{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:FMDAUJTO2DN4OA3DQ2B2PAY5FV","short_pith_number":"pith:FMDAUJTO","canonical_record":{"source":{"id":"2409.11860","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.IR","submitted_at":"2024-09-18T10:30:50Z","cross_cats_sorted":["cs.AI","cs.CL","cs.ET","cs.HC"],"title_canon_sha256":"5439c13114549997fb7a417db1dd607d253075516db7a9431816d7e0ffae4fcb","abstract_canon_sha256":"d9f39f50c5ebfa5111dc83208c89dd097b2b09b9dcd1591dd25261e07cf33d27"},"schema_version":"1.0"},"canonical_sha256":"2b060a266ed0dbc703638683a7831d2d5002b2d2af7d0d98e925961b201b9636","source":{"kind":"arxiv","id":"2409.11860","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.11860","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"arxiv_version","alias_value":"2409.11860v1","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.11860","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"pith_short_12","alias_value":"FMDAUJTO2DN4","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"pith_short_16","alias_value":"FMDAUJTO2DN4OA3D","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"pith_short_8","alias_value":"FMDAUJTO","created_at":"2026-07-05T09:08:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:FMDAUJTO2DN4OA3DQ2B2PAY5FV","target":"record","payload":{"canonical_record":{"source":{"id":"2409.11860","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.IR","submitted_at":"2024-09-18T10:30:50Z","cross_cats_sorted":["cs.AI","cs.CL","cs.ET","cs.HC"],"title_canon_sha256":"5439c13114549997fb7a417db1dd607d253075516db7a9431816d7e0ffae4fcb","abstract_canon_sha256":"d9f39f50c5ebfa5111dc83208c89dd097b2b09b9dcd1591dd25261e07cf33d27"},"schema_version":"1.0"},"canonical_sha256":"2b060a266ed0dbc703638683a7831d2d5002b2d2af7d0d98e925961b201b9636","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:08:39.033265Z","signature_b64":"BqjhQ9wa+0vkRR3yYThoVTTxQl5HuXzVhP5ZT7Fv3vVWf6IvL6UHRACUPy4wDXbLadSOuCQ2AcqQUp7p1fp5Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2b060a266ed0dbc703638683a7831d2d5002b2d2af7d0d98e925961b201b9636","last_reissued_at":"2026-07-05T09:08:39.032810Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:08:39.032810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2409.11860","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:08:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1Hw2wP9nat2r0/Jle5wNMaY9c3jRdK5v5BUeVBR/wFKBi7/4GeMXnF48FAziK01ce44WJaipI3chEuIKIqWtDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T17:36:39.187229Z"},"content_sha256":"ff66e0f92fdf7e1ca93e3d927130d6dea7fe07221c73dc2416bac22a9c72e10c","schema_version":"1.0","event_id":"sha256:ff66e0f92fdf7e1ca93e3d927130d6dea7fe07221c73dc2416bac22a9c72e10c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:FMDAUJTO2DN4OA3DQ2B2PAY5FV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Retrieve, Annotate, Evaluate, Repeat: Leveraging Multimodal LLMs for Large-Scale Product Retrieval Evaluation","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.ET","cs.HC"],"primary_cat":"cs.IR","authors_text":"Ana Peleteiro Ramallo, Josip Krapac, Kasra Hosseini, Roland Vollgraf, Thomas Kober, Weiwei Cheng","submitted_at":"2024-09-18T10:30:50Z","abstract_excerpt":"Evaluating production-level retrieval systems at scale is a crucial yet challenging task due to the limited availability of a large pool of well-trained human annotators. Large Language Models (LLMs) have the potential to address this scaling issue and offer a viable alternative to humans for the bulk of annotation tasks. In this paper, we propose a framework for assessing the product search engines in a large-scale e-commerce setting, leveraging Multimodal LLMs for (i) generating tailored annotation guidelines for individual queries, and (ii) conducting the subsequent annotation task. Our met"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.11860","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/2409.11860/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":"2026-07-05T09:08:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"IQ7VKGud6PtMhkj4/Jrl3yRbP/r0OiWYN6BWwdmtyQJnGgFcyF8yxsqDbuiUC5rpJ3uQ9IorOl+ZEv0euxs/Bg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T17:36:39.188264Z"},"content_sha256":"3af530c8fa374cc39a26692de82581a967fad1a0e65399076773f69274ad4493","schema_version":"1.0","event_id":"sha256:3af530c8fa374cc39a26692de82581a967fad1a0e65399076773f69274ad4493"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV/bundle.json","state_url":"https://pith.science/