Various adaptation methods, such as LoRA, prompts, and adapters, have been proposed to enhance the performance of pre-trained vision-language models in specific domains. As test samples in real-world applications usually differ from adaptation data, the robustness of these adaptation methods against distribution shifts are essential. In this study, we assess the robustness of 11 widely-used adaptation methods across 4 vision-language datasets under multimodal corruptions. Concretely, we introduce 7 benchmark datasets, including 96 visual and 87 textual corruptions, to investigate the robustness of different adaptation methods, the impact of available adaptation examples, and the influence of trainable parameter size during adaptation. Our analysis reveals that: 1) Adaptation methods are more sensitive to text corruptions than visual corruptions.

Recent advances in Vision-Language-Action (VLA) models have established a two-component architecture, where a pre-trained Vision-Language Model (VLM) encodes visual observations and task descriptions, and an action decoder maps these representations to continuous actions. Diffusion models have been widely adopted as action decoders due to their ability to model complex, multimodal action distributions. However, they require multiple iterative denoising steps at inference time or downstream techniques to speed up sampling, limiting their practicality in real-world settings where high-frequency control is crucial. In this work, we present NinA (Normalizing Flows in Action), a fast and expressive alternative to diffusion-based decoders for VLAs. NinA replaces the diffusion action decoder with a Normalizing Flow (NF) that enables one-shot sampling through an invertible transformation, significantly reducing inference time. We integrate NinA into the FLOWER VLA architecture and fine-tune on the LIBERO benchmark. Our experiments show that NinA matches the performance of its diffusion-based counterpart under the same training regime, while achieving substantially faster inference. These results suggest that NinA offers a promising path toward efficient, high-frequency VLA control without compromising performance.

A straightforward pipeline for zero-shot out-of-distribution (OOD) detection involves selecting potential OOD labels from an extensive semantic pool and then leveraging a pre-trained vision-language model to perform classification on both in-distribution (ID) and OOD labels. In this paper, we theorize that enhancing performance requires expanding the semantic pool, while increasing the expected probability of selected OOD labels being activated by OOD samples, and ensuring low mutual dependence among the activations of these OOD labels. A natural expansion manner is to adopt a larger lexicon; however, the inevitable introduction of numerous synonyms and uncommon words fails to meet the above requirements, indicating that viable expansion manners move beyond merely selecting words from a lexicon. Since OOD detection aims to correctly classify input images into ID/OOD class groups, we can "make up" OOD label candidates which are not standard class names but beneficial for the process. Observing that the original semantic pool is comprised of unmodified specific class names, we correspondingly construct a conjugated semantic pool (CSP) consisting of modified superclass names, each serving as a cluster center for samples sharing similar properties across different categories.

Various adaptation methods, such as LoRA, prompts, and adapters, have been proposed to enhance the performance of pre-trained vision-language models in specific domains. As test samples in real-world applications usually differ from adaptation data, the robustness of these adaptation methods against distribution shifts are essential. In this study, we assess the robustness of 11 widely-used adaptation methods across 4 vision-language datasets under multimodal corruptions. Concretely, we introduce 7 benchmark datasets, including 96 visual and 87 textual corruptions, to investigate the robustness of different adaptation methods, the impact of available adaptation examples, and the influence of trainable parameter size during adaptation. Our analysis reveals that: 1) Adaptation methods are more sensitive to text corruptions than visual corruptions.

State recognition of the environment and objects, such as the open/closed state of doors and the on/off of lights, is indispensable for robots that perform daily life support and security tasks. Until now, state recognition methods have been based on training neural networks from manual annotations, preparing special sensors for the recognition, or manually programming to extract features from point clouds or raw images. In contrast, we propose a robotic state recognition method using a pre-trained vision-language model, which is capable of Image-to-Text Retrieval (ITR) tasks. We prepare several kinds of language prompts in advance, calculate the similarity between these prompts and the current image by ITR, and perform state recognition. By applying the optimal weighting to each prompt using black-box optimization, state recognition can be performed with higher accuracy. Experiments show that this theory enables a variety of state recognitions by simply preparing multiple prompts without retraining neural networks or manual programming. In addition, since only prompts and their weights need to be prepared for each recognizer, there is no need to prepare multiple models, which facilitates resource management. It is possible to recognize the open/closed state of transparent doors, the state of whether water is running or not from a faucet, and even the qualitative state of whether a kitchen is clean or not, which have been challenging so far, through language.

