AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

MetaVLA: Unified Meta Co-Training for Efficient Embodied Adaptation

Chen Li, Zhantao Yang, Han Zhang, Fangyi Chen, Chenchen Zhu, Anudeepsekhar Bolimera, Marios Savvides

ICLR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Embodied multimodal LLM agents

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add auxiliary GR00T tasks to naive multi-task SFT (ablation)SFT-4LIBERO (OpenVLA) · LIBERO averagesuccess rate (%): 76.2 -> 8.6Hurt
train naive SFT with auxiliary tasks for more steps (187.5K vs 75K) (ablation)SFT-4LIBERO+5single+1bimanual · LIBERO averagesuccess rate (%): 8.6 -> 14.5Mixed
add Meta-Action-Reasoner context module with LIBERO context only (ablation)SFT-4LIBERO (OpenVLA) · LIBERO averagesuccess rate (%): 76.2 -> 77.8Improved
add six auxiliary tasks to the MetaVLA context bank (ablation)MetaVLA · LIBERO averagesuccess rate (%): 77.8 -> 79.3Improved
replace exotic auxiliary tasks with tasks already in OpenVLA pretraining (ablation)MetaVLA · LIBERO averagesuccess rate (%): 77.8 -> 74.4Hurt
add stochastic latent with KL regularization (ablation)MetaVLA · LIBERO-Longsuccess rate (%): 55.3 -> 53.0Hurt
increase context batch size from 4 to 32 (ablation)MetaVLA · LIBERO averagesuccess rate (%): 73.2 -> 77.8Improved
apply MetaVLA to NORA-Long without auxiliary tasks (ablation)NORA-Long · LIBERO averagesuccess rate (%): 87.3 -> 90.3Improved
add auxiliary tasks to naive SFT on NORA-Long (ablation)NORA-Long-SFT-4LIBERO · LIBERO averagesuccess rate (%): 87.3 -> 66.4Hurt
add auxiliary tasks to MetaVLA on NORA-Long (ablation)MetaVLA-NORA-Long · LIBERO averagesuccess rate (%): 90.3 -> 91.8Improved
train MetaVLA separately per suite (MetaVLA-EACH) instead of one co-trained model (ablation)MetaVLA · LIBERO-Long, 240K stepssuccess rate (%): 51.8 -> 55.8Improved
unify four suite models into one MetaVLA modelMetaVLA · LIBERO training costtraining steps: 240K -> 75KImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy