{"id":10867,"date":"2026-08-10T20:06:06","date_gmt":"2026-08-10T12:06:06","guid":{"rendered":"https:\/\/infernews.com\/blog\/world-action-model-for-end-to-end-autonomous-driving\/"},"modified":"2026-08-10T20:06:06","modified_gmt":"2026-08-10T12:06:06","slug":"world-action-model-for-end-to-end-autonomous-driving","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/world-action-model-for-end-to-end-autonomous-driving\/","title":{"rendered":"SimWAM\uff1a\u628a AI \u5f71\u7247\u8b8a\u6210\u81ea\u52d5\u99d5\u99db\u898f\u5283"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/08\/qualitative.jpg\" alt=\"Overview of the SimWAM architecture with isolated attention\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u83ef\u4e2d\u79d1\u6280\u5927\u5b78\u8207\u6771\u98a8\u7814\u7a76\u5718\u968a\u5408\u4f5c\u63d0\u51fa SimWAM\uff0c\u91cd\u9ede\u662f\u628a\u5f71\u7247\u751f\u6210\u7684\u80fd\u529b\u8f49\u6210\u81ea\u52d5\u99d5\u99db\u898f\u5283\u8a0a\u865f\uff0c\u800c\u4e0d\u662f\u5728\u63a8\u7406\u6642\u786c\u505a\u672a\u4f86\u5f71\u50cf\u751f\u6210\u3002\u5b83\u5c6c\u65bc World-Action Model (WAM)\uff0c\u76f4\u63a5\u8655\u7406\u7aef\u5230\u7aef\u81ea\u52d5\u99d5\u99db\u4e2d\u7684\u8ecc\u8de1\u9810\u6e2c\u8207\u898f\u5283\u554f\u984c\uff0c\u76ee\u6a19\u662f\u6e1b\u5c11\u5ef6\u9072\uff0c\u540c\u6642\u4fdd\u7559\u5c0d\u4ea4\u901a\u52d5\u614b\u7684\u7406\u89e3\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u8a13\u7df4\u6642\uff0c\u7cfb\u7d71\u6703\u628a\u9810\u8a13\u7df4\u5f71\u7247\u5c08\u5bb6\u548c\u8f15\u91cf\u52d5\u4f5c DiT \u4e00\u8d77\u505a joint flow matching\uff0c\u8b93\u672a\u4f86\u756b\u9762\u5b78\u5230\u7684\u52d5\u614b\u5148\u9a57\u6ef2\u5165\u52d5\u4f5c\u8868\u793a\u3002\u5169\u908a\u5171\u4eab\u6ce8\u610f\u529b\u6d41\uff0c\u4f46\u7528 isolated attention mask \u4ee4\u672a\u4f86\u7247\u6bb5\u548c\u52d5\u4f5c token \u4e0d\u6703\u4e92\u76f8\u5077\u770b\uff0c\u63a8\u7406\u6642\u5c31\u53ef\u4ee5\u76f4\u63a5\u4e1f\u6389\u5f71\u7247\u5206\u652f\uff0c\u53ea\u7559\u4e0b\u52d5\u4f5c\u5c08\u5bb6\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u7814\u7a76\u5718\u968a\u4f86\u81ea Huazhong University of Science &amp; Technology \u548c Dongfeng Research &amp; Development Institute\u3002<\/li>\n<li>\u63a8\u7406\u968e\u6bb5\u4e0d\u9700\u8981\u672a\u4f86\u5834\u666f\u751f\u6210\uff0c\u4e5f\u4e0d\u9700\u8981\u984d\u5916\u52d5\u4f5c\u6a21\u7d44\u3002<\/li>\n<li>\u52d5\u4f5c\u5c08\u5bb6\u53ef\u7368\u7acb\u7e2e\u653e\uff0c\u5f71\u7247\u9aa8\u5e79\u4ea6\u53ef\u66ff\u63db\uff0c\u7d50\u69cb\u76f8\u5c0d\u9748\u6d3b\u3002<\/li>\n<li>\u5728 NAVSIM \u4e0a\u53d6\u5f97 91.5 PDMS\uff0c\u4e26\u53ef zero-shot \u9077\u79fb\u5230 nuScenes\u3002<\/li>\n<li>\u5f37\u5316\u5b78\u7fd2\u90e8\u5206\u7528 FlowGRPO \u53bb\u512a\u5316\u7d44\u5408\u5f0f\u99d5\u99db\u734e\u52f5\uff0c\u53ea\u66f4\u65b0\u52d5\u4f5c\u5c08\u5bb6\u7684 LoRA adapters\u3002<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u5b83\u6700\u503c\u5f97\u7559\u610f\u7684\u5730\u65b9\uff0c\u4e0d\u5728\u65bc\u628a\u6a21\u578b\u5806\u5f97\u66f4\u5927\uff0c\u800c\u662f\u628a\u8a13\u7df4\u7528\u7684\u5f71\u7247\u8a0a\u865f\u548c\u90e8\u7f72\u7528\u7684\u898f\u5283\u5668\u5207\u958b\uff0c\u4ee4\u7cfb\u7d71\u5728\u4fdd\u7559\u52d5\u614b\u7406\u89e3\u7684\u540c\u6642\u6e1b\u5c11\u63a8\u7406\u8ca0\u64d4\u3002\u5b98\u65b9\u6c92\u6709\u5b8c\u6574\u5217\u51fa\u5b89\u88dd\u6b65\u9a5f\uff0c\u56e0\u6b64\u8f03\u9069\u5408\u5148\u6309\u8ad6\u6587\u8207\u7a0b\u5f0f\u5eab\u8aaa\u660e\u7406\u89e3\u5176\u8a13\u7df4\u548c\u90e8\u7f72\u6d41\u7a0b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/github.com\/H-EmbodVis\/SimWAM\" rel=\"noopener noreferrer\"><strong>GitHub<\/strong><\/a> \u00b7 <a href=\"https:\/\/huggingface.co\/H-EmbodVis\/SimWAM\" rel=\"noopener noreferrer\"><strong>\u6a21\u578b<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u83ef\u4e2d\u79d1\u5927\u8207\u6771\u98a8\u7814\u767c\u5718\u968a\u63d0\u51fa\u7684 SimWAM\uff0c\u5c07\u5f71\u7247\u5c08\u5bb6\u548c\u52d5\u4f5c\u5c08\u5bb6\u5171\u540c\u8a13\u7df4\uff0c\u518d\u4e1f\u6389\u5f71\u7247\u5206\u652f\u53ea\u4fdd\u7559\u898f\u5283\u5668\u3002\u5b83\u628a\u672a\u4f86\u756b\u9762\u5b78\u5230\u7684\u52d5\u614b\u5148\u9a57\uff0c\u8f49\u6210\u8f03\u4f4e\u5ef6\u9072\u7684\u8ecc\u8de1\u9810\u6e2c\u3002<\/p>\n","protected":false},"author":8,"featured_media":10866,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,149,189,204],"tags":[],"class_list":["post-10867","post","type-post","status-publish","format-standard","hentry","category-133","category-149","category-189","category-visionlanguageaction"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10867","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=10867"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/10867\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/10866"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=10867"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=10867"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=10867"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}