Foundation models (FMs) trained with different objectives and data learn diverse representations, making some more effective than others for specific downstream tasks. Existing adaptation strategies, such as parameter-efficient fine-tuning, focus on individual models and do not exploit the complementary strengths across models. Probing methods offer a promising alternative by extracting information from frozen models, but current techniques do not scale well with large feature sets and often rely on dataset-specific hyperparameter tuning. We propose Combined backBones (ComBo), a simple and scalable probing-based adapter that effectively integrates features from multiple models and layers. ComBo compresses activations from layers of one or more FMs into compact token-wise representations and processes them with a lightweight transformer for task-specific prediction. Crucially, ComBo does not require dataset-specific tuning or backpropagation through the backbone models. However, not all models are equally relevant for all tasks. To address this, we introduce a mechanism that leverages ComBo’s joint multi-backbone probing to efficiently evaluate each backbone’s task-relevance, enabling both practical model comparison and improved performance through selective adaptation. On the 19 tasks of the VTAB-1k benchmark, ComBo outperforms previous probing methods, matches or surpasses more expensive alternatives, such as distillation-based model merging, and enables efficient probing of tuned models. Our results demonstrate that ComBo offers a practical and general-purpose framework for combining diverse representations from multiple FMs.

This paper adapts a general dataset representation technique to produce robust Visual Place Recognition (VPR) descriptors, crucial to enable real-world mobile robot localisation. Two parallel lines of work on VPR have shown, on one side, that general-purpose off-the-shelf feature representations can provide robustness to domain shifts, and, on the other, that fused information from sequences of images improves performance. In our recent work on measuring domain gaps between image datasets, we proposed a Visual Distribution of Neuron Activations (VDNA) representation to represent datasets of images. This representation can naturally handle image sequences and provides a general and granular feature representation derived from a general-purpose model. Moreover, our representation is based on tracking neuron activation values over the list of images to represent and is not limited to a particular neural network layer, therefore having access to high- and low-level concepts. This work shows how VDNAs can be used for VPR by learning a very lightweight and simple encoder to generate task-specific descriptors. Our experiments show that our representation can allow for better robustness than current solutions to serious domain shifts away from the training data distribution, such as to indoor environments and aerial imagery.

Selecting appropriate datasets is critical in modern computer vision. However, no general-purpose tools exist to evaluate the extent to which two datasets differ. For this, we propose representing images - and by extension datasets - using Distributions of Neuron Activations (DNAs). DNAsfit distributions, such as histograms or Gaussians, to activations of neurons in a pre-trained feature extractor through which we pass the imager s) to represent. This extractor is frozen for all datasets, and we rely on its generally expressive power in feature space. By comparing two DNAs, we can evaluate the extent to which two datasets differ with granular control over the comparison attributes of interest, providing the ability to customise the way distances are measured to suit the requirements of the task at hand. Furthermore, DNAs are compact, representing datasets of any size with less than 15 megabytes. We demonstrate the value of DNAs by evaluating their applicability on several tasks, including conditional dataset comparison, synthetic image evaluation, and transfer learning, and across diverse datasets, ranging from synthetic cat images to celebrity faces and urban driving scenes.

Using multiple robots for exploring and mapping environments can provide improved robustness and performance, but it can be difficult to implement. In particular, limited communication bandwidth is a considerable constraint when a robot needs to determine if it has visited a location that was previously explored by another robot, as it requires for robots to share descriptions of places they have visited. One way to compress this description is to use constellations, groups of 3D points that correspond to the estimate of a set of relative object positions. Constellations maintain the same pattern from different viewpoints and can be robust to illumination changes or dynamic elements. We present a method to extract from these constellations compact spatial and semantic descriptors of the objects in a scene. We use this representation in a 2step decentralized loop closure verification: first, we distribute the compact semantic descriptors to determine which other robots might have seen scenes with similar objects; then we query matching robots with the full constellation to validate the match using geometric information. The proposed method requires less memory, is more interpretable than global image descriptors, and could be useful for other tasks and interactions with the environment. We validate our system's performance on a TUM RGB-D SLAM sequence and show its benefits in terms of bandwidth requirements.