{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:ZODXM66DIU7FBQ5DC5VOZ3LUEA","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":"2980f5f167d6bd6f17bb181a3e0ac3c68cdaed9c63463a0bc13d8e75bab75e94","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2020-10-05T16:37:52Z","title_canon_sha256":"0c1f4496f79b2bb17eaafac7e62dee7b97a22d340471a6fb1dc08c979554f80d"},"schema_version":"1.0","source":{"id":"2010.02140","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2010.02140","created_at":"2026-07-05T01:40:17Z"},{"alias_kind":"arxiv_version","alias_value":"2010.02140v1","created_at":"2026-07-05T01:40:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2010.02140","created_at":"2026-07-05T01:40:17Z"},{"alias_kind":"pith_short_12","alias_value":"ZODXM66DIU7F","created_at":"2026-07-05T01:40:17Z"},{"alias_kind":"pith_short_16","alias_value":"ZODXM66DIU7FBQ5D","created_at":"2026-07-05T01:40:17Z"},{"alias_kind":"pith_short_8","alias_value":"ZODXM66D","created_at":"2026-07-05T01:40:17Z"}],"graph_snapshots":[{"event_id":"sha256:a5f5e1425886c062c94dd665f6268f25f80eae58df5720d5bcae8bd8323b751d","target":"graph","created_at":"2026-07-05T01:40:17Z","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/2010.02140/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The lack of time-efficient and reliable evaluation methods hamper the development of conversational dialogue systems (chatbots). Evaluations requiring humans to converse with chatbots are time and cost-intensive, put high cognitive demands on the human judges, and yield low-quality results. In this work, we introduce \\emph{Spot The Bot}, a cost-efficient and robust evaluation framework that replaces human-bot conversations with conversations between bots. Human judges then only annotate for each entity in a conversation whether they think it is human or not (assuming there are humans participa","authors_text":"Aitor Soroa, Alvaro Rodrigo, Don Tuggener, Eneko Agirre, Jan Deriu, Jon Ander Campos, Mark Cieliebak, Pius von D\\\"aniken, Thiziri Belkacem","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2020-10-05T16:37:52Z","title":"Spot The Bot: A Robust and Efficient Framework for the Evaluation of Conversational Dialogue Systems"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2010.02140","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":"sha256:1a14511c881fd22816da4eb5215c1ed80a0c0001a005cdd6f1c1b7ec9b184397","target":"record","created_at":"2026-07-05T01:40:17Z","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":"2980f5f167d6bd6f17bb181a3e0ac3c68cdaed9c63463a0bc13d8e75bab75e94","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2020-10-05T16:37:52Z","title_canon_sha256":"0c1f4496f79b2bb17eaafac7e62dee7b97a22d340471a6fb1dc08c979554f80d"},"schema_version":"1.0","source":{"id":"2010.02140","kind":"arxiv","version":1}},"canonical_sha256":"cb87767bc3453e50c3a3176aeced74203a689052717c0fea8c00e8353ed7c0de","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cb87767bc3453e50c3a3176aeced74203a689052717c0fea8c00e8353ed7c0de","first_computed_at":"2026-07-05T01:40:17.776736Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T01:40:17.776736Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"itV9fNIqhKDz0GNk5Vf9aIQaoPl3eB0jttrNBPrP7B3QW4pXzm92fEbs/SFfUEergiFXF+pESiLs+0ai6MYTCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T01:40:17.777132Z","signed_message":"canonical_sha256_bytes"},"source_id":"2010.02140","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1a14511c881fd22816da4eb5215c1ed80a0c0001a005cdd6f1c1b7ec9b184397","sha256:a5f5e1425886c062c94dd665f6268f25f80eae58df5720d5bcae8bd8323b751d"],"state_sha256":"fd1ae8296ac5333c792f0e4368fc7dfe65ef2327ae9c560196b5f305c1ae1ce0"}