{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:GZ6GYSFJZCGCZRRVAOLOEIACB2","short_pith_number":"pith:GZ6GYSFJ","canonical_record":{"source":{"id":"2606.26175","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-24T10:32:48Z","cross_cats_sorted":[],"title_canon_sha256":"a3d256e0d7716201718f02ed07ed7b58053d3f1e15d872d05cc904e2c9762c66","abstract_canon_sha256":"22e241f9b63e474c6f4676415485681474f32d0993edbf010acd86f037648d81"},"schema_version":"1.0"},"canonical_sha256":"367c6c48a9c88c2cc6350396e220020eacce41680c36723a250b048554e551de","source":{"kind":"arxiv","id":"2606.26175","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.26175","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"arxiv_version","alias_value":"2606.26175v1","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26175","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"pith_short_12","alias_value":"GZ6GYSFJZCGC","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"pith_short_16","alias_value":"GZ6GYSFJZCGCZRRV","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"pith_short_8","alias_value":"GZ6GYSFJ","created_at":"2026-06-26T00:15:28Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:GZ6GYSFJZCGCZRRVAOLOEIACB2","target":"record","payload":{"canonical_record":{"source":{"id":"2606.26175","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-24T10:32:48Z","cross_cats_sorted":[],"title_canon_sha256":"a3d256e0d7716201718f02ed07ed7b58053d3f1e15d872d05cc904e2c9762c66","abstract_canon_sha256":"22e241f9b63e474c6f4676415485681474f32d0993edbf010acd86f037648d81"},"schema_version":"1.0"},"canonical_sha256":"367c6c48a9c88c2cc6350396e220020eacce41680c36723a250b048554e551de","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T00:15:28.524527Z","signature_b64":"2pmGHuifnRODRJRPJ9/rV7UFuY5Z6dG1tw0iHt84nM5tX1z6mdgTQPunekhQwezEgDeXLTS+o01vKoBXzhfdBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"367c6c48a9c88c2cc6350396e220020eacce41680c36723a250b048554e551de","last_reissued_at":"2026-06-26T00:15:28.524116Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T00:15:28.524116Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.26175","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-26T00:15:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XWBAZdrla1o41PZbTZAmHt6ba0Duwy4nG7x8aTokA/K9DYqpleJVjHbHU9KPEk/gDgGPx1s6VyDSdQj4XiNmCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-05T05:41:30.710858Z"},"content_sha256":"4cc81ce2a19f89c0f35f37b056d4693034da75f1c7ec3c570b29a387dc2835a6","schema_version":"1.0","event_id":"sha256:4cc81ce2a19f89c0f35f37b056d4693034da75f1c7ec3c570b29a387dc2835a6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:GZ6GYSFJZCGCZRRVAOLOEIACB2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"RMTL: Reinforced Micro-task Learning for Long-Horizon Manipulation with VLM Rewards","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.RO","authors_text":"An{\\i}l Can Ate\\c{s}, Cihan Topal, Orhan Kahraman","submitted_at":"2026-06-24T10:32:48Z","abstract_excerpt":"Reinforcement learning (RL) for robotic manipulation often requires manually designing a dense reward function, which is difficult to tune and often fragile, or learning a reward from human demonstrations or preferences, which can be expensive. A recent line of work uses pretrained vision-language models (VLMs) as zero-shot reward models, replacing these costs with a single text prompt. However, we argue that a single global prompt is too coarse for long-horizon manipulation tasks with randomized initial conditions. The single-prompt VLM reward is near-flat for much of the trajectory, making e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26175","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/2606.26175/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":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-26T00:15:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2VGyUnDC5p/oTQHnyzbHVwJQNmfGz/3bDadbOIpkh/d4mM9hOwL/otcJwj4JSE/OaRO8gFnFh+YvQFwiv3GgCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-05T05:41:30.711473Z"},"content_sha256":"7db48344ae75a0fe4beff60ca7f356c2db00a30281785917b0f54a0d1b481ae2","schema_version":"1.0","event_id":"sha256:7db48344ae75a0fe4beff60ca7f356c2db00a30281785917b0f54a0d1b481ae2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