{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:6MCM3MHUO3GROXHXXRNI524NAK","short_pith_number":"pith:6MCM3MHU","canonical_record":{"source":{"id":"2602.15752","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-02-17T17:30:53Z","cross_cats_sorted":[],"title_canon_sha256":"a89acae1b5fc8c9bf1e4a5038373d651560ffe4322cdd7110639a43c30e3ce54","abstract_canon_sha256":"7946dfa3f296150e470e7a5e0abb96ae97dd8d3c2e2ea2fbefe4335acd84f0a1"},"schema_version":"1.0"},"canonical_sha256":"f304cdb0f476cd175cf7bc5a8eeb8d029df68ba4ae3b8d9687504058952cd31a","source":{"kind":"arxiv","id":"2602.15752","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.15752","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"arxiv_version","alias_value":"2602.15752v3","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.15752","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"pith_short_12","alias_value":"6MCM3MHUO3GR","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"pith_short_16","alias_value":"6MCM3MHUO3GROXHX","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"pith_short_8","alias_value":"6MCM3MHU","created_at":"2026-05-20T01:05:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:6MCM3MHUO3GROXHXXRNI524NAK","target":"record","payload":{"canonical_record":{"source":{"id":"2602.15752","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-02-17T17:30:53Z","cross_cats_sorted":[],"title_canon_sha256":"a89acae1b5fc8c9bf1e4a5038373d651560ffe4322cdd7110639a43c30e3ce54","abstract_canon_sha256":"7946dfa3f296150e470e7a5e0abb96ae97dd8d3c2e2ea2fbefe4335acd84f0a1"},"schema_version":"1.0"},"canonical_sha256":"f304cdb0f476cd175cf7bc5a8eeb8d029df68ba4ae3b8d9687504058952cd31a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:05:09.432502Z","signature_b64":"4UZL+JbswdEHEu8UYBSWi80AQuflvWCw9lkMLLEzX1WKN4Y+NToW4JU6Abp2CH9zu/ySeRUHtmTHYt4Q2ronAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f304cdb0f476cd175cf7bc5a8eeb8d029df68ba4ae3b8d9687504058952cd31a","last_reissued_at":"2026-05-20T01:05:09.431764Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:05:09.431764Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.15752","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T01:05:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0n90gnL/A51oS2TdonXiTgF4tuX7Fp2raaznY6xISQFMfSk+7pwyL+KZpq7gbnqShwqb0MhbK0InipMfIeOJBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T01:15:02.364664Z"},"content_sha256":"1328800580b3c0015a4a07f382c022053e4c9d8b53c98b8d15859198f35579e1","schema_version":"1.0","event_id":"sha256:1328800580b3c0015a4a07f382c022053e4c9d8b53c98b8d15859198f35579e1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:6MCM3MHUO3GROXHXXRNI524NAK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Beyond Match Maximization and Fairness: Retention-Optimized Two-Sided Matching","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MRet maximizes retention in two-sided matching by learning personalized curves that predict how recommendations affect each user's decision to stay.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Koichi Tanaka, Masahiro Nomura, Ren Kishimoto, Rikiya Takehi, Riku Togashi, Yoji Tomita, Yuta Saito","submitted_at":"2026-02-17T17:30:53Z","abstract_excerpt":"On two-sided matching platforms such as online dating and recruiting, recommendation algorithms often aim to maximize the total number of matches. However, this objective creates an imbalance, where some users receive far too many matches while many others receive very few and eventually abandon the platform. Retaining users is crucial for many platforms, such as those that depend heavily on subscriptions. Some may use fairness objectives to solve the problem of match maximization. However, fairness in itself is not the ultimate objective for many platforms, as users do not suddenly reward the"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"empirical evaluations on synthetic and real-world datasets from a major online dating platform show that MRet achieves higher user retention, since conventional methods optimize matches or fairness rather than retention.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"that personalized retention curves learned from each user's profile and interaction history accurately capture how future recommendations will affect that user's decision to stay or leave, and that the joint optimization over both sides correctly allocates limited matches to maximize aggregate retention.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MRet is a dynamic LTR algorithm that maximizes platform-wide user retention in two-sided matching by jointly optimizing retention gains for both the recipient and the recommended users via learned personalized retention curves.