{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:GQPRER7NMOKNVWIT6L5UDA6CR7","short_pith_number":"pith:GQPRER7N","schema_version":"1.0","canonical_sha256":"341f1247ed6394dad913f2fb4183c28fee49610b4954541f400f09f08261115f","source":{"kind":"arxiv","id":"1912.02761","version":2},"attestation_state":"computed","paper":{"title":"Measuring Social Bias in Knowledge Graph Embeddings","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Arpit Mittal, Christos Christodoulopoulos, Dave Palfrey, Joseph Fisher","submitted_at":"2019-12-05T17:53:25Z","abstract_excerpt":"It has recently been shown that word embeddings encode social biases, with a harmful impact on downstream tasks. However, to this point there has been no similar work done in the field of graph embeddings. We present the first study on social bias in knowledge graph embeddings, and propose a new metric suitable for measuring such bias. We conduct experiments on Wikidata and Freebase, and show that, as with word embeddings, harmful social biases related to professions are encoded in the embeddings with respect to gender, religion, ethnicity and nationality. For example, graph embeddings encode "},"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":"1912.02761","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-12-05T17:53:25Z","cross_cats_sorted":[],"title_canon_sha256":"62d4ca9367e6d4b7760b8d0064715da426f6af7d35b5118626df88031a1e5f6e","abstract_canon_sha256":"0fc7a022d7e51a2ffeb40bf9dbf6ed8d982ee6462c7df8f0ce1e349043b8cfbc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:01:01.520903Z","signature_b64":"RX7gZd/W7+QdH1CySC7LysX15vY83/Ol5WSWF4ufcczrZGXaJVSgwL65/VT1XOKeoD9owI1fF7+6KAe1IPggCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"341f1247ed6394dad913f2fb4183c28fee49610b4954541f400f09f08261115f","last_reissued_at":"2026-07-05T01:01:01.520481Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:01:01.520481Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Measuring Social Bias in Knowledge Graph Embeddings","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Arpit Mittal, Christos Christodoulopoulos, Dave Palfrey, Joseph Fisher","submitted_at":"2019-12-05T17:53:25Z","abstract_excerpt":"It has recently been shown that word embeddings encode social biases, with a harmful impact on downstream tasks. However, to this point there has been no similar work done in the field of graph embeddings. We present the first study on social bias in knowledge graph embeddings, and propose a new metric suitable for measuring such bias. We conduct experiments on Wikidata and Freebase, and show that, as with word embeddings, harmful social biases related to professions are encoded in the embeddings with respect to gender, religion, ethnicity and nationality. For example, graph embeddings encode "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1912.02761","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/1912.02761/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"1912.02761","created_at":"2026-07-05T01:01:01.520542+00:00"},{"alias_kind":"arxiv_version","alias_value":"1912.02761v2","created_at":"2026-07-05T01:01:01.520542+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1912.02761","created_at":"2026-07-05T01:01:01.520542+00:00"},{"alias_kind":"pith_short_12","alias_value":"GQPRER7NMOKN","created_at":"2026-07-05T01:01:01.520542+00:00"},{"alias_kind":"pith_short_16","alias_value":"GQPRER7NMOKNVWIT","created_at":"2026-07-05T01:01:01.520542+00:00"},{"alias_kind":"pith_short_8","alias_value":"GQPRER7N","created_at":"2026-07-05T01:01:01.520542+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.16383","citing_title":"Filling in the Blanks? A Systematic Review and Theoretical Conceptualisation for Measuring WikiData Content Gaps","ref_index":24,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7","json":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7.json","graph_json":"https://pith.science/api/pith-number/GQPRER7NMOKNVWIT6L5UDA6CR7/graph.json","events_json":"https://pith.science/api/pith-number/GQPRER7NMOKNVWIT6L5UDA6CR7/events.json","paper":"https://pith.science/paper/GQPRER7N"},"agent_actions":{"view_html":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7","download_json":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7.json","view_paper":"https://pith.science/paper/GQPRER7N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1912.02761&json=true","fetch_graph":"https://pith.science/api/pith-number/GQPRER7NMOKNVWIT6L5UDA6CR7/graph.json","fetch_events":"https://pith.science/api/pith-number/GQPRER7NMOKNVWIT6L5UDA6CR7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7/action/storage_attestation","attest_author":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7/action/author_attestation","sign_citation":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7/action/citation_signature","submit_replication":"https://pith.science/pith/GQPRER7NMOKNVWIT6L5UDA6CR7/action/replication_record"}},"created_at":"2026-07-05T01:01:01.520542+00:00","updated_at":"2026-07-05T01:01:01.520542+00:00"}