EVLP:Learning Unified Embodied Vision-Language Planner with Reinforced Supervised Fine-Tuning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Cai, Xinyan, Wu, Shiguang, Chi, Dafeng, Zhuang, Yuzheng, Quan, Xingyue, Hao, Jianye, Guan, Qiang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911255369875456
author Cai, Xinyan
Wu, Shiguang
Chi, Dafeng
Zhuang, Yuzheng
Quan, Xingyue
Hao, Jianye
Guan, Qiang
author_facet Cai, Xinyan
Wu, Shiguang
Chi, Dafeng
Zhuang, Yuzheng
Quan, Xingyue
Hao, Jianye
Guan, Qiang
contents In complex embodied long-horizon manipulation tasks, effective task decomposition and execution require synergistic integration of textual logical reasoning and visual-spatial imagination to ensure efficient and accurate operation. Current methods fail to adopt a unified generation framework for multimodal planning, lead to inconsistent in multimodal planning. To address this challenge, we present \textbf{EVLP (Embodied Vision-Language Planner)}, an innovative multimodal unified generation framework that jointly models linguistic reasoning and visual generation. Our approach achieves multimodal planning for long-horizon tasks through a novel training pipeline incorporating dynamic pretraining and reinforced alignment. Our core innovations consist of three key components: \textbf{1) Unified Multimodal Generation Framework}: For understanding, We integrate semantic information with spatial features to provide comprehensive visual perception. For generation, we directly learn the joint distribution of discrete images for one-step visual synthesis, enabling coordinated language-visual modeling through learnable cross-modal attention mechanisms. \textbf{2) Dynamic Perception Pretraining}: We propose a bidirectional dynamic alignment strategy employing inverse dynamics tasks and forward dynamics tasks, effectively strengthening multimodal correlations within a unified feature space. \textbf{3) Reinforced Supervised Fine-Tuning}: While conducting instruction-based fine-tuning in the unified generation space, we construct a reinforce loss to align the spatial logic between textual actions and generated images, enabling the model to acquire spatio-awared multimodal planning capabilities.
format Preprint
id arxiv_https___arxiv_org_abs_2511_05553
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle EVLP:Learning Unified Embodied Vision-Language Planner with Reinforced Supervised Fine-Tuning
Cai, Xinyan
Wu, Shiguang
Chi, Dafeng
Zhuang, Yuzheng
Quan, Xingyue
Hao, Jianye
Guan, Qiang
Computer Vision and Pattern Recognition
Artificial Intelligence
In complex embodied long-horizon manipulation tasks, effective task decomposition and execution require synergistic integration of textual logical reasoning and visual-spatial imagination to ensure efficient and accurate operation. Current methods fail to adopt a unified generation framework for multimodal planning, lead to inconsistent in multimodal planning. To address this challenge, we present \textbf{EVLP (Embodied Vision-Language Planner)}, an innovative multimodal unified generation framework that jointly models linguistic reasoning and visual generation. Our approach achieves multimodal planning for long-horizon tasks through a novel training pipeline incorporating dynamic pretraining and reinforced alignment. Our core innovations consist of three key components: \textbf{1) Unified Multimodal Generation Framework}: For understanding, We integrate semantic information with spatial features to provide comprehensive visual perception. For generation, we directly learn the joint distribution of discrete images for one-step visual synthesis, enabling coordinated language-visual modeling through learnable cross-modal attention mechanisms. \textbf{2) Dynamic Perception Pretraining}: We propose a bidirectional dynamic alignment strategy employing inverse dynamics tasks and forward dynamics tasks, effectively strengthening multimodal correlations within a unified feature space. \textbf{3) Reinforced Supervised Fine-Tuning}: While conducting instruction-based fine-tuning in the unified generation space, we construct a reinforce loss to align the spatial logic between textual actions and generated images, enabling the model to acquire spatio-awared multimodal planning capabilities.
title EVLP:Learning Unified Embodied Vision-Language Planner with Reinforced Supervised Fine-Tuning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2511.05553