{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WA5YOM4JC2FDETQYSEZNFCYU3A","short_pith_number":"pith:WA5YOM4J","schema_version":"1.0","canonical_sha256":"b03b873389168a324e189132d28b14d83f43833b2dbdfb3b989ccf05a15aa93f","source":{"kind":"arxiv","id":"2603.20990","version":2},"attestation_state":"computed","paper":{"title":"ECI: Effective Contrastive Information to Evaluate Hard-Negatives","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.IR","authors_text":"Aarush Sinha, Aman Bansal, Rahul Seetharaman","submitted_at":"2026-03-22T00:21:05Z","abstract_excerpt":"Hard-negative source selection for dense retrieval is usually decided only after fine-tuning and downstream evaluation. We propose Effective Contrastive Information (ECI), a training-free diagnostic that ranks candidate negative sources using frozen target-encoder embeddings. ECI is training-free, not label-free: each scored example requires a query, a labeled positive, and an explicit candidate negative. $\\mathrm{ECI}_{\\mathrm{sem}}$ builds a weighted residual information matrix from target consistency, semantic locality, lexical residuality, and a log-determinant diversity objective. On MS M"},"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":"2603.20990","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.IR","submitted_at":"2026-03-22T00:21:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"89fa387efb37b0f36d44f0cdca5d15049b6cac165b148c9da4cfd237976ac9c4","abstract_canon_sha256":"665ea58c5a6c3d213779928c9e59f2c43cb9dfb60ee3e4acd0c134552a873cc8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:37.079286Z","signature_b64":"W2tmUEY3d+crCf1BTwDaGUuNLBRrTPn4VZ+v7VNxH3wsfObL/g+E92XjV8xqZ7z5IhDOwy9wRCeiFxrpxu4VCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b03b873389168a324e189132d28b14d83f43833b2dbdfb3b989ccf05a15aa93f","last_reissued_at":"2026-06-05T01:14:37.078465Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:37.078465Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ECI: Effective Contrastive Information to Evaluate Hard-Negatives","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.IR","authors_text":"Aarush Sinha, Aman Bansal, Rahul Seetharaman","submitted_at":"2026-03-22T00:21:05Z","abstract_excerpt":"Hard-negative source selection for dense retrieval is usually decided only after fine-tuning and downstream evaluation. We propose Effective Contrastive Information (ECI), a training-free diagnostic that ranks candidate negative sources using frozen target-encoder embeddings. ECI is training-free, not label-free: each scored example requires a query, a labeled positive, and an explicit candidate negative. $\\mathrm{ECI}_{\\mathrm{sem}}$ builds a weighted residual information matrix from target consistency, semantic locality, lexical residuality, and a log-determinant diversity objective. On MS M"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.20990","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/2603.20990/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":"2603.20990","created_at":"2026-06-05T01:14:37.078600+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.20990v2","created_at":"2026-06-05T01:14:37.078600+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.20990","created_at":"2026-06-05T01:14:37.078600+00:00"},{"alias_kind":"pith_short_12","alias_value":"WA5YOM4JC2FD","created_at":"2026-06-05T01:14:37.078600+00:00"},{"alias_kind":"pith_short_16","alias_value":"WA5YOM4JC2FDETQY","created_at":"2026-06-05T01:14:37.078600+00:00"},{"alias_kind":"pith_short_8","alias_value":"WA5YOM4J","created_at":"2026-06-05T01:14:37.078600+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/WA5YOM4JC2FDETQYSEZNFCYU3A","json":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A.json","graph_json":"https://pith.science/api/pith-number/WA5YOM4JC2FDETQYSEZNFCYU3A/graph.json","events_json":"https://pith.science/api/pith-number/WA5YOM4JC2FDETQYSEZNFCYU3A/events.json","paper":"https://pith.science/paper/WA5YOM4J"},"agent_actions":{"view_html":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A","download_json":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A.json","view_paper":"https://pith.science/paper/WA5YOM4J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.20990&json=true","fetch_graph":"https://pith.science/api/pith-number/WA5YOM4JC2FDETQYSEZNFCYU3A/graph.json","fetch_events":"https://pith.science/api/pith-number/WA5YOM4JC2FDETQYSEZNFCYU3A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A/action/storage_attestation","attest_author":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A/action/author_attestation","sign_citation":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A/action/citation_signature","submit_replication":"https://pith.science/pith/WA5YOM4JC2FDETQYSEZNFCYU3A/action/replication_record"}},"created_at":"2026-06-05T01:14:37.078600+00:00","updated_at":"2026-06-05T01:14:37.078600+00:00"}