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MRet maximizes retention in two-sided matching by learning personalized curves that predict how recommendations affect each user's decision to stay.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2f078c9fd0c8ad13c4d15902d0e6d837eafdd1a5dabf0a62cad56dffba253b71"},"source":{"id":"2602.15752","kind":"arxiv","version":3},"verdict":{"id":"77fe7a3a-d941-4f49-9252-d4cc007ead57","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T21:35:34.872894Z","strongest_claim":"empirical evaluations on synthetic and real-world datasets from a major online dating platform show that MRet achieves higher user retention, since conventional methods optimize matches or fairness rather than retention.","one_line_summary":"MRet is a dynamic LTR algorithm that maximizes platform-wide user retention in two-sided matching by jointly optimizing retention gains for both the recipient and the recommended users via learned personalized retention curves.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"that personalized retention curves learned from each user's profile and interaction history accurately capture how future recommendations will affect that user's decision to stay or leave, and that the joint optimization over both sides correctly allocates limited matches to maximize aggregate retention.","pith_extraction_headline":"MRet maximizes retention in two-sided matching by learning personalized curves that predict how recommendations affect each user's decision to stay."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2602.15752/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":"77fe7a3a-d941-4f49-9252-d4cc007ead57"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-20T01:05:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iNlji7L53W/+HUC7J6rmy+qMe4O5TCD0GwFIez4wDxTDVchPd22LusNxq3rKMU+rdn/uIQ+92/6vSVyv1c+IDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-09T01:15:02.365612Z"},"content_sha256":"1cb1893e9b36d0891586858e78455395930edd93c9791db960a27167c054d094","schema_version":"1.0","event_id":"sha256:1cb1893e9b36d0891586858e78455395930edd93c9791db960a27167c054d094"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6MCM3MHUO3GROXHXXRNI524NAK/bundle.json","state_url":"https://pith.science/pith/6MCM3MHUO3GROXHXXRNI524NAK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6MCM3MHUO3GROXHXXRNI524NAK/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":"2026-06-09T01:15:02Z","links":{"resolver":"https://pith.science/pith/6MCM3MHUO3GROXHXXRNI524NAK","bundle":"https://pith.science/pith/6MCM3MHUO3GROXHXXRNI524NAK/bundle.json","state":"https://pith.science/pith/6MCM3MHUO3GROXHXXRNI524NAK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6MCM3MHUO3GROXHXXRNI524NAK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:6MCM3MHUO3GROXHXXRNI524NAK","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":"7946dfa3f296150e470e7a5e0abb96ae97dd8d3c2e2ea2fbefe4335acd84f0a1","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-02-17T17:30:53Z","title_canon_sha256":"a89acae1b5fc8c9bf1e4a5038373d651560ffe4322cdd7110639a43c30e3ce54"},"schema_version":"1.0","source":{"id":"2602.15752","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.15752","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"arxiv_version","alias_value":"2602.15752v3","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.15752","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"pith_short_12","alias_value":"6MCM3MHUO3GR","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"pith_short_16","alias_value":"6MCM3MHUO3GROXHX","created_at":"2026-05-20T01:05:09Z"},{"alias_kind":"pith_short_8","alias_value":"6MCM3MHU","created_at":"2026-05-20T01:05:09Z"}],"graph_snapshots":[{"event_id":"sha256:1cb1893e9b36d0891586858e78455395930edd93c9791db960a27167c054d094","target":"graph","created_at":"2026-05-20T01:05:09Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"empirical evaluations on synthetic and real-world datasets from a major online dating platform show that MRet achieves higher user retention, since conventional methods optimize matches or fairness rather than retention."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"that personalized retention curves learned from each user's profile and interaction history accurately capture how future recommendations will affect that user's decision to stay or leave, and that the joint optimization over both sides correctly allocates limited matches to maximize aggregate retention."