{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/rethinking-self-driving-multi-task-knowledge-1","title":"Rethinking Self-driving: Multi-task Knowledge for Better Generalization and Accident Explanation Ability","arxiv_id":"1809.11100","date":"2018-09-28","proceeding":null,"authors":["Zhihao Li","Toshiyuki Motoyoshi","Kazuma Sasaki","Tetsuya OGATA","Shigeki SUGANO"],"abstract":"Current end-to-end deep learning driving models have two problems: (1) Poor\ngeneralization ability of unobserved driving environment when diversity of\ntraining driving dataset is limited (2) Lack of accident explanation ability\nwhen driving models don't work as expected. To tackle these two problems,\nrooted on the believe that knowledge of associated easy task is benificial for\naddressing difficult task, we proposed a new driving model which is composed of\nperception module for \\textit{see and think} and driving module for\n\\textit{behave}, and trained it with multi-task perception-related basic\nknowledge and driving knowledge stepwisely.\n  Specifically segmentation map and depth map (pixel level understanding of\nimages) were considered as \\textit{what \\& where} and \\textit{how far}\nknowledge for tackling easier driving-related perception problems before\ngenerating final control commands for difficult driving task. The results of\nexperiments demonstrated the effectiveness of multi-task perception knowledge\nfor better generalization and accident explanation ability. With our method the\naverage sucess rate of finishing most difficult navigation tasks in untrained\ncity of CoRL test surpassed current benchmark method for 15 percent in trained\nweather and 20 percent in untrained weathers. Demonstration video link is:\nhttps://www.youtube.com/watch?v=N7ePnnZZwdE","url_abs":"http://arxiv.org/abs/1809.11100v1","url_pdf":"http://arxiv.org/pdf/1809.11100v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"rethinking-self-driving-multi-task-knowledge-1","repo_url":"https://github.com/jackspp/rethinking-self-driving","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"diversity","task_name":"Diversity"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1809.11100","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}