# SVLL: Staged Vision-Language Learning for Physically Grounded Embodied Task Planning

arXiv: https://arxiv.org/abs/2603.11563 · HTML: https://mlsanity.com/paper/2603.11563/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Embodied and 3D vision-language models, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 26 of its 37 references that are in the map.
- First posted on arXiv 12 Mar 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Embodied and 3D vision-language models](https://mlsanity.com/mllm/embodied-and-3d-vision-language-models/)
- LLM alignment / RLHF / preference optimisation / [Reward modeling and RLHF](https://mlsanity.com/alignment/reward-modeling-and-rlhf/)

## Builds on (most cited first, 26 of 26)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [Do As I Can, Not As I Say: Grounding Language in Robotic Affordances](https://mlsanity.com/paper/2204.01691/) (2022)
- [Code as Policies: Language Model Programs for Embodied Control](https://mlsanity.com/paper/2209.07753/) (2022)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [Large Language Models](https://mlsanity.com/paper/2307.05782/) (2023)
- [Multimodal Chain-of-Thought Reasoning in Language Models](https://mlsanity.com/paper/2302.00923/) (2023)
- [Eureka: Human-Level Reward Design via Coding Large Language Models](https://mlsanity.com/paper/2310.12931/) (2023)
- [CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Models](https://mlsanity.com/paper/2503.22020/) (2025)
- [Robotic Control via Embodied Chain-of-Thought Reasoning](https://mlsanity.com/paper/2407.08693/) (2024)
- [A Survey on Vision-Language-Action Models for Embodied AI](https://mlsanity.com/paper/2405.14093/) (2024)
- [Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization](https://mlsanity.com/paper/2411.10442/) (2024)
- [Text2Motion: From Natural Language Instructions to Feasible Plans](https://mlsanity.com/paper/2303.12153/) (2023)
- [Hi Robot: Open-Ended Instruction Following with Hierarchical Vision-Language-Action Models](https://mlsanity.com/paper/2502.19417/) (2025)
- [Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents](https://mlsanity.com/paper/2403.02502/) (2024)
- [Embodied Task Planning with Large Language Models](https://mlsanity.com/paper/2307.01848/) (2023)
- [RedStar: Does Scaling Long-CoT Data Unlock Better Slow-Reasoning Systems?](https://mlsanity.com/paper/2501.11284/) (2025)
- [Context-Aware Planning and Environment-Aware Memory for Instruction Following Embodied Agents](https://mlsanity.com/paper/2308.07241/) (2023)
- [World Modeling Makes a Better Planner: Dual Preference Optimization for Embodied Task Planning](https://mlsanity.com/paper/2503.10480/) (2025)
- [Aux-Think: Exploring Reasoning Strategies for Data-Efficient Vision-Language Navigation](https://mlsanity.com/paper/2505.11886/) (2025)
- [Socratic Planner: Self-QA-Based Zero-Shot Planning for Embodied Instruction Following](https://mlsanity.com/paper/2404.15190/) (2024)
- [Reward Evolution with Graph-of-Thoughts: A Bi-Level Language Model Framework for Reinforcement Learning](https://mlsanity.com/paper/2509.16136/) (2025)
- [ReFineVLA: Reasoning-Aware Teacher-Guided Transfer Fine-Tuning](https://mlsanity.com/paper/2505.19080/) (2025)
- [Efficacy of Image Similarity as a Metric for Augmenting Small Dataset Retinal Image Segmentation](https://mlsanity.com/paper/2507.04862/) (2025)

## Built on by (most cited first, 1 of 1)

- [Getting Space Right: Agentic Rewards for Image Generation](https://mlsanity.com/paper/2610.14852/) (2026)

## Abstract (arXiv)

Embodied task planning demands vision-language models to generate action sequences that are both visually grounded and causally coherent over time. However, existing training paradigms face a critical trade-off: joint end-to-end training often leads to premature temporal binding, while standard reinforcement learning methods suffer from optimization instability. To bridge this gap, we present Staged Vision-Language Learning (SVLL), a unified three-stage framework for robust, physically-grounded embodied planning. In the first two stages, SVLL decouples spatial grounding from temporal reasoning, establishing robust visual dependency before introducing sequential action history. In the final stage, we identify a key limitation of standard Direct Preference Optimization (DPO), its purely relative nature -- optimizing only the preference gap between winning and losing trajectories while neglecting absolute likelihood constraints on optimal path, often yields unsafe or hallucinated behaviors. To address this, we further introduce Bias-DPO, a novel alignment objective that injects an inductive bias toward expert trajectories by explicitly maximizing likelihood on ground-truth actions while penalizing overconfident hallucinations. By anchoring the policy to the expert manifold and mitigating causal misalignment, SVLL, powered by Bias-DPO, ensures strict adherence to environmental affordances and effectively suppresses physically impossible shortcuts. Finally, extensive experiments on the interactive AI2-THOR benchmark and real-world robotic deployments demonstrate that SVLL outperforms both state-of-the-art open-source (e.g., Qwen2.5-VL-7B) and closed-source models (e.g., GPT-4o, Gemini-2.0-flash) in task success rate, while significantly reducing physical constraint violations.