{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:KX6KFLCS75SI37SHTHNOFITV7M","short_pith_number":"pith:KX6KFLCS","schema_version":"1.0","canonical_sha256":"55fca2ac52ff648dfe4799dae2a275fb13aeb1bb03dff4224d3a1a45216ebb38","source":{"kind":"arxiv","id":"2102.10336","version":2},"attestation_state":"computed","paper":{"title":"Physical Reasoning Using Dynamics-Aware Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Anton Bakhtin, Eltayeb Ahmed, Laurens van der Maaten, Rohit Girdhar","submitted_at":"2021-02-20T12:56:16Z","abstract_excerpt":"A common approach to solving physical reasoning tasks is to train a value learner on example tasks. A limitation of such an approach is that it requires learning about object dynamics solely from reward values assigned to the final state of a rollout of the environment. This study aims to address this limitation by augmenting the reward value with self-supervised signals about object dynamics. Specifically, we train the model to characterize the similarity of two environment rollouts, jointly with predicting the outcome of the reasoning task. This similarity can be defined as a distance measur"},"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":"2102.10336","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2021-02-20T12:56:16Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"ab7f4bb76304f3784c3d9f0ffa7634a2bf774540b9beca80830fe7f5e00c1645","abstract_canon_sha256":"7cb7b9100eb4844cd1c9b80ccc043a981cabc71981d6a9a1df518c30068bd574"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:10:52.191415Z","signature_b64":"+uDw/SndcJRUStWZHFBTcABFsjJo0PHKktgdO/THOoE0cFO4tBgXdoQhh9vyJGa2Qx2A092vCIwrgC+8JP8RBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"55fca2ac52ff648dfe4799dae2a275fb13aeb1bb03dff4224d3a1a45216ebb38","last_reissued_at":"2026-07-05T03:10:52.191011Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:10:52.191011Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Physical Reasoning Using Dynamics-Aware Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Anton Bakhtin, Eltayeb Ahmed, Laurens van der Maaten, Rohit Girdhar","submitted_at":"2021-02-20T12:56:16Z","abstract_excerpt":"A common approach to solving physical reasoning tasks is to train a value learner on example tasks. A limitation of such an approach is that it requires learning about object dynamics solely from reward values assigned to the final state of a rollout of the environment. This study aims to address this limitation by augmenting the reward value with self-supervised signals about object dynamics. Specifically, we train the model to characterize the similarity of two environment rollouts, jointly with predicting the outcome of the reasoning task. This similarity can be defined as a distance measur"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.10336","kind":"arxiv","version":2},"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/2102.10336/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":"2102.10336","created_at":"2026-07-05T03:10:52.191073+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.10336v2","created_at":"2026-07-05T03:10:52.191073+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.10336","created_at":"2026-07-05T03:10:52.191073+00:00"},{"alias_kind":"pith_short_12","alias_value":"KX6KFLCS75SI","created_at":"2026-07-05T03:10:52.191073+00:00"},{"alias_kind":"pith_short_16","alias_value":"KX6KFLCS75SI37SH","created_at":"2026-07-05T03:10:52.191073+00:00"},{"alias_kind":"pith_short_8","alias_value":"KX6KFLCS","created_at":"2026-07-05T03:10:52.191073+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06522","citing_title":"Bridging Physical Reasoning and Task Generalization via Visual Action Outcome Reasoning Alignment","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M","json":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M.json","graph_json":"https://pith.science/api/pith-number/KX6KFLCS75SI37SHTHNOFITV7M/graph.json","events_json":"https://pith.science/api/pith-number/KX6KFLCS75SI37SHTHNOFITV7M/events.json","paper":"https://pith.science/paper/KX6KFLCS"},"agent_actions":{"view_html":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M","download_json":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M.json","view_paper":"https://pith.science/paper/KX6KFLCS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.10336&json=true","fetch_graph":"https://pith.science/api/pith-number/KX6KFLCS75SI37SHTHNOFITV7M/graph.json","fetch_events":"https://pith.science/api/pith-number/KX6KFLCS75SI37SHTHNOFITV7M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M/action/storage_attestation","attest_author":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M/action/author_attestation","sign_citation":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M/action/citation_signature","submit_replication":"https://pith.science/pith/KX6KFLCS75SI37SHTHNOFITV7M/action/replication_record"}},"created_at":"2026-07-05T03:10:52.191073+00:00","updated_at":"2026-07-05T03:10:52.191073+00:00"}