{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:VR75DMQ7N3BWEYDGRI2VLXTZY3","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":"7f8fb2debb4121cb60fd8c9e1997395fc32152ee419161dea64c5c7eff8c21e7","cross_cats_sorted":["q-bio.BM"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CE","submitted_at":"2026-05-09T03:55:15Z","title_canon_sha256":"a75216692bf65b82c5279a2c9a3a6ba14a51bb279bdf004450d75f0076a70e0f"},"schema_version":"1.0","source":{"id":"2605.08659","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.08659","created_at":"2026-05-20T00:03:15Z"},{"alias_kind":"arxiv_version","alias_value":"2605.08659v2","created_at":"2026-05-20T00:03:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.08659","created_at":"2026-05-20T00:03:15Z"},{"alias_kind":"pith_short_12","alias_value":"VR75DMQ7N3BW","created_at":"2026-05-20T00:03:15Z"},{"alias_kind":"pith_short_16","alias_value":"VR75DMQ7N3BWEYDG","created_at":"2026-05-20T00:03:15Z"},{"alias_kind":"pith_short_8","alias_value":"VR75DMQ7","created_at":"2026-05-20T00:03:15Z"}],"graph_snapshots":[{"event_id":"sha256:cac39e6863b0edda3aa6992c890920d242035fe1540a87c97445dbf34f419acd","target":"graph","created_at":"2026-05-20T00:03:15Z","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":"Across decoding sweeps, SGRPO expands the utility-diversity Pareto frontier and achieves the best frontier-level metrics relative to pretrained generators, GRPO, and memory-assisted GRPO when applicable."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That supergroup sampling with leave-one-out diversity contributions provides a stable, unbiased signal for redistributing set-level rewards to individual rollouts without introducing artifacts in policy optimization or depending on specific diversity metrics."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"SGRPO expands the utility-diversity Pareto frontier in biomolecular design by using supergroup sampling and leave-one-out diversity rewards combined with utility signals."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Supergroup Relative Policy Optimization expands the utility-diversity Pareto frontier for biomolecular generators by directly rewarding set-level diversity."}],"snapshot_sha256":"ceb92b76a7c4a85325837a6af15e08a98e17ffb5a7a5cdb99adb097e68412a52"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"3d287ef768fa55a71c505bf91476c8782ffa57efd01fdb1d0307e2bbb2e94773"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T22:36:35.227881Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T14:31:17.953922Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T10:52:18.332137Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.08659/integrity.json","findings":[],"snapshot_sha256":"5b843a286c965236fd2931f5009534692dfe09cf8d1792761f3fcd76db037c3c","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Biomolecular generators are often adapted with reward feedback to improve task-specific utility, but pushing utility alone can concentrate generation on a narrow family of candidates. Maintaining diversity is difficult because sample diversity is a set-level property. We introduce Supergroup Relative Policy Optimization (SGRPO), a flexible GRPO-style framework that directly constructs rewards from set-level diversity. For each condition, SGRPO samples a supergroup of candidate sets, compares their diversity under the same condition, and redistributes the group diversity reward to individual ro","authors_text":"Bin Feng, Hao Li, He Cao, Shenghua Gao, Xiangru Tang, Xinwu Ye, Yu Li, Zijing Liu","cross_cats":["q-bio.BM"],"headline":"Supergroup Relative Policy Optimization expands the utility-diversity Pareto frontier for biomolecular generators by directly rewarding set-level diversity.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CE","submitted_at":"2026-05-09T03:55:15Z","title":"Pushing Biomolecular Utility-Diversity Frontiers with Supergroup Relative Policy Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.08659","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-12T01:06:16.303088Z","id":"1551846e-d72c-4100-b774-317f3b8370cd","model_set":{"reader":"grok-4.3"},"one_line_summary":"SGRPO expands the utility-diversity Pareto frontier in biomolecular design by using supergroup sampling and leave-one-out diversity rewards combined with utility signals.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Supergroup Relative Policy Optimization expands the utility-diversity Pareto frontier for biomolecular generators by directly rewarding set-level diversity.","strongest_claim":"Across decoding sweeps, SGRPO expands the utility-diversity Pareto frontier and achieves the best frontier-level metrics relative to pretrained generators, GRPO, and memory-assisted GRPO when applicable.","weakest_assumption":"That supergroup sampling with leave-one-out diversity contributions provides a stable, unbiased signal for redistributing set-level rewards to individual rollouts without introducing artifacts in policy optimization or depending on specific diversity metrics."}},"verdict_id":"1551846e-d72c-4100-b774-317f3b8370cd"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:54ef08dab1860ec90a69620f3c66d5e4ccfd7c380207e16710211d2474128291","target":"record","created_at":"2026-05-20T00:03:15Z","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":"7f8fb2debb4121cb60fd8c9e1997395fc32152ee419161dea64c5c7eff8c21e7","cross_cats_sorted":["q-bio.BM"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CE","submitted_at":"2026-05-09T03:55:15Z","title_canon_sha256":"a75216692bf65b82c5279a2c9a3a6ba14a51bb279bdf004450d75f0076a70e0f"},"schema_version":"1.0","source":{"id":"2605.08659","kind":"arxiv","version":2}},"canonical_sha256":"ac7fd1b21f6ec36260668a3555de79c6ef602b51d9b1f2c1308a4363b250866e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ac7fd1b21f6ec36260668a3555de79c6ef602b51d9b1f2c1308a4363b250866e","first_computed_at":"2026-05-20T00:03:15.383856Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:03:15.383856Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"d0Vcl1omXwqt7CF0Q0mBn4vA2oPEsvMVT/nJ63/FV98h0H33TKaVuFIz+RKaKsoC3iYfSOtYWGPktqr411XoAg==","signature_status":"signed_v1","signed_at":"2026-05-20T00:03:15.384754Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.08659","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:54ef08dab1860ec90a69620f3c66d5e4ccfd7c380207e16710211d2474128291","sha256:cac39e6863b0edda3aa6992c890920d242035fe1540a87c97445dbf34f419acd"],"state_sha256":"7ae40eb00fd6b45bedc8930100418799bf4501f0619c4526f0a87be277df16c6"}