{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KZEWYUAQRQEGRTWI5LD65MKBD2","short_pith_number":"pith:KZEWYUAQ","schema_version":"1.0","canonical_sha256":"56496c50108c0868cec8eac7eeb1411eb824d09be6588daa28e952bccde30bbd","source":{"kind":"arxiv","id":"2505.19356","version":2},"attestation_state":"computed","paper":{"title":"Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.IR","authors_text":"Kidist Amde Mekonnen, Maarten de Rijke, Yosef Worku Alemneh","submitted_at":"2025-05-25T23:06:20Z","abstract_excerpt":"Neural retrieval methods using transformer-based pre-trained language models have advanced multilingual and cross-lingual retrieval. However, their effectiveness for low-resource, morphologically rich languages such as Amharic remains underexplored due to data scarcity and suboptimal tokenization. We address this gap by introducing Amharic-specific dense retrieval models based on pre-trained Amharic BERT and RoBERTa backbones. Our proposed RoBERTa-Base-Amharic-Embed model (110M parameters) achieves a 17.6% relative improvement in MRR@10 and a 9.86% gain in Recall@10 over the strongest multilin"},"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":"2505.19356","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2025-05-25T23:06:20Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"ba04b45eb177f302baf153f4db5c6d198063b8a17e51a40e861649dc05d85f51","abstract_canon_sha256":"ed121a0b87396301142dac63fc91cf2ec97d8871ef9dbc34651cd3df2a10e49f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:19:07.470850Z","signature_b64":"tdRmhKDr2b3P+35p/9ypTFgKEZuuy9s+BX9UN6Mp3+avyi33OPl8FA9p5L2lRTqvjZMeECpurp+NQNbPg1HgCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"56496c50108c0868cec8eac7eeb1411eb824d09be6588daa28e952bccde30bbd","last_reissued_at":"2026-07-05T11:19:07.470402Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:19:07.470402Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimized Text Embedding Models and Benchmarks for Amharic Passage Retrieval","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.IR","authors_text":"Kidist Amde Mekonnen, Maarten de Rijke, Yosef Worku Alemneh","submitted_at":"2025-05-25T23:06:20Z","abstract_excerpt":"Neural retrieval methods using transformer-based pre-trained language models have advanced multilingual and cross-lingual retrieval. However, their effectiveness for low-resource, morphologically rich languages such as Amharic remains underexplored due to data scarcity and suboptimal tokenization. We address this gap by introducing Amharic-specific dense retrieval models based on pre-trained Amharic BERT and RoBERTa backbones. Our proposed RoBERTa-Base-Amharic-Embed model (110M parameters) achieves a 17.6% relative improvement in MRR@10 and a 9.86% gain in Recall@10 over the strongest multilin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.19356","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/2505.19356/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":"2505.19356","created_at":"2026-07-05T11:19:07.470454+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.19356v2","created_at":"2026-07-05T11:19:07.470454+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19356","created_at":"2026-07-05T11:19:07.470454+00:00"},{"alias_kind":"pith_short_12","alias_value":"KZEWYUAQRQEG","created_at":"2026-07-05T11:19:07.470454+00:00"},{"alias_kind":"pith_short_16","alias_value":"KZEWYUAQRQEGRTWI","created_at":"2026-07-05T11:19:07.470454+00:00"},{"alias_kind":"pith_short_8","alias_value":"KZEWYUAQ","created_at":"2026-07-05T11:19:07.470454+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/KZEWYUAQRQEGRTWI5LD65MKBD2","json":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2.json","graph_json":"https://pith.science/api/pith-number/KZEWYUAQRQEGRTWI5LD65MKBD2/graph.json","events_json":"https://pith.science/api/pith-number/KZEWYUAQRQEGRTWI5LD65MKBD2/events.json","paper":"https://pith.science/paper/KZEWYUAQ"},"agent_actions":{"view_html":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2","download_json":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2.json","view_paper":"https://pith.science/paper/KZEWYUAQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.19356&json=true","fetch_graph":"https://pith.science/api/pith-number/KZEWYUAQRQEGRTWI5LD65MKBD2/graph.json","fetch_events":"https://pith.science/api/pith-number/KZEWYUAQRQEGRTWI5LD65MKBD2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2/action/storage_attestation","attest_author":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2/action/author_attestation","sign_citation":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2/action/citation_signature","submit_replication":"https://pith.science/pith/KZEWYUAQRQEGRTWI5LD65MKBD2/action/replication_record"}},"created_at":"2026-07-05T11:19:07.470454+00:00","updated_at":"2026-07-05T11:19:07.470454+00:00"}