{"ID":2835612,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.23369","arxiv_id":"2511.23369","title":"SimScale: Learning to Drive via Real-World Simulation at Scale","abstract":"Achieving fully autonomous driving systems requires learning rational decisions in a wide span of scenarios, including safety-critical and out-of-distribution ones. However, such cases are underrepresented in real-world corpus collected by human experts. To complement for the lack of data diversity, we introduce a novel and scalable simulation framework capable of synthesizing massive unseen states upon existing driving logs. Our pipeline utilizes advanced neural rendering with a reactive environment to generate high-fidelity multi-view observations controlled by the perturbed ego trajectory. Furthermore, we develop a pseudo-expert trajectory generation mechanism for these newly simulated states to provide action supervision. Upon the synthesized data, we find that a simple co-training strategy on both real-world and simulated samples can lead to significant improvements in both robustness and generalization for various planning methods on challenging real-world benchmarks, up to +8.6 EPDMS on navhard and +2.9 on navtest. More importantly, such policy improvement scales smoothly by increasing simulation data only, even without extra real-world data streaming in. We further reveal several crucial findings of such a sim-real learning system, which we term SimScale, including the design of pseudo-experts and the scaling properties for different policy architectures. Simulation data and code have been released at https://github.com/OpenDriveLab/SimScale.","short_abstract":"Achieving fully autonomous driving systems requires learning rational decisions in a wide span of scenarios, including safety-critical and out-of-distribution ones. However, such cases are underrepresented in real-world corpus collected by human experts. To complement for the lack of data diversity, we introduce a nove...","url_abs":"https://arxiv.org/abs/2511.23369","url_pdf":"https://arxiv.org/pdf/2511.23369v3","authors":"[\"Haochen Tian\",\"Tianyu Li\",\"Haochen Liu\",\"Jiazhi Yang\",\"Yihang Qiu\",\"Guang Li\",\"Junli Wang\",\"Yinfeng Gao\",\"Zhang Zhang\",\"Liang Wang\",\"Hangjun Ye\",\"Tieniu Tan\",\"Long Chen\",\"Hongyang Li\"]","published":"2025-11-28T17:17:38Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.RO\"]","methods":"[]","has_code":false,"code_links":[{"ID":606523,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2835612,"paper_url":"https://arxiv.org/abs/2511.23369","paper_title":"SimScale: Learning to Drive via Real-World Simulation at Scale","repo_url":"https://github.com/OpenDriveLab/SimScale","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}