{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2016:BHOWV3D6AQAVD2526IS5OIOHPV","short_pith_number":"pith:BHOWV3D6","schema_version":"1.0","canonical_sha256":"09dd6aec7e040151ebbaf225d721c77d75cdc1636f0234d10916745b722d52ce","source":{"kind":"arxiv","id":"1611.01578","version":2},"attestation_state":"computed","paper":{"title":"Neural Architecture Search with Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.NE"],"primary_cat":"cs.LG","authors_text":"Barret Zoph, Quoc V. Le","submitted_at":"2016-11-05T00:41:37Z","abstract_excerpt":"Neural networks are powerful and flexible models that work well for many difficult learning tasks in image, speech and natural language understanding. Despite their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set. On the CIFAR-10 dataset, our method, starting from scratch, can design a novel network architecture that rivals the best human-invented architecture in ter"},"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":"1611.01578","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2016-11-05T00:41:37Z","cross_cats_sorted":["cs.AI","cs.NE"],"title_canon_sha256":"9d648c1b3ad6bd59423ea0fca91598684cf9101473fe7ab3cacfb8c4c475157a","abstract_canon_sha256":"38e92b38fb05a04fc848bbaed5304c60c63a467143233ee13306b23c1ee6838f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:50:42.333661Z","signature_b64":"HYsDS+cYKxcLjxMFdmNJFBkc2M5ESaH2ChzZj5nw6Pk7+dYTwBGEc2Gl4lLqYTfSaaxYxNKt63mej2IhyA/aAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"09dd6aec7e040151ebbaf225d721c77d75cdc1636f0234d10916745b722d52ce","last_reissued_at":"2026-05-18T00:50:42.332950Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:50:42.332950Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neural Architecture Search with Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.NE"],"primary_cat":"cs.LG","authors_text":"Barret Zoph, Quoc V. Le","submitted_at":"2016-11-05T00:41:37Z","abstract_excerpt":"Neural networks are powerful and flexible models that work well for many difficult learning tasks in image, speech and natural language understanding. Despite their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set. On the CIFAR-10 dataset, our method, starting from scratch, can design a novel network architecture that rivals the best human-invented architecture in ter"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1611.01578","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":"1611.01578","created_at":"2026-05-18T00:50:42.333062+00:00"},{"alias_kind":"arxiv_version","alias_value":"1611.01578v2","created_at":"2026-05-18T00:50:42.333062+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1611.01578","created_at":"2026-05-18T00:50:42.333062+00:00"},{"alias_kind":"pith_short_12","alias_value":"BHOWV3D6AQAV","created_at":"2026-05-18T12:30:07.202191+00:00"},{"alias_kind":"pith_short_16","alias_value":"BHOWV3D6AQAVD252","created_at":"2026-05-18T12:30:07.202191+00:00"},{"alias_kind":"pith_short_8","alias_value":"BHOWV3D6","created_at":"2026-05-18T12:30:07.202191+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":19,"internal_anchor_count":8,"sample":[{"citing_arxiv_id":"2605.17046","citing_title":"1GC-7RC: One Graphic Card -- Seven Research Challenges! How Good Are AI Agents at Doing Your Job?","ref_index":57,"is_internal_anchor":true},{"citing_arxiv_id":"2507.20906","citing_title":"Soft Head Selection for Injecting ICL-Derived Task Embeddings","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2509.21010","citing_title":"Bridging the phenotype-target gap for molecular generation via multi-objective reinforcement learning","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2510.14235","citing_title":"Spiking Neural Network Architecture Search: A Survey","ref_index":78,"is_internal_anchor":true},{"citing_arxiv_id":"2512.02459","citing_title":"Neural Architecture Search of Time-to-First-Spike-Coded Spiking Neural Networks for Efficient Eye-based Emotion Recognition","ref_index":43,"is_internal_anchor":true},{"citing_arxiv_id":"2512.12448","citing_title":"Optimized Architectures for Kolmogorov-Arnold Networks","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2601.15127","citing_title":"DeepFedNAS: Efficient Hardware-Aware Architecture Adaptation for Heterogeneous IoT Federations via Pareto-Guided Supernet Training","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00830","citing_title":"Efficient Accelerated Graph Edit Distance Computation on GPU","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08083","citing_title":"LLMs Improving LLMs: Agentic Discovery for Test-Time Scaling","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26211","citing_title":"OMEGA: Optimizing Machine Learning by Evaluating Generated Algorithms","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00970","citing_title":"Split and Aggregation Learning for Foundation Models Over Mobile Embodied AI Network (MEAN): A Comprehensive Survey","ref_index":119,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12005","citing_title":"BayMOTH: Bayesian optiMizatiOn with meTa-lookahead -- a simple approacH","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10560","citing_title":"Heterogeneous Connectivity in Sparse Networks: Fan-in Profiles, Gradient Hierarchy, and Topological Equilibria","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.07526","citing_title":"From LLM to Silicon: RL-Driven ASIC Architecture Exploration for On-Device AI Inference","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08083","citing_title":"LLMs Improving LLMs: Agentic Discovery for Test-Time Scaling","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07379","citing_title":"RELO: Reinforcement Learning to Localize for Visual Object Tracking","ref_index":252,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05550","citing_title":"AutoSOTA: An End-to-End Automated Research System for State-of-the-Art AI Model Discovery","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2410.02713","citing_title":"LLaVA-Video: Video Instruction Tuning With Synthetic Data","ref_index":90,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16228","citing_title":"TRON: Trainable, architecture-reconfigurable random optical neural networks","ref_index":63,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV","json":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV.json","graph_json":"https://pith.science/api/pith-number/BHOWV3D6AQAVD2526IS5OIOHPV/graph.json","events_json":"https://pith.science/api/pith-number/BHOWV3D6AQAVD2526IS5OIOHPV/events.json","paper":"https://pith.science/paper/BHOWV3D6"},"agent_actions":{"view_html":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV","download_json":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV.json","view_paper":"https://pith.science/paper/BHOWV3D6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1611.01578&json=true","fetch_graph":"https://pith.science/api/pith-number/BHOWV3D6AQAVD2526IS5OIOHPV/graph.json","fetch_events":"https://pith.science/api/pith-number/BHOWV3D6AQAVD2526IS5OIOHPV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV/action/storage_attestation","attest_author":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV/action/author_attestation","sign_citation":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV/action/citation_signature","submit_replication":"https://pith.science/pith/BHOWV3D6AQAVD2526IS5OIOHPV/action/replication_record"}},"created_at":"2026-05-18T00:50:42.333062+00:00","updated_at":"2026-05-18T00:50:42.333062+00:00"}