Domain Arithmetic: One-Shot VLA Adaptation under Environmental Shifts

• URL: http://arxiv.org/abs/2607.00666v1
• Date: Wed, 01 Jul 2026 09:13:40 GMT
• Title: Domain Arithmetic: One-Shot VLA Adaptation under Environmental Shifts
• Abstract summary: We propose an analogy-based method that adapts VLA models under environmental shifts through weight vector arithmetic.<n>Unlike prior approaches, DART requires collecting only a single demonstration, enabling efficient adaptation.<n>In both simulated and real-world experiments, DART outperforms existing VLA adaptation methods in one-shot scenarios.
• Score: 24.352848169805426
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Vision-Language-Action (VLA) models often fail to perform the same learned tasks under environmental shifts, such as changes in camera pose and shifts to a different but similar robot (e.g., from Panda to UR5e). Adapting these models to the shifted environment (i.e., target domain) often requires training on multiple demonstrations for each task, which are costly to collect. To reduce the burden of data curation and training, we propose an analogy-based method that adapts VLA models under environmental shifts through weight vector arithmetic with domain-specific information addition, named Domain ARiThmetic (DART). Unlike prior approaches, DART requires collecting only a single demonstration, enabling efficient adaptation. To accurately isolate domain-specific information for addition, DART performs subspace alignment between singular components in weight vectors to filter out noisy components. In both simulated and real-world experiments, DART outperforms existing VLA adaptation methods in one-shot scenarios across diverse visual and embodiment shifts. Code is available at https://github.com/snumprlab/dart.

Related papers

• Domain-Division based Progressive Learning for Source-Free Domain Adaptation [73.44692421206774]
  Source-free domain adaptation requires only a source pre-trained model and an unlabeled target domain.<n>We propose a domain-division based progressive learning method named DPL.<n>Our approach consists of two alternating stages, each beginning with the division of the target domain into easy-to-adapt and hard-to-adapt.
  arXiv  Detail & Related papers  (2026-07-31T09:21:56Z)
• EUDA: An Efficient Unsupervised Domain Adaptation via Self-Supervised Vision Transformer [21.59850502993888]
  Unsupervised domain adaptation (UDA) aims to mitigate the domain shift issue, where the distribution of training (source) data differs from that of testing (target) data. Many models have been developed to tackle this problem, and recently vision transformers (ViTs) have shown promising results. This paper introduces an efficient model that reduces trainable parameters and allows for adjustable complexity.
  arXiv  Detail & Related papers  (2024-07-31T03:29:28Z)
• Leveraging Normalization Layer in Adapters With Progressive Learning and Adaptive Distillation for Cross-Domain Few-Shot Learning [27.757318834190443]
  Cross-domain few-shot learning presents a formidable challenge, as models must be trained on base classes and tested on novel classes from various domains with only a few samples at hand. We introduce a novel generic framework that leverages normalization layer in adapters with Progressive Learning and Adaptive Distillation (ProLAD) We deploy two strategies: a progressive training of the two adapters and an adaptive distillation technique derived from features determined by the model solely with the adapter devoid of a normalization layer.
  arXiv  Detail & Related papers  (2023-12-18T15:02:14Z)
• SALUDA: Surface-based Automotive Lidar Unsupervised Domain Adaptation [62.889835139583965]
  We introduce an unsupervised auxiliary task of learning an implicit underlying surface representation simultaneously on source and target data. As both domains share the same latent representation, the model is forced to accommodate discrepancies between the two sources of data. Our experiments demonstrate that our method achieves a better performance than the current state of the art, both in real-to-real and synthetic-to-real scenarios.
  arXiv  Detail & Related papers  (2023-04-06T17:36:23Z)
• Progressive Transformation Learning for Leveraging Virtual Images in Training [21.590496842692744]
  We introduce Progressive Transformation Learning (PTL) to augment a training dataset by adding transformed virtual images with enhanced realism. PTL takes a novel approach that progressively iterates the following three steps: 1) select a subset from a pool of virtual images according to the domain gap, 2) transform the selected virtual images to enhance realism, and 3) add the transformed virtual images to the training set while removing them from the pool. Experiments show that PTL results in a substantial performance increase over the baseline, especially in the small data and the cross-domain regime.
  arXiv  Detail & Related papers  (2022-11-03T13:04:15Z)
• Exploiting Instance-based Mixed Sampling via Auxiliary Source Domain Supervision for Domain-adaptive Action Detection [75.38704117155909]
  We propose a novel domain adaptive action detection approach and a new adaptation protocol. Self-training combined with cross-domain mixed sampling has shown remarkable performance gain in UDA context. We name our proposed framework as domain-adaptive action instance mixing (DA-AIM)
  arXiv  Detail & Related papers  (2022-09-28T22:03:25Z)
• The Norm Must Go On: Dynamic Unsupervised Domain Adaptation by Normalization [10.274423413222763]
  Domain adaptation is crucial to adapt a learned model to new scenarios, such as domain shifts or changing data distributions. Current approaches usually require a large amount of labeled or unlabeled data from the shifted domain. We propose Dynamic Unsupervised Adaptation (DUA) to overcome this problem.
  arXiv  Detail & Related papers  (2021-12-01T12:43:41Z)
• Semi-Supervised Domain Adaptation with Prototypical Alignment and Consistency Learning [86.6929930921905]
  This paper studies how much it can help address domain shifts if we further have a few target samples labeled. To explore the full potential of landmarks, we incorporate a prototypical alignment (PA) module which calculates a target prototype for each class from the landmarks. Specifically, we severely perturb the labeled images, making PA non-trivial to achieve and thus promoting model generalizability.
  arXiv  Detail & Related papers  (2021-04-19T08:46:08Z)
• One to Many: Adaptive Instrument Segmentation via Meta Learning and Dynamic Online Adaptation in Robotic Surgical Video [71.43912903508765]
  MDAL is a dynamic online adaptive learning scheme for instrument segmentation in robot-assisted surgery. It learns the general knowledge of instruments and the fast adaptation ability through the video-specific meta-learning paradigm. It outperforms other state-of-the-art methods on two datasets.
  arXiv  Detail & Related papers  (2021-03-24T05:02:18Z)
• Adaptive Risk Minimization: Learning to Adapt to Domain Shift [109.87561509436016]
  A fundamental assumption of most machine learning algorithms is that the training and test data are drawn from the same underlying distribution. In this work, we consider the problem setting of domain generalization, where the training data are structured into domains and there may be multiple test time shifts. We introduce the framework of adaptive risk minimization (ARM), in which models are directly optimized for effective adaptation to shift by learning to adapt on the training domains.
  arXiv  Detail & Related papers  (2020-07-06T17:59:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.