{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:BAUIC6QIT7OE5RFVPM5ZKH3PUW","short_pith_number":"pith:BAUIC6QI","schema_version":"1.0","canonical_sha256":"0828817a089fdc4ec4b57b3b951f6fa5b05584877cf24e14824c4dede426a8db","source":{"kind":"arxiv","id":"2002.11963","version":1},"attestation_state":"computed","paper":{"title":"Plannable Approximations to MDP Homomorphisms: Equivariance under Actions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Elise van der Pol, Frans A. Oliehoek, Max Welling, Thomas Kipf","submitted_at":"2020-02-27T08:29:10Z","abstract_excerpt":"This work exploits action equivariance for representation learning in reinforcement learning. Equivariance under actions states that transitions in the input space are mirrored by equivalent transitions in latent space, while the map and transition functions should also commute. We introduce a contrastive loss function that enforces action equivariance on the learned representations. We prove that when our loss is zero, we have a homomorphism of a deterministic Markov Decision Process (MDP). Learning equivariant maps leads to structured latent spaces, allowing us to build a model on which we p"},"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":"2002.11963","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-27T08:29:10Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"64863a09c3e230d752c85ad005ca04c18a6e3605dd711f7f67404303f0dd901c","abstract_canon_sha256":"edf62e32d60b9f628ee0ec68c7896ba89114c33c064462338bbdae6c03f568d4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:44:10.574456Z","signature_b64":"wYejrNw1fa72ryDmVwKvt9eBu8cKZ1gHxqVBYHfCuX62Ycz0fc2pOc3E4+1bdGDPfNaKSpIwT51Bhsu22GxnBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0828817a089fdc4ec4b57b3b951f6fa5b05584877cf24e14824c4dede426a8db","last_reissued_at":"2026-07-05T00:44:10.574007Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:44:10.574007Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Plannable Approximations to MDP Homomorphisms: Equivariance under Actions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Elise van der Pol, Frans A. Oliehoek, Max Welling, Thomas Kipf","submitted_at":"2020-02-27T08:29:10Z","abstract_excerpt":"This work exploits action equivariance for representation learning in reinforcement learning. Equivariance under actions states that transitions in the input space are mirrored by equivalent transitions in latent space, while the map and transition functions should also commute. We introduce a contrastive loss function that enforces action equivariance on the learned representations. We prove that when our loss is zero, we have a homomorphism of a deterministic Markov Decision Process (MDP). Learning equivariant maps leads to structured latent spaces, allowing us to build a model on which we p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.11963","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/2002.11963/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":"2002.11963","created_at":"2026-07-05T00:44:10.574064+00:00"},{"alias_kind":"arxiv_version","alias_value":"2002.11963v1","created_at":"2026-07-05T00:44:10.574064+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.11963","created_at":"2026-07-05T00:44:10.574064+00:00"},{"alias_kind":"pith_short_12","alias_value":"BAUIC6QIT7OE","created_at":"2026-07-05T00:44:10.574064+00:00"},{"alias_kind":"pith_short_16","alias_value":"BAUIC6QIT7OE5RFV","created_at":"2026-07-05T00:44:10.574064+00:00"},{"alias_kind":"pith_short_8","alias_value":"BAUIC6QI","created_at":"2026-07-05T00:44:10.574064+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.09035","citing_title":"Advantage-Guided Diffusion for Model-Based Reinforcement Learning","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW","json":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW.json","graph_json":"https://pith.science/api/pith-number/BAUIC6QIT7OE5RFVPM5ZKH3PUW/graph.json","events_json":"https://pith.science/api/pith-number/BAUIC6QIT7OE5RFVPM5ZKH3PUW/events.json","paper":"https://pith.science/paper/BAUIC6QI"},"agent_actions":{"view_html":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW","download_json":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW.json","view_paper":"https://pith.science/paper/BAUIC6QI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2002.11963&json=true","fetch_graph":"https://pith.science/api/pith-number/BAUIC6QIT7OE5RFVPM5ZKH3PUW/graph.json","fetch_events":"https://pith.science/api/pith-number/BAUIC6QIT7OE5RFVPM5ZKH3PUW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW/action/storage_attestation","attest_author":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW/action/author_attestation","sign_citation":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW/action/citation_signature","submit_replication":"https://pith.science/pith/BAUIC6QIT7OE5RFVPM5ZKH3PUW/action/replication_record"}},"created_at":"2026-07-05T00:44:10.574064+00:00","updated_at":"2026-07-05T00:44:10.574064+00:00"}