{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:QV3YXPL67TCNOQA2XFIYFLX6AY","short_pith_number":"pith:QV3YXPL6","canonical_record":{"source":{"id":"2509.26627","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-30T17:58:20Z","cross_cats_sorted":["cs.LG","cs.RO"],"title_canon_sha256":"205e080e3732218b51c38aefd0c15dc79ff876deab59b3e83ec591ba644593e2","abstract_canon_sha256":"c9d9a9c42031db5545492f30d1bbabce9fbeadbaf91422eb03a1d09b6662a63d"},"schema_version":"1.0"},"canonical_sha256":"85778bbd7efcc4d7401ab95182aefe0622e68648cdd0f433b253135bd5c3f71e","source":{"kind":"arxiv","id":"2509.26627","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.26627","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"arxiv_version","alias_value":"2509.26627v3","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.26627","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"pith_short_12","alias_value":"QV3YXPL67TCN","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"pith_short_16","alias_value":"QV3YXPL67TCNOQA2","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"pith_short_8","alias_value":"QV3YXPL6","created_at":"2026-05-21T01:05:10Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:QV3YXPL67TCNOQA2XFIYFLX6AY","target":"record","payload":{"canonical_record":{"source":{"id":"2509.26627","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-30T17:58:20Z","cross_cats_sorted":["cs.LG","cs.RO"],"title_canon_sha256":"205e080e3732218b51c38aefd0c15dc79ff876deab59b3e83ec591ba644593e2","abstract_canon_sha256":"c9d9a9c42031db5545492f30d1bbabce9fbeadbaf91422eb03a1d09b6662a63d"},"schema_version":"1.0"},"canonical_sha256":"85778bbd7efcc4d7401ab95182aefe0622e68648cdd0f433b253135bd5c3f71e","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:05:10.196907Z","signature_b64":"KirZXEWktO83dvi45+Zr0mLwa+zgzaIB8K6ifm/7KWHzEN7qW/005Ei1y1S+VJbMYE4COBdQfdNVk8nJnXD/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"85778bbd7efcc4d7401ab95182aefe0622e68648cdd0f433b253135bd5c3f71e","last_reissued_at":"2026-05-21T01:05:10.196001Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:05:10.196001Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2509.26627","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":"2026-05-21T01:05:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1eA/hRCHr1AsLfZKhwiOUyMzZDNhrEn68Yd1bDFARLuFM7JO1mEsRkJ3XrwUJviuaygOvJROtAwHT4JALqiFAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-23T02:03:57.383722Z"},"content_sha256":"ae82719f6415c1ef137cc57263e062958f79997b7df5dfc241ff8f9a8220b005","schema_version":"1.0","event_id":"sha256:ae82719f6415c1ef137cc57263e062958f79997b7df5dfc241ff8f9a8220b005"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:QV3YXPL67TCNOQA2XFIYFLX6AY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"TimeRewarder: Learning Dense Reward from Passive Videos via Frame-wise Temporal Distance","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Modeling temporal distances between frames in passive videos yields step-wise proxy rewards that let reinforcement learning succeed on most Meta-World tasks with far fewer interactions than sparse or hand-designed rewards.","cross_cats":["cs.LG","cs.RO"],"primary_cat":"cs.AI","authors_text":"Chuan Wen, Dinesh Jayaraman, Yang Gao, Yihang Hu, Yuyang Liu","submitted_at":"2025-09-30T17:58:20Z","abstract_excerpt":"Designing dense rewards is crucial for reinforcement learning (RL), yet in robotics it often demands extensive manual effort and lacks scalability. One promising solution is to view task progress as a dense reward signal, as it quantifies the degree to which actions advance the system toward task completion over time. We present TimeRewarder, a simple yet effective reward learning method that derives progress estimation signals from passive videos, including robot demonstrations and human videos, by modeling temporal distances between frame pairs. We then demonstrate how TimeRewarder can suppl"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"TimeRewarder supplies step-wise proxy rewards from passive videos that enable nearly perfect success in 9/10 Meta-World tasks with only 200,000 interactions per task, outperforming both prior methods and manually designed environment dense rewards.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That modeling temporal distances between frame pairs in passive videos (robot demos or human videos) produces a reliable proxy for task progress that generalizes to guide RL without additional supervision or task-specific tuning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TimeRewarder derives progress-based dense rewards from passive videos via frame-wise temporal distance modeling and uses them as proxy rewards to boost RL success on Meta-World tasks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Modeling temporal distances between frames in passive videos yields step-wise proxy rewards that let reinforcement learning succeed on most Meta-World tasks with far fewer interactions than sparse or hand-designed rewards.