# Offline Semantic Guidance for Efficient Vision-Language-Action Policy Distillation

arXiv: https://arxiv.org/abs/2605.16241 · HTML: https://mlsanity.com/paper/2605.16241/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Generalist VLA models and benchmarks, one of 13 directions in Vision-language-action models.
- Builds on 13 of its 20 references that are in the map.
- First posted on arXiv 15 May 2026.

## Placed in

- Vision-language-action models / [Generalist VLA models and benchmarks](https://mlsanity.com/vla/generalist-vla-models-and-benchmarks/)
- Knowledge distillation / [Multimodal distillation](https://mlsanity.com/distill/multimodal-distillation/)

## Builds on (most cited first, 13 of 13)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models](https://mlsanity.com/paper/2307.05973/) (2023)
- [What Matters in Learning from Offline Human Demonstrations for Robot Manipulation](https://mlsanity.com/paper/2108.03298/) (2021)
- [Learning by Cheating](https://mlsanity.com/paper/1912.12294/) (2019)
- [Long-CLIP: Unlocking the Long-Text Capability of CLIP](https://mlsanity.com/paper/2403.15378/) (2024)
- [RoboAgent: Generalization and Efficiency in Robot Manipulation via Semantic Augmentations and Action Chunking](https://mlsanity.com/paper/2309.01918/) (2023)
- [CEED-VLA: Consistency Vision-Language-Action Model with Early-Exit Decoding](https://mlsanity.com/paper/2506.13725/) (2025)
- [Don't Blind Your VLA: Aligning Visual Representations for OOD Generalization](https://mlsanity.com/paper/2510.25616/) (2025)
- [Mixture of Experts in Large Language Models](https://mlsanity.com/paper/2507.11181/) (2025)
- [ActDistill: General Action-Guided Self-Derived Distillation for Efficient Vision-Language-Action Models](https://mlsanity.com/paper/2511.18082/) (2025)

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

- [Anatomy of a Closed-Loop Collapse: A Causal Case Study of a Compressed VLA Policy](https://mlsanity.com/paper/2609.23048/) (2026)

## Abstract (arXiv)

Billion-parameter Vision-Language-Action (VLA) policies have recently shown impressive performance in robotic manipulation, yet their size and inference cost remain major obstacles for real-time closed-loop control. We introduce VLA-AD, a distillation framework that uses a Vision-Language Model as an offline semantic supervisor to transfer large VLA teachers into lightweight student policies. Instead of relying only on low-level action imitation, VLA-AD augments teacher-provided 7-DoF action targets with high-level semantic guidance, including task phase anchors and multi-frame operating-direction descriptions. These auxiliary signals are used only during training: at test time, the student policy runs independently, with neither the VLA teacher nor the VLM required. We evaluate VLA-AD on three LIBERO benchmark suites. Using OpenVLA-7B as the teacher, our method produces a 158M-parameter student, yielding a 44× reduction in model size while matching the teacher with only a 0.27% average relative gap. The resulting policy runs at 12.5 Hz on an RTX 4090, achieving a 3.28× inference speedup over OpenVLA-7B. We further show that the same semantic distillation pipeline generalizes to a different π0.5-4B teacher, where the student outperforms the teacher on two suites and remains within 0.53% on textttliberogoal. Additional analysis indicates that phase-level supervision and multi-frame directional cues make the student less sensitive to noisy teacher actions, such as erroneous high-frequency gripper changes. Overall, VLA-AD demonstrates that offline semantic guidance from VLMs can substantially improve the efficiency, robustness, and deployability of VLA policy distillation.