{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MZ2Q6ZWWEBZ2M4XGYHF72E5SOH","short_pith_number":"pith:MZ2Q6ZWW","schema_version":"1.0","canonical_sha256":"66750f66d62073a672e6c1cbfd13b271d5568d74846ccc3c164489e69622a1e9","source":{"kind":"arxiv","id":"2503.07608","version":1},"attestation_state":"computed","paper":{"title":"AlphaDrive: Unleashing the Power of VLMs in Autonomous Driving via Reinforcement Learning and Reasoning","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Reinforcement learning with tailored rewards and a two-stage strategy improves vision-language models for autonomous driving planning.","cross_cats":["cs.RO"],"primary_cat":"cs.CV","authors_text":"Bo Jiang, Qian Zhang, Shaoyu Chen, Wenyu Liu, Xinggang Wang","submitted_at":"2025-03-10T17:59:42Z","abstract_excerpt":"OpenAI o1 and DeepSeek R1 achieve or even surpass human expert-level performance in complex domains like mathematics and science, with reinforcement learning (RL) and reasoning playing a crucial role. In autonomous driving, recent end-to-end models have greatly improved planning performance but still struggle with long-tailed problems due to limited common sense and reasoning abilities. Some studies integrate vision-language models (VLMs) into autonomous driving, but they typically rely on pre-trained models with simple supervised fine-tuning (SFT) on driving data, without further exploration "},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2503.07608","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-10T17:59:42Z","cross_cats_sorted":["cs.RO"],"title_canon_sha256":"ffbef9e71aead583bc8fd0ecc333658dd73e98f7124c6a20764b1750183a9904","abstract_canon_sha256":"251251be26a88dc4ba247abe36d2d3fb3589f84bb6d3ebad2a96ee22e776a2fb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:46.739040Z","signature_b64":"Ogl9N7mZ0au+lknR6+T89Dx0a5IZoaNfXd0YuwxTL1O/PBjeLZ7Uf4QTKzcD2O+c8DLH3DGAnCvM0FpNbLoIAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66750f66d62073a672e6c1cbfd13b271d5568d74846ccc3c164489e69622a1e9","last_reissued_at":"2026-05-17T23:38:46.738477Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:46.738477Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AlphaDrive: Unleashing the Power of VLMs in Autonomous Driving via Reinforcement Learning and Reasoning","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"Reinforcement learning with tailored rewards and a two-stage strategy improves vision-language models for autonomous driving planning.","cross_cats":["cs.RO"],"primary_cat":"cs.CV","authors_text":"Bo Jiang, Qian Zhang, Shaoyu Chen, Wenyu Liu, Xinggang Wang","submitted_at":"2025-03-10T17:59:42Z","abstract_excerpt":"OpenAI o1 and DeepSeek R1 achieve or even surpass human expert-level performance in complex domains like mathematics and science, with reinforcement learning (RL) and reasoning playing a crucial role. In autonomous driving, recent end-to-end models have greatly improved planning performance but still struggle with long-tailed problems due to limited common sense and reasoning abilities. Some studies integrate vision-language models (VLMs) into autonomous driving, but they typically rely on pre-trained models with simple supervised fine-tuning (SFT) on driving data, without further exploration "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"AlphaDrive significantly improves both planning performance and training efficiency compared to using only SFT or without reasoning, and following RL training exhibits emergent multimodal planning capabilities.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the four GRPO-based RL rewards and two-stage training strategy produce generalizable, safe improvements on real-world driving data rather than overfitting to the training distribution.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"AlphaDrive uses GRPO-based RL rewards and two-stage SFT+RL training on VLMs to improve autonomous driving planning performance and efficiency while producing emergent multimodal capabilities.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Reinforcement learning with tailored rewards and a two-stage strategy improves vision-language models for autonomous driving planning.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"0f635f6f218a8fe37b2ed36a843b8b0ed939e4aa0dd1f623f4fea8311164831e"},"source":{"id":"2503.07608","kind":"arxiv","version":1},"verdict":{"id":"8b8c67fa-e1a6-49b9-b03a-3a5416fde7cf","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T20:03:30.788130Z","strongest_claim":"AlphaDrive significantly improves both planning performance and training efficiency compared to using only SFT or without reasoning, and following RL training exhibits emergent multimodal planning capabilities.","one_line_summary":"AlphaDrive uses GRPO-based RL rewards and two-stage SFT+RL training on VLMs to improve autonomous driving planning performance and efficiency while producing emergent multimodal capabilities.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the four GRPO-based RL rewards and two-stage training strategy produce generalizable, safe improvements on real-world driving data rather than overfitting to the training distribution.","pith_extraction_headline":"Reinforcement learning with tailored rewards and a two-stage strategy improves vision-language models for autonomous driving planning."