{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:UUI7464NP7SIJQ3YN7YO45DTHS","short_pith_number":"pith:UUI7464N","schema_version":"1.0","canonical_sha256":"a511fe7b8d7fe484c3786ff0ee74733c9d97f7217333d7a84aea7b488f7443bc","source":{"kind":"arxiv","id":"2302.11793","version":2},"attestation_state":"computed","paper":{"title":"Revisiting the Gumbel-Softmax in MADDPG","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.MA","stat.ML"],"primary_cat":"cs.LG","authors_text":"Callum Rhys Tilbury, Filippos Christianos, Stefano V. Albrecht","submitted_at":"2023-02-23T06:13:51Z","abstract_excerpt":"MADDPG is an algorithm in multi-agent reinforcement learning (MARL) that extends the popular single-agent method, DDPG, to multi-agent scenarios. Importantly, DDPG is an algorithm designed for continuous action spaces, where the gradient of the state-action value function exists. For this algorithm to work in discrete action spaces, discrete gradient estimation must be performed. For MADDPG, the Gumbel-Softmax (GS) estimator is used -- a reparameterisation which relaxes a discrete distribution into a similar continuous one. This method, however, is statistically biased, and a recent MARL bench"},"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":"2302.11793","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-02-23T06:13:51Z","cross_cats_sorted":["cs.AI","cs.MA","stat.ML"],"title_canon_sha256":"46b786613518f3dd495fde55993eef2c5f23cb331977f7da9441d92ce8a9dac8","abstract_canon_sha256":"277f26dd1a848ea4a116f2899c2b9e398fb6f77919c7015687666c1a731a1745"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:20:41.007357Z","signature_b64":"AHD3kUO5oSrDxwIgqMTp4g/LPTgrvPbnRI+JVBMhkbSX4EypjIcZRxbav3llTCEJoXAy6MnhVMo5tHVhgVb2Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a511fe7b8d7fe484c3786ff0ee74733c9d97f7217333d7a84aea7b488f7443bc","last_reissued_at":"2026-07-05T06:20:41.006945Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:20:41.006945Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Revisiting the Gumbel-Softmax in MADDPG","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.MA","stat.ML"],"primary_cat":"cs.LG","authors_text":"Callum Rhys Tilbury, Filippos Christianos, Stefano V. Albrecht","submitted_at":"2023-02-23T06:13:51Z","abstract_excerpt":"MADDPG is an algorithm in multi-agent reinforcement learning (MARL) that extends the popular single-agent method, DDPG, to multi-agent scenarios. Importantly, DDPG is an algorithm designed for continuous action spaces, where the gradient of the state-action value function exists. For this algorithm to work in discrete action spaces, discrete gradient estimation must be performed. For MADDPG, the Gumbel-Softmax (GS) estimator is used -- a reparameterisation which relaxes a discrete distribution into a similar continuous one. This method, however, is statistically biased, and a recent MARL bench"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.11793","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/2302.11793/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":"2302.11793","created_at":"2026-07-05T06:20:41.007000+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.11793v2","created_at":"2026-07-05T06:20:41.007000+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.11793","created_at":"2026-07-05T06:20:41.007000+00:00"},{"alias_kind":"pith_short_12","alias_value":"UUI7464NP7SI","created_at":"2026-07-05T06:20:41.007000+00:00"},{"alias_kind":"pith_short_16","alias_value":"UUI7464NP7SIJQ3Y","created_at":"2026-07-05T06:20:41.007000+00:00"},{"alias_kind":"pith_short_8","alias_value":"UUI7464N","created_at":"2026-07-05T06:20:41.007000+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS","json":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS.json","graph_json":"https://pith.science/api/pith-number/UUI7464NP7SIJQ3YN7YO45DTHS/graph.json","events_json":"https://pith.science/api/pith-number/UUI7464NP7SIJQ3YN7YO45DTHS/events.json","paper":"https://pith.science/paper/UUI7464N"},"agent_actions":{"view_html":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS","download_json":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS.json","view_paper":"https://pith.science/paper/UUI7464N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.11793&json=true","fetch_graph":"https://pith.science/api/pith-number/UUI7464NP7SIJQ3YN7YO45DTHS/graph.json","fetch_events":"https://pith.science/api/pith-number/UUI7464NP7SIJQ3YN7YO45DTHS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS/action/storage_attestation","attest_author":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS/action/author_attestation","sign_citation":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS/action/citation_signature","submit_replication":"https://pith.science/pith/UUI7464NP7SIJQ3YN7YO45DTHS/action/replication_record"}},"created_at":"2026-07-05T06:20:41.007000+00:00","updated_at":"2026-07-05T06:20:41.007000+00:00"}