For example, the robot must recognize whether a door is open, a light is on, water is running, a fire is burning, and so on. In order to change the robot's behavior based on the recognition results, state recognition is usually performed with discrete values of about two or three options. Until now, appropriate individual methods have been used for each state to be recognized, such as direct processing of images or point clouds by human programming [3, 4], creating a dataset with annotations and training neural networks [5], or detecting the state by installing new sensors [6, 7]. However, these methods require us to manually program the process for each state recognition, to train neural networks one by one, and to increase the number of sensors installed. In addition, this will increase the number of programs and trained models needed for each state recognition, which will cause problems in management of source code and computer resource. To cope with these problems, a single program or model should be able to recognize multiple states. In this study, we propose a method to easily recognize various environmental states in a unified manner and through the spoken language (Figure 1). In order to perform state recognition through the spoken language, we use pre-trained large-scale vision-language models (VLMs) [8-12]. Currently, VLMs are being used for map generation [13, 14], scene understanding [15-17], and feature extraction for behav-Corresponding author.

Various robot navigation methods have been developed, but they are mainly based on Simultaneous Localization and Mapping (SLAM), reinforcement learning, etc., which require prior map construction or learning. In this study, we consider the simplest method that does not require any map construction or learning, and execute open-vocabulary navigation of robots without any prior knowledge to do this. We applied an omnidirectional camera and pre-trained vision-language models to the robot. The omnidirectional camera provides a uniform view of the surroundings, thus eliminating the need for complicated exploratory behaviors including trajectory generation. By applying multiple pre-trained vision-language models to this omnidirectional image and incorporating reflective behaviors, we show that navigation becomes simple and does not require any prior setup. Interesting properties and limitations of our method are discussed based on experiments with the mobile robot Fetch.

Recognition of the current state is indispensable for the operation of a robot. There are various states to be recognized, such as whether an elevator door is open or closed, whether an object has been grasped correctly, and whether the TV is turned on or off. Until now, these states have been recognized by programmatically describing the state of a point cloud or raw image, by annotating and learning images, by using special sensors, etc. In contrast to these methods, we apply Visual Question Answering (VQA) from a Pre-Trained Vision-Language Model (PTVLM) trained on a large-scale dataset, to such binary state recognition. This idea allows us to intuitively describe state recognition in language without any re-training, thereby improving the recognition ability of robots in a simple and general way. We summarize various techniques in questioning methods and image processing, and clarify their properties through experiments.

For robots to follow instructions from people, they must be able to connect the rich semantic information in human vocabulary, e.g. "can you get me the pink stuffed whale?" to their sensory observations and actions. This brings up a notably difficult challenge for robots: while robot learning approaches allow robots to learn many different behaviors from first-hand experience, it is impractical for robots to have first-hand experiences that span all of this semantic information. We would like a robot's policy to be able to perceive and pick up the pink stuffed whale, even if it has never seen any data interacting with a stuffed whale before. Fortunately, static data on the internet has vast semantic information, and this information is captured in pre-trained vision-language models. In this paper, we study whether we can interface robot policies with these pre-trained models, with the aim of allowing robots to complete instructions involving object categories that the robot has never seen first-hand. We develop a simple approach, which we call Manipulation of Open-World Objects (MOO), which leverages a pre-trained vision-language model to extract object-identifying information from the language command and image, and conditions the robot policy on the current image, the instruction, and the extracted object information. In a variety of experiments on a real mobile manipulator, we find that MOO generalizes zero-shot to a wide range of novel object categories and environments. In addition, we show how MOO generalizes to other, non-language-based input modalities to specify the object of interest such as finger pointing, and how it can be further extended to enable open-world navigation and manipulation. The project's website and evaluation videos can be found at https://robot-moo.github.io/

In recent years, a number of models that learn the relations between vision and language from large datasets have been released. These models perform a variety of tasks, such as answering questions about images, retrieving sentences that best correspond to images, and finding regions in images that correspond to phrases. Although there are some examples, the connection between these pre-trained vision-language models and robotics is still weak. If they are directly connected to robot motions, they lose their versatility due to the embodiment of the robot and the difficulty of data collection, and become inapplicable to a wide range of bodies and situations. Therefore, in this study, we categorize and summarize the methods to utilize the pre-trained vision-language models flexibly and easily in a way that the robot can understand, without directly connecting them to robot motions. We discuss how to use these models for robot motion selection and motion planning without re-training the models. We consider five types of methods to extract information understandable for robots, and show the results of state recognition, object recognition, affordance recognition, relation recognition, and anomaly detection based on the combination of these five methods. We expect that this study will add flexibility and ease-of-use, as well as new applications, to the recognition behavior of existing robots.