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2a99a60c9c271b9c4dc5274637a2388c0396cfe0029f428c5e8f46fed5300eab"},"source":{"id":"2509.26627","kind":"arxiv","version":3},"verdict":{"id":"acee4e24-4c9e-43c7-bd9a-f005ae8e45e0","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-18T11:25:42.306445Z","strongest_claim":"TimeRewarder supplies step-wise proxy rewards from passive videos that enable nearly perfect success in 9/10 Meta-World tasks with only 200,000 interactions per task, outperforming both prior methods and manually designed environment dense rewards.","one_line_summary":"TimeRewarder derives progress-based dense rewards from passive videos via frame-wise temporal distance modeling and uses them as proxy rewards to boost RL success on Meta-World tasks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That modeling temporal distances between frame pairs in passive videos (robot demos or human videos) produces a reliable proxy for task progress that generalizes to guide RL without additional supervision or task-specific tuning.","pith_extraction_headline":"Modeling temporal distances between frames in passive videos yields step-wise proxy rewards that let reinforcement learning succeed on most Meta-World tasks with far fewer interactions than sparse or hand-designed rewards."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2509.26627/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":"acee4e24-4c9e-43c7-bd9a-f005ae8e45e0"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-21T01:05:10Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xxqwrt132KUFZyOLa60O7RSdHdT5xCOFmeyu/KI6341hm63/xYXPVDYSGaFvRW/otMEFWDCTs6ZHS5AuJsVwBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-23T02:03:57.384213Z"},"content_sha256":"e58284dbf2a814a63e8f3d944667c98def002b0ad1477de288c7d1bedeb726a3","schema_version":"1.0","event_id":"sha256:e58284dbf2a814a63e8f3d944667c98def002b0ad1477de288c7d1bedeb726a3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/QV3YXPL67TCNOQA2XFIYFLX6AY/bundle.json","state_url":"https://pith.science/pith/QV3YXPL67TCNOQA2XFIYFLX6AY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/QV3YXPL67TCNOQA2XFIYFLX6AY/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":"2026-05-23T02:03:57Z","links":{"resolver":"https://pith.science/pith/QV3YXPL67TCNOQA2XFIYFLX6AY","bundle":"https://pith.science/pith/QV3YXPL67TCNOQA2XFIYFLX6AY/bundle.json","state":"https://pith.science/pith/QV3YXPL67TCNOQA2XFIYFLX6AY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/QV3YXPL67TCNOQA2XFIYFLX6AY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:QV3YXPL67TCNOQA2XFIYFLX6AY","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":"c9d9a9c42031db5545492f30d1bbabce9fbeadbaf91422eb03a1d09b6662a63d","cross_cats_sorted":["cs.LG","cs.RO"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-30T17:58:20Z","title_canon_sha256":"205e080e3732218b51c38aefd0c15dc79ff876deab59b3e83ec591ba644593e2"},"schema_version":"1.0","source":{"id":"2509.26627","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2509.26627","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"arxiv_version","alias_value":"2509.26627v3","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.26627","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"pith_short_12","alias_value":"QV3YXPL67TCN","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"pith_short_16","alias_value":"QV3YXPL67TCNOQA2","created_at":"2026-05-21T01:05:10Z"},{"alias_kind":"pith_short_8","alias_value":"QV3YXPL6","created_at":"2026-05-21T01:05:10Z"}],"graph_snapshots":[{"event_id":"sha256:e58284dbf2a814a63e8f3d944667c98def002b0ad1477de288c7d1bedeb726a3","target":"graph","created_at":"2026-05-21T01:05:10Z","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":"TimeRewarder supplies step-wise proxy rewards from passive videos that enable nearly perfect success in 9/10 Meta-World tasks with only 200,000 interactions per task, outperforming both prior methods and manually designed environment dense rewards."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That modeling temporal distances between frame pairs in passive videos (robot demos or human videos) produces a reliable proxy for task progress that generalizes to guide RL without additional supervision or task-specific tuning."