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"MRet is a dynamic LTR algorithm that maximizes platform-wide user retention in two-sided matching by jointly optimizing retention gains for both the recipient and the recommended users via learned personalized retention curves."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"MRet maximizes retention in two-sided matching by learning personalized curves that predict how recommendations affect each user's decision to stay."}],"snapshot_sha256":"2f078c9fd0c8ad13c4d15902d0e6d837eafdd1a5dabf0a62cad56dffba253b71"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2602.15752/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"On two-sided matching platforms such as online dating and recruiting, recommendation algorithms often aim to maximize the total number of matches. However, this objective creates an imbalance, where some users receive far too many matches while many others receive very few and eventually abandon the platform. Retaining users is crucial for many platforms, such as those that depend heavily on subscriptions. Some may use fairness objectives to solve the problem of match maximization. However, fairness in itself is not the ultimate objective for many platforms, as users do not suddenly reward the","authors_text":"Koichi Tanaka, Masahiro Nomura, Ren Kishimoto, Rikiya Takehi, Riku Togashi, Yoji Tomita, Yuta Saito","cross_cats":[],"headline":"MRet maximizes retention in two-sided matching by learning personalized curves that predict how recommendations affect each user's decision to stay.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-02-17T17:30:53Z","title":"Beyond Match Maximization and Fairness: Retention-Optimized Two-Sided Matching"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.15752","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-15T21:35:34.872894Z","id":"77fe7a3a-d941-4f49-9252-d4cc007ead57","model_set":{"reader":"grok-4.3"},"one_line_summary":"MRet is a dynamic LTR algorithm that maximizes platform-wide user retention in two-sided matching by jointly optimizing retention gains for both the recipient and the recommended users via learned personalized retention curves.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"MRet maximizes retention in two-sided matching by learning personalized curves that predict how recommendations affect each user's decision to stay.","strongest_claim":"empirical evaluations on synthetic and real-world datasets from a major online dating platform show that MRet achieves higher user retention, since conventional methods optimize matches or fairness rather than retention.","weakest_assumption":"that personalized retention curves learned from each user's profile and interaction history accurately capture how future recommendations will affect that user's decision to stay or leave, and that the joint optimization over both sides correctly allocates limited matches to maximize aggregate retention."}},"verdict_id":"77fe7a3a-d941-4f49-9252-d4cc007ead57"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1328800580b3c0015a4a07f382c022053e4c9d8b53c98b8d15859198f35579e1","target":"record","created_at":"2026-05-20T01:05:09Z","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":"7946dfa3f296150e470e7a5e0abb96ae97dd8d3c2e2ea2fbefe4335acd84f0a1","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-02-17T17:30:53Z","title_canon_sha256":"a89acae1b5fc8c9bf1e4a5038373d651560ffe4322cdd7110639a43c30e3ce54"},"schema_version":"1.0","source":{"id":"2602.15752","kind":"arxiv","version":3}},"canonical_sha256":"f304cdb0f476cd175cf7bc5a8eeb8d029df68ba4ae3b8d9687504058952cd31a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f304cdb0f476cd175cf7bc5a8eeb8d029df68ba4ae3b8d9687504058952cd31a","first_computed_at":"2026-05-20T01:05:09.431764Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T01:05:09.431764Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"4UZL+JbswdEHEu8UYBSWi80AQuflvWCw9lkMLLEzX1WKN4Y+NToW4JU6Abp2CH9zu/ySeRUHtmTHYt4Q2ronAw==","signature_status":"signed_v1","signed_at":"2026-05-20T01:05:09.432502Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.15752","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1328800580b3c0015a4a07f382c022053e4c9d8b53c98b8d15859198f35579e1","sha256:1cb1893e9b36d0891586858e78455395930edd93c9791db960a27167c054d094"],"state_sha256":"cbbbdb6151749259660bdeded95f01fdb727bc1e543b54732ca06256c4893fdf"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"V2FPHjDdbPS2tWFgKm1m7vUNLtpGyMtQ/4/I7v4qArxz8lkywrzxJBY9QRdzgdbF8Er4XOdPIIz4E048SOP7Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-09T01:15:02.369811Z","bundle_sha256":"db4a90afd1f4b3a5ec398345748eeb8b06e1a7f25b9f6f2ab8bf11eef392cfa5"}}