},"references":{"count":49,"sample":[{"doi":"","year":null,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2022,"title":"Flamingo: a visual language model for few-shot learning","work_id":"2d29aa49-7f72-4532-8c66-e33ed3d6d8a8","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond","work_id":"cbc2bb21-b6bb-46c0-80bf-107e195ffe10","ref_index":3,"cited_arxiv_id":"2308.12966","is_internal_anchor":true},{"doi":"","year":2005,"title":"Meteor: An automatic metric for mt evaluation with improved correlation with hu- man judgments","work_id":"90de9967-cc22-427f-91fb-ed50f063376c","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control","work_id":"ff438a8a-8003-4fae-9131-acd418b3597b","ref_index":5,"cited_arxiv_id":"2307.15818","is_internal_anchor":true}],"resolved_work":49,"snapshot_sha256":"a80272b7b288b32d0835e4f514004ed6eafe32c0ace5e6a9eca451ddc76446f5","internal_anchors":16},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4050fc7d31763a3a0ce57228bfdcbd91b98fa13dafc1b3731d305dae64b84142"},"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":"2503.07608","created_at":"2026-05-17T23:38:46.738583+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.07608v1","created_at":"2026-05-17T23:38:46.738583+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.07608","created_at":"2026-05-17T23:38:46.738583+00:00"},{"alias_kind":"pith_short_12","alias_value":"MZ2Q6ZWWEBZ2","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"MZ2Q6ZWWEBZ2M4XG","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"MZ2Q6ZWW","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":26,"internal_anchor_count":26,"sample":[{"citing_arxiv_id":"2605.21139","citing_title":"Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2604.06750","citing_title":"How Well Do Vision-Language Models Understand Sequential Driving Scenes? A Sensitivity Study","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14201","citing_title":"MAPLE: Latent Multi-Agent Play for End-to-End Autonomous Driving","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2605.21139","citing_title":"Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.13646","citing_title":"Causality-Aware End-to-End Autonomous Driving via Ego-Centric Joint Scene Modeling","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19524","citing_title":"SafeAlign-VLA: A Negative-Enhanced Safe Alignment Framework for Risk-Aware Autonomous Driving","ref_index":31,"is_internal_anchor":true},{"citing_arxiv_id":"2507.04049","citing_title":"DIVER: Reinforced Diffusion Breaks Imitation Bottlenecks in End-to-End Autonomous Driving","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2507.01925","citing_title":"A Survey on Vision-Language-Action Models: An Action Tokenization Perspective","ref_index":279,"is_internal_anchor":true},{"citing_arxiv_id":"2602.09514","citing_title":"EcoGym: Evaluating LLMs for Long-Horizon Plan-and-Execute in Interactive Economies","ref_index":15,"is_internal_anchor":true},{"citing_arxiv_id":"2603.13842","citing_title":"Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14201","citing_title":"MAPLE: Latent Multi-Agent Play for End-to-End Autonomous Driving","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2506.13757","citing_title":"AutoVLA: A Vision-Language-Action Model for End-to-End Autonomous Driving with Adaptive Reasoning and Reinforcement Fine-Tuning","ref_index":38,"is_internal_anchor":true},{"citing_arxiv_id":"2605.13646","citing_title":"Causality-Aware End-to-End Autonomous Driving via Ego-Centric Joint Scene Modeling","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2604.00813","citing_title":"DVGT-2: Vision-Geometry-Action Model for Autonomous Driving at Scale","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2604.03497","citing_title":"Sim2Real-AD: A Modular Sim-to-Real Framework for Deploying VLM-Guided Reinforcement Learning in Real-World Autonomous Driving","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10564","citing_title":"DeepSight: Long-Horizon World Modeling via Latent States Prediction for End-to-End Autonomous Driving","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2604.11734","citing_title":"SCORP: Scene-Consistent Multi-agent Diffusion Planning with Stable Online Reinforcement Post-Training for Cooperative Driving","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2604.19710","citing_title":"SpanVLA: Efficient Action Bridging and Learning from Negative-Recovery Samples for Vision-Language-Action Model","ref_index":32,"is_internal_anchor":true},{"citing_arxiv_id":"2604.11734","citing_title":"SCORP: Scene-Consistent Multi-agent Diffusion Planning with Stable Online Reinforcement Post-Training for Cooperative Driving","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2604.12656","citing_title":"FeaXDrive: Feasibility-aware Trajectory-Centric Diffusion Planning for End-to-End Autonomous Driving","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2604.09059","citing_title":"Learning Vision-Language-Action World Models for Autonomous Driving","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2604.08031","citing_title":"Open-Ended Instruction Realization with LLM-Enabled Multi-Planner Scheduling in Autonomous Vehicles","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2604.06750","citing_title":"How Well Do Vision-Language Models Understand Sequential Driving Scenes? A Sensitivity Study","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2604.04857","citing_title":"The Blind Spot of Adaptation: Quantifying and Mitigating Forgetting in Fine-tuned Driving Models","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15308","citing_title":"RAD-2: Scaling Reinforcement Learning in a Generator-Discriminator Framework","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH","json":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH.json","graph_json":"https://pith.science/api/pith-number/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/graph.json","events_json":"https://pith.science/api/pith-number/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/events.json","paper":"https://pith.science/paper/MZ2Q6ZWW"},"agent_actions":{"view_html":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH","download_json":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH.json","view_paper":"https://pith.science/paper/MZ2Q6ZWW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.07608&json=true","fetch_graph":"https://pith.science/api/pith-number/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/graph.json","fetch_events":"https://pith.science/api/pith-number/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/action/storage_attestation","attest_author":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/action/author_attestation","sign_citation":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/action/citation_signature","submit_replication":"https://pith.science/pith/MZ2Q6ZWWEBZ2M4XGYHF72E5SOH/action/replication_record"}},"created_at":"2026-05-17T23:38:46.738583+00:00","updated_at":"2026-05-17T23:38:46.738583+00:00"}