2/bundle.json","state_url":"https://pith.science/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2/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":"2026-07-05T05:41:30Z","links":{"resolver":"https://pith.science/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2","bundle":"https://pith.science/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2/bundle.json","state":"https://pith.science/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/GZ6GYSFJZCGCZRRVAOLOEIACB2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:GZ6GYSFJZCGCZRRVAOLOEIACB2","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":"22e241f9b63e474c6f4676415485681474f32d0993edbf010acd86f037648d81","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-24T10:32:48Z","title_canon_sha256":"a3d256e0d7716201718f02ed07ed7b58053d3f1e15d872d05cc904e2c9762c66"},"schema_version":"1.0","source":{"id":"2606.26175","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.26175","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"arxiv_version","alias_value":"2606.26175v1","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26175","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"pith_short_12","alias_value":"GZ6GYSFJZCGC","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"pith_short_16","alias_value":"GZ6GYSFJZCGCZRRV","created_at":"2026-06-26T00:15:28Z"},{"alias_kind":"pith_short_8","alias_value":"GZ6GYSFJ","created_at":"2026-06-26T00:15:28Z"}],"graph_snapshots":[{"event_id":"sha256:7db48344ae75a0fe4beff60ca7f356c2db00a30281785917b0f54a0d1b481ae2","target":"graph","created_at":"2026-06-26T00:15:28Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2606.26175/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement learning (RL) for robotic manipulation often requires manually designing a dense reward function, which is difficult to tune and often fragile, or learning a reward from human demonstrations or preferences, which can be expensive. A recent line of work uses pretrained vision-language models (VLMs) as zero-shot reward models, replacing these costs with a single text prompt. However, we argue that a single global prompt is too coarse for long-horizon manipulation tasks with randomized initial conditions. The single-prompt VLM reward is near-flat for much of the trajectory, making e","authors_text":"An{\\i}l Can Ate\\c{s}, Cihan Topal, Orhan Kahraman","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-24T10:32:48Z","title":"RMTL: Reinforced Micro-task Learning for Long-Horizon Manipulation with VLM Rewards"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26175","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:4cc81ce2a19f89c0f35f37b056d4693034da75f1c7ec3c570b29a387dc2835a6","target":"record","created_at":"2026-06-26T00:15:28Z","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":"22e241f9b63e474c6f4676415485681474f32d0993edbf010acd86f037648d81","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-06-24T10:32:48Z","title_canon_sha256":"a3d256e0d7716201718f02ed07ed7b58053d3f1e15d872d05cc904e2c9762c66"},"schema_version":"1.0","source":{"id":"2606.26175","kind":"arxiv","version":1}},"canonical_sha256":"367c6c48a9c88c2cc6350396e220020eacce41680c36723a250b048554e551de","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"367c6c48a9c88c2cc6350396e220020eacce41680c36723a250b048554e551de","first_computed_at":"2026-06-26T00:15:28.524116Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-26T00:15:28.524116Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2pmGHuifnRODRJRPJ9/rV7UFuY5Z6dG1tw0iHt84nM5tX1z6mdgTQPunekhQwezEgDeXLTS+o01vKoBXzhfdBQ==","signature_status":"signed_v1","signed_at":"2026-06-26T00:15:28.524527Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.26175","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4cc81ce2a19f89c0f35f37b056d4693034da75f1c7ec3c570b29a387dc2835a6","sha256:7db48344ae75a0fe4beff60ca7f356c2db00a30281785917b0f54a0d1b481ae2"],"state_sha256":"b2a36e382fb88081c62ede25b4773844544eaf8ddfed731f25a80b5f58fe82a8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tm/0csrJXbVoUhacPxP49TICsREdECx5Sjdf/62cyg9Afv6okfG/WoLCmS4y9Q88Lv1FHAes+HDZEVLiOP7GAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-05T05:41:30.714592Z","bundle_sha256":"1a50736b60e2cef4736a2b4d31c7d4a332b73ad4c00df7e8d81192aef339643f"}}