{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:5DWTKVQRY2T7OXWNJJTNUGXCSO","short_pith_number":"pith:5DWTKVQR","schema_version":"1.0","canonical_sha256":"e8ed355611c6a7f75ecd4a66da1ae2938aae05b975b7e873b3742f6821b24607","source":{"kind":"arxiv","id":"1804.00645","version":2},"attestation_state":"computed","paper":{"title":"Universal Planning Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.RO","stat.ML"],"primary_cat":"cs.LG","authors_text":"Allan Jabri, Aravind Srinivas, Chelsea Finn, Pieter Abbeel, Sergey Levine","submitted_at":"2018-04-02T17:51:53Z","abstract_excerpt":"A key challenge in complex visuomotor control is learning abstract representations that are effective for specifying goals, planning, and generalization. To this end, we introduce universal planning networks (UPN). UPNs embed differentiable planning within a goal-directed policy. This planning computation unrolls a forward model in a latent space and infers an optimal action plan through gradient descent trajectory optimization. The plan-by-gradient-descent process and its underlying representations are learned end-to-end to directly optimize a supervised imitation learning objective. We find "},"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":"1804.00645","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-04-02T17:51:53Z","cross_cats_sorted":["cs.AI","cs.CV","cs.RO","stat.ML"],"title_canon_sha256":"71999c00f19fe0e3e4feb5905539c91551beace3c3d2cea7763f32136df8bf99","abstract_canon_sha256":"026ce49aeb7e6f533823f9801708a81312d4c44ef1a643ea7022cd11372b2eab"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:19:16.157945Z","signature_b64":"shYxqD4hr3bT3eIC8CONwHCD1tNvLuhn/UFbJDpxJ4glgQVG3Z9eTSlGkdsuwEKb2urr77P6K528oBQL7bpUCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e8ed355611c6a7f75ecd4a66da1ae2938aae05b975b7e873b3742f6821b24607","last_reissued_at":"2026-05-18T00:19:16.157263Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:19:16.157263Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Universal Planning Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.RO","stat.ML"],"primary_cat":"cs.LG","authors_text":"Allan Jabri, Aravind Srinivas, Chelsea Finn, Pieter Abbeel, Sergey Levine","submitted_at":"2018-04-02T17:51:53Z","abstract_excerpt":"A key challenge in complex visuomotor control is learning abstract representations that are effective for specifying goals, planning, and generalization. To this end, we introduce universal planning networks (UPN). UPNs embed differentiable planning within a goal-directed policy. This planning computation unrolls a forward model in a latent space and infers an optimal action plan through gradient descent trajectory optimization. The plan-by-gradient-descent process and its underlying representations are learned end-to-end to directly optimize a supervised imitation learning objective. We find "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1804.00645","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":""},"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":"1804.00645","created_at":"2026-05-18T00:19:16.157387+00:00"},{"alias_kind":"arxiv_version","alias_value":"1804.00645v2","created_at":"2026-05-18T00:19:16.157387+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1804.00645","created_at":"2026-05-18T00:19:16.157387+00:00"},{"alias_kind":"pith_short_12","alias_value":"5DWTKVQRY2T7","created_at":"2026-05-18T12:32:08.215937+00:00"},{"alias_kind":"pith_short_16","alias_value":"5DWTKVQRY2T7OXWN","created_at":"2026-05-18T12:32:08.215937+00:00"},{"alias_kind":"pith_short_8","alias_value":"5DWTKVQR","created_at":"2026-05-18T12:32:08.215937+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"1906.09624","citing_title":"On the Feasibility of Learning, Rather than Assuming, Human Biases for Reward Inference","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"1906.01820","citing_title":"Risks from Learned Optimization in Advanced Machine Learning Systems","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2010.02193","citing_title":"Mastering Atari with Discrete World Models","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"1912.01603","citing_title":"Dream to Control: Learning Behaviors by Latent Imagination","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO","json":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO.json","graph_json":"https://pith.science/api/pith-number/5DWTKVQRY2T7OXWNJJTNUGXCSO/graph.json","events_json":"https://pith.science/api/pith-number/5DWTKVQRY2T7OXWNJJTNUGXCSO/events.json","paper":"https://pith.science/paper/5DWTKVQR"},"agent_actions":{"view_html":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO","download_json":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO.json","view_paper":"https://pith.science/paper/5DWTKVQR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1804.00645&json=true","fetch_graph":"https://pith.science/api/pith-number/5DWTKVQRY2T7OXWNJJTNUGXCSO/graph.json","fetch_events":"https://pith.science/api/pith-number/5DWTKVQRY2T7OXWNJJTNUGXCSO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO/action/storage_attestation","attest_author":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO/action/author_attestation","sign_citation":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO/action/citation_signature","submit_replication":"https://pith.science/pith/5DWTKVQRY2T7OXWNJJTNUGXCSO/action/replication_record"}},"created_at":"2026-05-18T00:19:16.157387+00:00","updated_at":"2026-05-18T00:19:16.157387+00:00"}