{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:YF3WE2F6MYLAB3BE2Z4HEDAT6L","short_pith_number":"pith:YF3WE2F6","schema_version":"1.0","canonical_sha256":"c1776268be661600ec24d678720c13f2e46d10721d01e35272665a4901f8625f","source":{"kind":"arxiv","id":"2606.22652","version":1},"attestation_state":"computed","paper":{"title":"A Markov Chain Approach to Preference Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Takuya Koriyama, Tengyuan Liang","submitted_at":"2026-06-21T19:56:56Z","abstract_excerpt":"We propose Markov Chain from Human Feedback (MCHF), an elementary approach for aligning generative models from pairwise human preferences. Unlike Reinforcement Learning from Human Feedback (RLHF), which reduces comparisons to a scalar reward, and Nash Learning from Human Feedback (NLHF), which preserves pairwise utilities through a KL-regularized minimax optimization, MCHF uses pairwise preferences directly to define a transition mechanism over model outputs. Given a pairwise utility $U(x,y)$, which quantifies human preference for $y$ over $x$, and a reference probability distribution $\\mu_{\\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":"2606.22652","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-21T19:56:56Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"e469532c0745df46482ffa8467216d611d5479887559a4c92e75d9e82939b513","abstract_canon_sha256":"a770920161e117d9739be56afbd63d8361854995175d7d14f436409034c638a8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:44.009875Z","signature_b64":"HLvxPhVOvm/fw/1Hg8fIB5F1ne9wzEMYal5KBev6DHvzNOB21Mn/7el54D80IJKEyo6zX1jDwn87erTiEPR2Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c1776268be661600ec24d678720c13f2e46d10721d01e35272665a4901f8625f","last_reissued_at":"2026-06-23T02:13:44.009494Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:44.009494Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Markov Chain Approach to Preference Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Takuya Koriyama, Tengyuan Liang","submitted_at":"2026-06-21T19:56:56Z","abstract_excerpt":"We propose Markov Chain from Human Feedback (MCHF), an elementary approach for aligning generative models from pairwise human preferences. Unlike Reinforcement Learning from Human Feedback (RLHF), which reduces comparisons to a scalar reward, and Nash Learning from Human Feedback (NLHF), which preserves pairwise utilities through a KL-regularized minimax optimization, MCHF uses pairwise preferences directly to define a transition mechanism over model outputs. Given a pairwise utility $U(x,y)$, which quantifies human preference for $y$ over $x$, and a reference probability distribution $\\mu_{\\m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22652","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/2606.22652/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":"2606.22652","created_at":"2026-06-23T02:13:44.009552+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22652v1","created_at":"2026-06-23T02:13:44.009552+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22652","created_at":"2026-06-23T02:13:44.009552+00:00"},{"alias_kind":"pith_short_12","alias_value":"YF3WE2F6MYLA","created_at":"2026-06-23T02:13:44.009552+00:00"},{"alias_kind":"pith_short_16","alias_value":"YF3WE2F6MYLAB3BE","created_at":"2026-06-23T02:13:44.009552+00:00"},{"alias_kind":"pith_short_8","alias_value":"YF3WE2F6","created_at":"2026-06-23T02:13:44.009552+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/YF3WE2F6MYLAB3BE2Z4HEDAT6L","json":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L.json","graph_json":"https://pith.science/api/pith-number/YF3WE2F6MYLAB3BE2Z4HEDAT6L/graph.json","events_json":"https://pith.science/api/pith-number/YF3WE2F6MYLAB3BE2Z4HEDAT6L/events.json","paper":"https://pith.science/paper/YF3WE2F6"},"agent_actions":{"view_html":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L","download_json":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L.json","view_paper":"https://pith.science/paper/YF3WE2F6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22652&json=true","fetch_graph":"https://pith.science/api/pith-number/YF3WE2F6MYLAB3BE2Z4HEDAT6L/graph.json","fetch_events":"https://pith.science/api/pith-number/YF3WE2F6MYLAB3BE2Z4HEDAT6L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L/action/storage_attestation","attest_author":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L/action/author_attestation","sign_citation":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L/action/citation_signature","submit_replication":"https://pith.science/pith/YF3WE2F6MYLAB3BE2Z4HEDAT6L/action/replication_record"}},"created_at":"2026-06-23T02:13:44.009552+00:00","updated_at":"2026-06-23T02:13:44.009552+00:00"}