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV/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":"2026-08-08T17:36:39Z","links":{"resolver":"https://pith.science/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV","bundle":"https://pith.science/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV/bundle.json","state":"https://pith.science/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FMDAUJTO2DN4OA3DQ2B2PAY5FV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:FMDAUJTO2DN4OA3DQ2B2PAY5FV","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":"d9f39f50c5ebfa5111dc83208c89dd097b2b09b9dcd1591dd25261e07cf33d27","cross_cats_sorted":["cs.AI","cs.CL","cs.ET","cs.HC"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.IR","submitted_at":"2024-09-18T10:30:50Z","title_canon_sha256":"5439c13114549997fb7a417db1dd607d253075516db7a9431816d7e0ffae4fcb"},"schema_version":"1.0","source":{"id":"2409.11860","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2409.11860","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"arxiv_version","alias_value":"2409.11860v1","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.11860","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"pith_short_12","alias_value":"FMDAUJTO2DN4","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"pith_short_16","alias_value":"FMDAUJTO2DN4OA3D","created_at":"2026-07-05T09:08:39Z"},{"alias_kind":"pith_short_8","alias_value":"FMDAUJTO","created_at":"2026-07-05T09:08:39Z"}],"graph_snapshots":[{"event_id":"sha256:3af530c8fa374cc39a26692de82581a967fad1a0e65399076773f69274ad4493","target":"graph","created_at":"2026-07-05T09:08:39Z","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/2409.11860/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Evaluating production-level retrieval systems at scale is a crucial yet challenging task due to the limited availability of a large pool of well-trained human annotators. Large Language Models (LLMs) have the potential to address this scaling issue and offer a viable alternative to humans for the bulk of annotation tasks. In this paper, we propose a framework for assessing the product search engines in a large-scale e-commerce setting, leveraging Multimodal LLMs for (i) generating tailored annotation guidelines for individual queries, and (ii) conducting the subsequent annotation task. Our met","authors_text":"Ana Peleteiro Ramallo, Josip Krapac, Kasra Hosseini, Roland Vollgraf, Thomas Kober, Weiwei Cheng","cross_cats":["cs.AI","cs.CL","cs.ET","cs.HC"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.IR","submitted_at":"2024-09-18T10:30:50Z","title":"Retrieve, Annotate, Evaluate, Repeat: Leveraging Multimodal LLMs for Large-Scale Product Retrieval Evaluation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.11860","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:ff66e0f92fdf7e1ca93e3d927130d6dea7fe07221c73dc2416bac22a9c72e10c","target":"record","created_at":"2026-07-05T09:08:39Z","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":"d9f39f50c5ebfa5111dc83208c89dd097b2b09b9dcd1591dd25261e07cf33d27","cross_cats_sorted":["cs.AI","cs.CL","cs.ET","cs.HC"],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.IR","submitted_at":"2024-09-18T10:30:50Z","title_canon_sha256":"5439c13114549997fb7a417db1dd607d253075516db7a9431816d7e0ffae4fcb"},"schema_version":"1.0","source":{"id":"2409.11860","kind":"arxiv","version":1}},"canonical_sha256":"2b060a266ed0dbc703638683a7831d2d5002b2d2af7d0d98e925961b201b9636","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2b060a266ed0dbc703638683a7831d2d5002b2d2af7d0d98e925961b201b9636","first_computed_at":"2026-07-05T09:08:39.032810Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:08:39.032810Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"BqjhQ9wa+0vkRR3yYThoVTTxQl5HuXzVhP5ZT7Fv3vVWf6IvL6UHRACUPy4wDXbLadSOuCQ2AcqQUp7p1fp5Aw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:08:39.033265Z","signed_message":"canonical_sha256_bytes"},"source_id":"2409.11860","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ff66e0f92fdf7e1ca93e3d927130d6dea7fe07221c73dc2416bac22a9c72e10c","sha256:3af530c8fa374cc39a26692de82581a967fad1a0e65399076773f69274ad4493"],"state_sha256":"4052d01f69f8fcad35962fb2731f9962c0553224d35f804091867f3fc54cf6c1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KZiYaW3RnRrdLONT8ofWh8YWXSgai7JpgofWTPZmnsr94hgVv2gb1AECWpHYqGjzVaznnnc7MpK3zXjWq0vMAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T17:36:39.195419Z","bundle_sha256":"d6e4c8bb87bed7ee52a6f21099695cdc25a8420dc7d4a15df5e863a1f3b68cde"}}