{"id":92,"date":"2025-12-11T20:25:54","date_gmt":"2025-12-11T20:25:54","guid":{"rendered":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/?page_id=92"},"modified":"2025-12-12T01:27:15","modified_gmt":"2025-12-12T01:27:15","slug":"future-work","status":"publish","type":"page","link":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/","title":{"rendered":"Future Work"},"content":{"rendered":"\n<p><strong>1. VLM-Guided Noise Steering<\/strong><\/p>\n\n\n\n<p>We plan to integrate VLM-guided instructions into the noise-steering process. A key challenge is mapping natural-language guidance to the latent noise or action space.<\/p>\n\n\n\n<p><strong>2. IRL-Based Observation\u2192Reward Network<\/strong><\/p>\n\n\n\n<p>NAVSIM is not a real simulator, which limits online RL. We aim to use inverse reinforcement learning (IRL) to learn an observation-to-reward model, enabling more realistic RL training.<\/p>\n\n\n\n<p><strong>3. Diffusion Rollout for Data Augmentation<\/strong><\/p>\n\n\n\n<p>Human driving logs are biased toward high-reward behaviors. We plan to use diffusion models to rollout diverse trajectories and generate richer, more varied training data.<\/p>\n\n\n\n<p><strong>4. Exploring Alternative RL Methods (e.g., GRPO [6])<\/strong><\/p>\n\n\n\n<p>Diffusion policies generate entire trajectories at each timestep, making group-advantage methods like GRPO a promising direction for improving learning stability and efficiency.<\/p>\n\n\n\n<p><strong>5. Flow-Matching as an Alternative Action Head<\/strong><\/p>\n\n\n\n<p>Flow-matching action heads are increasingly popular in robotic manipulation. We plan to explore using a flow-matching-based action head for autonomous driving.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>1. VLM-Guided Noise Steering We plan to integrate VLM-guided instructions into the noise-steering process. A key challenge is mapping natural-language guidance to the latent noise or action space. 2. IRL-Based Observation\u2192Reward Network NAVSIM is not a real simulator, which limits online RL. We aim to use inverse reinforcement learning (IRL) to learn an observation-to-reward model, [&hellip;]<\/p>\n","protected":false},"author":245,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"footnotes":""},"class_list":["post-92","page","type-page","status-publish","hentry"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.4 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Future Work - Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Future Work - Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models\" \/>\n<meta property=\"og:description\" content=\"1. VLM-Guided Noise Steering We plan to integrate VLM-guided instructions into the noise-steering process. A key challenge is mapping natural-language guidance to the latent noise or action space. 2. IRL-Based Observation\u2192Reward Network NAVSIM is not a real simulator, which limits online RL. We aim to use inverse reinforcement learning (IRL) to learn an observation-to-reward model, [&hellip;]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/\" \/>\n<meta property=\"og:site_name\" content=\"Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models\" \/>\n<meta property=\"article:modified_time\" content=\"2025-12-12T01:27:15+00:00\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"1 minute\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/future-work\\\/\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/future-work\\\/\",\"name\":\"Future Work - Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/#website\"},\"datePublished\":\"2025-12-11T20:25:54+00:00\",\"dateModified\":\"2025-12-12T01:27:15+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/future-work\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/future-work\\\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/future-work\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Future Work\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/#website\",\"url\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/\",\"name\":\"Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/mscvprojects.ri.cmu.edu\\\/2025fteam24\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Future Work - Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/","og_locale":"en_US","og_type":"article","og_title":"Future Work - Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models","og_description":"1. VLM-Guided Noise Steering We plan to integrate VLM-guided instructions into the noise-steering process. A key challenge is mapping natural-language guidance to the latent noise or action space. 2. IRL-Based Observation\u2192Reward Network NAVSIM is not a real simulator, which limits online RL. We aim to use inverse reinforcement learning (IRL) to learn an observation-to-reward model, [&hellip;]","og_url":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/","og_site_name":"Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models","article_modified_time":"2025-12-12T01:27:15+00:00","twitter_card":"summary_large_image","twitter_misc":{"Est. reading time":"1 minute"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/","url":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/","name":"Future Work - Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models","isPartOf":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/#website"},"datePublished":"2025-12-11T20:25:54+00:00","dateModified":"2025-12-12T01:27:15+00:00","breadcrumb":{"@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/future-work\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/"},{"@type":"ListItem","position":2,"name":"Future Work"}]},{"@type":"WebSite","@id":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/#website","url":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/","name":"Reinforcement Learning for Noise Steering in Diffusion-Based Driving Models","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"}]}},"_links":{"self":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/pages\/92","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/users\/245"}],"replies":[{"embeddable":true,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/comments?post=92"}],"version-history":[{"count":4,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/pages\/92\/revisions"}],"predecessor-version":[{"id":133,"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/pages\/92\/revisions\/133"}],"wp:attachment":[{"href":"https:\/\/mscvprojects.ri.cmu.edu\/2025fteam24\/wp-json\/wp\/v2\/media?parent=92"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}