{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:F27RRWIRF4354IVTI6KPHVE4Q5","short_pith_number":"pith:F27RRWIR","schema_version":"1.0","canonical_sha256":"2ebf18d9112f37de22b34794f3d49c876103139a3928f60fa13c613fd37688dc","source":{"kind":"arxiv","id":"2606.06272","version":1},"attestation_state":"computed","paper":{"title":"Your GFlowNet Secretly Learns an Optimal Transport Plan","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Denis Belomestny, Ian Maksimov, Nikita Morozov, Sergey Samsonov","submitted_at":"2026-06-04T15:14:24Z","abstract_excerpt":"Generative Flow Networks (GFlowNets) are a framework for sampling structured objects via stochastic trajectories in a directed graph. In this work, we establish a theoretical connection between non-acyclic GFlowNets and optimal transport (OT). We show that fixing the initial flow distribution in a minimum-flow GFlowNet reduces its objective to a Kantorovich OT problem with graph-induced shortest path costs. At the optimum, the learned GFlowNet policy therefore encodes an optimal transport plan from the source distribution to the target distribution: we show that sampling trajectories from the "},"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":"2606.06272","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-04T15:14:24Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f50e3b0b01cf88f3a24a860d39eda4fc48a4a2b65694f150e238fd00b0225e75","abstract_canon_sha256":"d2274cc14300d1cbd985814e3f6688415ed7b59f56075e21cc1a4c1cc06094d5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:15:40.194558Z","signature_b64":"qVUc787yh1CTdB6nT/Y5ohiUztDa2nU87JSRPrY2WAyfuB2MDP1zWb4KVJelL/2DTwfCd0RTsz9UFwzd7gL9Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2ebf18d9112f37de22b34794f3d49c876103139a3928f60fa13c613fd37688dc","last_reissued_at":"2026-06-05T01:15:40.194175Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:15:40.194175Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Your GFlowNet Secretly Learns an Optimal Transport Plan","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Denis Belomestny, Ian Maksimov, Nikita Morozov, Sergey Samsonov","submitted_at":"2026-06-04T15:14:24Z","abstract_excerpt":"Generative Flow Networks (GFlowNets) are a framework for sampling structured objects via stochastic trajectories in a directed graph. In this work, we establish a theoretical connection between non-acyclic GFlowNets and optimal transport (OT). We show that fixing the initial flow distribution in a minimum-flow GFlowNet reduces its objective to a Kantorovich OT problem with graph-induced shortest path costs. At the optimum, the learned GFlowNet policy therefore encodes an optimal transport plan from the source distribution to the target distribution: we show that sampling trajectories from the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.06272","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/2606.06272/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":"2606.06272","created_at":"2026-06-05T01:15:40.194238+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.06272v1","created_at":"2026-06-05T01:15:40.194238+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.06272","created_at":"2026-06-05T01:15:40.194238+00:00"},{"alias_kind":"pith_short_12","alias_value":"F27RRWIRF435","created_at":"2026-06-05T01:15:40.194238+00:00"},{"alias_kind":"pith_short_16","alias_value":"F27RRWIRF4354IVT","created_at":"2026-06-05T01:15:40.194238+00:00"},{"alias_kind":"pith_short_8","alias_value":"F27RRWIR","created_at":"2026-06-05T01:15:40.194238+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/F27RRWIRF4354IVTI6KPHVE4Q5","json":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5.json","graph_json":"https://pith.science/api/pith-number/F27RRWIRF4354IVTI6KPHVE4Q5/graph.json","events_json":"https://pith.science/api/pith-number/F27RRWIRF4354IVTI6KPHVE4Q5/events.json","paper":"https://pith.science/paper/F27RRWIR"},"agent_actions":{"view_html":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5","download_json":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5.json","view_paper":"https://pith.science/paper/F27RRWIR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.06272&json=true","fetch_graph":"https://pith.science/api/pith-number/F27RRWIRF4354IVTI6KPHVE4Q5/graph.json","fetch_events":"https://pith.science/api/pith-number/F27RRWIRF4354IVTI6KPHVE4Q5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5/action/storage_attestation","attest_author":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5/action/author_attestation","sign_citation":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5/action/citation_signature","submit_replication":"https://pith.science/pith/F27RRWIRF4354IVTI6KPHVE4Q5/action/replication_record"}},"created_at":"2026-06-05T01:15:40.194238+00:00","updated_at":"2026-06-05T01:15:40.194238+00:00"}