functions, we introduce a simple change to common loss functions used for multi-modal embeddings. That, combined with fine-tuning and use of augmented data, yields significant gains in retrieval performance. We showcase our approach, VSE++, on MS-COCO and Flickr30K datasets, using ablation studies and comparisons with existing methods. On MS-COCO our approach outperforms state-of-the-art methods by 8"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1707.05612","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":"2026-05-18T00:09:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Zb8ofnNdRw21PEe8LloAs18dAAo8Q3vzTvDNkrrB3IDQmnNHy7ePm6eapu1cUNTrhQY1A7aRR5j1yozCKNO8DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T15:16:52.220249Z"},"content_sha256":"d1ddf0678e1625586f42d61304cc2ea8e5205fb787cb6aff7942e3027081ef4c","schema_version":"1.0","event_id":"sha256:d1ddf0678e1625586f42d61304cc2ea8e5205fb787cb6aff7942e3027081ef4c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BWYZ5PAJSARALRMON6YFVTQVFD/bundle.json","state_url":"https://pith.science/pith/BWYZ5PAJSARALRMON6YFVTQVFD/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BWYZ5PAJSARALRMON6YFVTQVFD/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":"2026-05-26T15:16:52Z","links":{"resolver":"https://pith.science/pith/BWYZ5PAJSARALRMON6YFVTQVFD","bundle":"https://pith.science/pith/BWYZ5PAJSARALRMON6YFVTQVFD/bundle.json","state":"https://pith.science/pith/BWYZ5PAJSARALRMON6YFVTQVFD/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BWYZ5PAJSARALRMON6YFVTQVFD/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2017:BWYZ5PAJSARALRMON6YFVTQVFD","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":"9c5c7c56ce1b9b3416611767ded9ac8b8bcea192d6b3f87fc6f3af7e3b75d59d","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-07-18T13:51:32Z","title_canon_sha256":"f6aafd43e8d5afb01ef53afc75c809df67dd98c8f119ec9bb3133aea7fa087da"},"schema_version":"1.0","source":{"id":"1707.05612","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1707.05612","created_at":"2026-05-18T00:09:36Z"},{"alias_kind":"arxiv_version","alias_value":"1707.05612v4","created_at":"2026-05-18T00:09:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1707.05612","created_at":"2026-05-18T00:09:36Z"},{"alias_kind":"pith_short_12","alias_value":"BWYZ5PAJSARA","created_at":"2026-05-18T12:31:08Z"},{"alias_kind":"pith_short_16","alias_value":"BWYZ5PAJSARALRMO","created_at":"2026-05-18T12:31:08Z"},{"alias_kind":"pith_short_8","alias_value":"BWYZ5PAJ","created_at":"2026-05-18T12:31:08Z"}],"graph_snapshots":[{"event_id":"sha256:d1ddf0678e1625586f42d61304cc2ea8e5205fb787cb6aff7942e3027081ef4c","target":"graph","created_at":"2026-05-18T00:09:36Z","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":"We present a new technique for learning visual-semantic embeddings for cross-modal retrieval. Inspired by hard negative mining, the use of hard negatives in structured prediction, and ranking loss functions, we introduce a simple change to common loss functions used for multi-modal embeddings. That, combined with fine-tuning and use of augmented data, yields significant gains in retrieval performance. We showcase our approach, VSE++, on MS-COCO and Flickr30K datasets, using ablation studies and comparisons with existing methods. On MS-COCO our approach outperforms state-of-the-art methods by 8","authors_text":"David J. Fleet, Fartash Faghri, Jamie Ryan Kiros, Sanja Fidler","cross_cats":["cs.CL","cs.CV"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-07-18T13:51:32Z","title":"VSE++: Improving Visual-Semantic Embeddings with Hard Negatives"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1707.05612","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":"sha256:9540e8b1811a1093d22b2381d7d0d72160ec4609b42fb81d5480b0e3e5f305b3","target":"record","created_at":"2026-05-18T00:09:36Z","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":"9c5c7c56ce1b9b3416611767ded9ac8b8bcea192d6b3f87fc6f3af7e3b75d59d","cross_cats_sorted":["cs.CL","cs.CV"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-07-18T13:51:32Z","title_canon_sha256":"f6aafd43e8d5afb01ef53afc75c809df67dd98c8f119ec9bb3133aea7fa087da"},"schema_version":"1.0","source":{"id":"1707.05612","kind":"arxiv","version":4}},"canonical_sha256":"0db19ebc09902205c58e6fb05ace1528e458bac2c4ee722a55e0eadb6e4143e3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0db19ebc09902205c58e6fb05ace1528e458bac2c4ee722a55e0eadb6e4143e3","first_computed_at":"2026-05-18T00:09:36.556030Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:09:36.556030Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"sGbUbpplgZnkwHisfhB8r06h5thRZxl2WXZhvqRtE8AQOgtmjnJnFVUNQNrf4ROflMDrqcfY2wnGYbvGS/5VAQ==","signature_status":"signed_v1","signed_at":"2026-05-18T00:09:36.556546Z","signed_message":"canonical_sha256_bytes"},"source_id":"1707.05612","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9540e8b1811a1093d22b2381d7d0d72160ec4609b42fb81d5480b0e3e5f305b3","sha256:d1ddf0678e1625586f42d61304cc2ea8e5205fb787cb6aff7942e3027081ef4c"],"state_sha256":"5320a05a70b8675cb4763708ed98b3c91001cf716c7d5816f4a448a995be9f2c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"P4dNr6b/0M/fv8lXRiC1lkxQn1xtWO+Bmi6iF+NGj3btT1BNaDwdrvSZja3gaRtG3AdSYDbbNgvZZMvUIcZQCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T15:16:52.222842Z","bundle_sha256":"14fff5642db5531db5a6a988643aabf6cbedf398a655dc72f76255de7b9d7126"}}