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"TimeRewarder derives progress-based dense rewards from passive videos via frame-wise temporal distance modeling and uses them as proxy rewards to boost RL success on Meta-World tasks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Modeling temporal distances between frames in passive videos yields step-wise proxy rewards that let reinforcement learning succeed on most Meta-World tasks with far fewer interactions than sparse or hand-designed rewards."}],"snapshot_sha256":"2a99a60c9c271b9c4dc5274637a2388c0396cfe0029f428c5e8f46fed5300eab"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2509.26627/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Designing dense rewards is crucial for reinforcement learning (RL), yet in robotics it often demands extensive manual effort and lacks scalability. One promising solution is to view task progress as a dense reward signal, as it quantifies the degree to which actions advance the system toward task completion over time. We present TimeRewarder, a simple yet effective reward learning method that derives progress estimation signals from passive videos, including robot demonstrations and human videos, by modeling temporal distances between frame pairs. We then demonstrate how TimeRewarder can suppl","authors_text":"Chuan Wen, Dinesh Jayaraman, Yang Gao, Yihang Hu, Yuyang Liu","cross_cats":["cs.LG","cs.RO"],"headline":"Modeling temporal distances between frames in passive videos yields step-wise proxy rewards that let reinforcement learning succeed on most Meta-World tasks with far fewer interactions than sparse or hand-designed rewards.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-30T17:58:20Z","title":"TimeRewarder: Learning Dense Reward from Passive Videos via Frame-wise Temporal Distance"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.26627","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-18T11:25:42.306445Z","id":"acee4e24-4c9e-43c7-bd9a-f005ae8e45e0","model_set":{"reader":"grok-4.3"},"one_line_summary":"TimeRewarder derives progress-based dense rewards from passive videos via frame-wise temporal distance modeling and uses them as proxy rewards to boost RL success on Meta-World tasks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Modeling temporal distances between frames in passive videos yields step-wise proxy rewards that let reinforcement learning succeed on most Meta-World tasks with far fewer interactions than sparse or hand-designed rewards.","strongest_claim":"TimeRewarder supplies step-wise proxy rewards from passive videos that enable nearly perfect success in 9/10 Meta-World tasks with only 200,000 interactions per task, outperforming both prior methods and manually designed environment dense rewards.","weakest_assumption":"That modeling temporal distances between frame pairs in passive videos (robot demos or human videos) produces a reliable proxy for task progress that generalizes to guide RL without additional supervision or task-specific tuning."}},"verdict_id":"acee4e24-4c9e-43c7-bd9a-f005ae8e45e0"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:ae82719f6415c1ef137cc57263e062958f79997b7df5dfc241ff8f9a8220b005","target":"record","created_at":"2026-05-21T01:05:10Z","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":"c9d9a9c42031db5545492f30d1bbabce9fbeadbaf91422eb03a1d09b6662a63d","cross_cats_sorted":["cs.LG","cs.RO"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-09-30T17:58:20Z","title_canon_sha256":"205e080e3732218b51c38aefd0c15dc79ff876deab59b3e83ec591ba644593e2"},"schema_version":"1.0","source":{"id":"2509.26627","kind":"arxiv","version":3}},"canonical_sha256":"85778bbd7efcc4d7401ab95182aefe0622e68648cdd0f433b253135bd5c3f71e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"85778bbd7efcc4d7401ab95182aefe0622e68648cdd0f433b253135bd5c3f71e","first_computed_at":"2026-05-21T01:05:10.196001Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-21T01:05:10.196001Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KirZXEWktO83dvi45+Zr0mLwa+zgzaIB8K6ifm/7KWHzEN7qW/005Ei1y1S+VJbMYE4COBdQfdNVk8nJnXD/Ag==","signature_status":"signed_v1","signed_at":"2026-05-21T01:05:10.196907Z","signed_message":"canonical_sha256_bytes"},"source_id":"2509.26627","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ae82719f6415c1ef137cc57263e062958f79997b7df5dfc241ff8f9a8220b005","sha256:e58284dbf2a814a63e8f3d944667c98def002b0ad1477de288c7d1bedeb726a3"],"state_sha256":"a80cbdd1b3b20cf2754840b0b3dd62385d825eaa93272a095a24c7f0c035e7f7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"U19Oju01zK4aC0VLvbtCmARltlPqWqGbGxRsgB72m83bgMqFMJpLUyj3VCQCi0SCc7jjgZACRm8Th4r+dSXiBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-23T02:03:57.386783Z","bundle_sha256":"ba4743eb3deca2926141db10c377fe59c2aa20f82155fb6c5653ff30b6f24197"}}