{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:OTAYUAZCS63UHCMTOP53BP2NNU","short_pith_number":"pith:OTAYUAZC","schema_version":"1.0","canonical_sha256":"74c18a032297b743899373fbb0bf4d6d076e4f1dd44ca5681d6f01b002807f62","source":{"kind":"arxiv","id":"2101.10504","version":1},"attestation_state":"computed","paper":{"title":"On the Evaluation of Vision-and-Language Navigation Instructions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.AI","authors_text":"Alexander Ku, Eugene Ie, Jason Baldridge, Ming Zhao, Peter Anderson, Su Wang, Vihan Jain","submitted_at":"2021-01-26T01:03:49Z","abstract_excerpt":"Vision-and-Language Navigation wayfinding agents can be enhanced by exploiting automatically generated navigation instructions. However, existing instruction generators have not been comprehensively evaluated, and the automatic evaluation metrics used to develop them have not been validated. Using human wayfinders, we show that these generators perform on par with or only slightly better than a template-based generator and far worse than human instructors. Furthermore, we discover that BLEU, ROUGE, METEOR and CIDEr are ineffective for evaluating grounded navigation instructions. To improve ins"},"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":"2101.10504","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2021-01-26T01:03:49Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"9ac0a8a7d56c0d36ffa462048fa0f0af2c4d61a60cb4a0fd633168f8a701fc94","abstract_canon_sha256":"0b0496a7f7d2f7df0c9b4c2f290bde017808a97d43b9fef81c9213dd03143d52"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:09:59.227374Z","signature_b64":"+c43pbsh427AqZgJI/xA/bUu/Nt2P4dYhOWQjTeeccGRSOvRio1daIj08iu9nsKDuPCJGKlKoJSwq7d42EQZBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"74c18a032297b743899373fbb0bf4d6d076e4f1dd44ca5681d6f01b002807f62","last_reissued_at":"2026-07-05T02:09:59.226978Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:09:59.226978Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Evaluation of Vision-and-Language Navigation Instructions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.AI","authors_text":"Alexander Ku, Eugene Ie, Jason Baldridge, Ming Zhao, Peter Anderson, Su Wang, Vihan Jain","submitted_at":"2021-01-26T01:03:49Z","abstract_excerpt":"Vision-and-Language Navigation wayfinding agents can be enhanced by exploiting automatically generated navigation instructions. However, existing instruction generators have not been comprehensively evaluated, and the automatic evaluation metrics used to develop them have not been validated. Using human wayfinders, we show that these generators perform on par with or only slightly better than a template-based generator and far worse than human instructors. Furthermore, we discover that BLEU, ROUGE, METEOR and CIDEr are ineffective for evaluating grounded navigation instructions. To improve ins"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.10504","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/2101.10504/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":"2101.10504","created_at":"2026-07-05T02:09:59.227040+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.10504v1","created_at":"2026-07-05T02:09:59.227040+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.10504","created_at":"2026-07-05T02:09:59.227040+00:00"},{"alias_kind":"pith_short_12","alias_value":"OTAYUAZCS63U","created_at":"2026-07-05T02:09:59.227040+00:00"},{"alias_kind":"pith_short_16","alias_value":"OTAYUAZCS63UHCMT","created_at":"2026-07-05T02:09:59.227040+00:00"},{"alias_kind":"pith_short_8","alias_value":"OTAYUAZC","created_at":"2026-07-05T02:09:59.227040+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU","json":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU.json","graph_json":"https://pith.science/api/pith-number/OTAYUAZCS63UHCMTOP53BP2NNU/graph.json","events_json":"https://pith.science/api/pith-number/OTAYUAZCS63UHCMTOP53BP2NNU/events.json","paper":"https://pith.science/paper/OTAYUAZC"},"agent_actions":{"view_html":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU","download_json":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU.json","view_paper":"https://pith.science/paper/OTAYUAZC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.10504&json=true","fetch_graph":"https://pith.science/api/pith-number/OTAYUAZCS63UHCMTOP53BP2NNU/graph.json","fetch_events":"https://pith.science/api/pith-number/OTAYUAZCS63UHCMTOP53BP2NNU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU/action/storage_attestation","attest_author":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU/action/author_attestation","sign_citation":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU/action/citation_signature","submit_replication":"https://pith.science/pith/OTAYUAZCS63UHCMTOP53BP2NNU/action/replication_record"}},"created_at":"2026-07-05T02:09:59.227040+00:00","updated_at":"2026-07-05T02:09:59.227040+00:00"}