Composed image retrieval (CIR) aims to retrieve a target image that depicts a reference image modified by a textual description. While recent vision-language models (VLMs) achieve promising CIR performance by embedding images and text into a shared space for retrieval, they often fail to reason about what to preserve and what to change. This limitation hinders interpretability and yields suboptimal results, particularly in fine-grained domains like fashion. In this paper, we introduce FIRE-CIR, a model that brings compositional reasoning and interpretability to fashion CIR. Instead of relying solely on embedding similarity, FIRE-CIR performs question-driven visual reasoning: it automatically generates attribute-focused visual questions derived from the modification text, and verifies the corresponding visual evidence in both reference and candidate images. To train such a reasoning system, we automatically construct a large-scale fashion-specific visual question answering dataset, containing questions requiring either single- or dual-image analysis. During retrieval, our model leverages this explicit reasoning to re-rank candidate results, filtering out images inconsistent with the intended modifications. Experimental results on the Fashion IQ benchmark show that FIRE-CIR outperforms state-of-the-art methods in retrieval accuracy. It also provides interpretable, attribute-level insights into retrieval decisions.

Endowing the configuration space of a robot with an appropriate metric structure and characterizing and computing the corresponding geodesics are central issues in motion planning. As recently observed in de Mont-Marin et al. (A minimum swept-volume metric structure for configuration space. In: IEEE international conference on robotics and automation (ICRA), pp 3686–3692, 2023), the geodesics of SE(2) equipped with the so called minimum swept-volume distance exhibit in practice a behavior akin to the turnpike property in optimal control, with transient phases separated by a longer steady state close to prototypical trajectories, the turnpikes (Trélat and Zuazua (2015) J Differ Equ 258(1):81–114). This presentation gives a theoretical counterpoint to this empirical observation with a formal definition of geodesic turnpikes using vector fields on Finsler manifolds, a simple differential characterization of geodesics in the case where the manifold is a Lie group and the Finsler distance is left-invariant, and, in the case where the corresponding operator is also reversible, a conjecture characterizing the turnpikes by vector fields satisfying simple conditions in the corresponding Lie algebras. As a proof of concept, closed-form (resp. numerical) procedures for computing the vector fields predicted by this conjecture are given for SE(2) equipped with the left-invariant Riemannian (resp. minimum swept-volume) distance introduced in Žefran et al. (IEEE Trans Robot Autom 14(4):576–589, 1998) (resp. de Mont-Marin et al. (A minimum swept-volume metric structure for configuration space. In: IEEE international conference on robotics and automation (ICRA), pp 3686–3692, 2023)) for rectangular shapes. The solutions empirically match, in both cases, the observed turnpike behavior of the corresponding geodesics. In the minimum swept-volume distance case, using the turnpikes for initialization also yields an order of magnitude speedup in computing geodesics.

Building deep learning models that can reason about their environment requires capturing its underlying dynamics. Joint-Embedded Predictive Architectures (JEPA) provide a promising framework to model such dynamics by learning representations and predictors through a self-supervised prediction objective. However, their ability to support effective action planning remains limited. We propose an approach to enhance planning with JEPA world models by shaping their representation space so that the negative goal-conditioned value function for a reaching cost in a given environment is approximated by a distance (or quasi-distance) between state embeddings. We introduce a practical method to enforce this constraint during training and show that it leads to significantly improved planning performance compared to standard JEPA models on simple control tasks.

The {\em stop gradient} and {\em exponential moving average} iterative procedures are commonly used in non-contrastive approaches to self-supervised learning to avoid representation collapse, with excellent performance in downstream applications in practice. This presentation investigates these procedures from the dual viewpoints of optimization and dynamical systems. We show that, in general, although they {\em do not} optimize the original objective, or {\em any} other smooth function, they {\em do} avoid collapse. Following [Tian et al. 2021], but without any of the extra assumptions used in their proofs, we then show using a dynamical system perspective that, in the linear case, minimizing the original objective function without the use of a stop gradient or exponential moving average {\em always} leads to collapse. Conversely, we characterize explicitly the equilibria of the dynamical systems associated with these two procedures in this linear setting as algebraic varieties in their parameter space, and show that they are, in general, {\em asymptotically stable}. Our theoretical findings are illustrated by empirical experiments with real and synthetic data.

Non-contrastive self-supervised learning (SSL) is an effective framework for predictive representation learning, but popular (and in practice effective) methods such as SimSiam, BYOL, I-JEPA or DINO, which rely on a form of self-distillation to train a teacher-student network, remain poorly understood as they typically do not minimize a well-defined objective. We analyze the dynamics of a variant of the Joint Embedding Predictive Architecture (JEPA) using a regularized linear regressor to predict the learned representations of two views of the data from one another, and fully characterize its stability: non-collapsed stable equilibria align with leading nonlinear canonical correlation subspaces, while collapsed equilibria may also be stable attractors. Motivated by this result, we introduce PEIRA, a non-contrastive SSL method with an explicit objective defined through the trace of the optimal linear regressor. We show that its only stable equilibria are nontrivial global minimizers and recover the same canonical correlation subspaces, with regularization selecting the effective dimension. Experiments on ImageNet-1K and CIFAR-10 show PEIRA is competitive with VICReg and LeJEPA baselines, and qualitative empirical results support the theory.

We present a unifying theoretical result that con- nects two foundational principles in robotics: the Signorini law for point contacts, which underpins many simulation methods for preventing object interpenetration, and the center of pres- sure (also known as the zero-moment point), a key concept in optimization-based locomotion control. Our contribution is the planar Signorini condition, a conic complementarity formulation that models general planar contacts between rigid bodies. We prove that this formulation is equivalent to enforcing the punctual Signorini law across an entire contact surface, thereby bridging the gap between discrete and continuous contact models. A geometric interpretation reveals that the framework naturally captures three physical regimes —stick- ing, separating, and tilting— within a unified complementarity structure. This leads to a principled extension of the classical center of pressure, which we refer to as the extended center of pressure. By establishing this connection, our work provides a mathematically consistent and computationally tractable foundation for handling planar contacts, with implications for both the accurate simulation of contact dynamics and the design of next-generation control and optimization algorithms in locomotion and manipulation.

This paper addresses the problem of reconstructing a scene online at the level of objects given an RGB-D video sequence. While current object-aware neural implicit representations hold promise, they are limited in online reconstruction efficiency and shape completion. Our main contributions to alleviate the above limitations are twofold. First, we propose a feature grid interpolation mechanism to continuously update grid-based object-centric neural implicit representations as new object parts are revealed. Second, we construct an object library with previously mapped objects in advance and leverage the corresponding shape priors to initialize geometric object models in new videos, subsequently completing them with novel views as well as synthesized past views to avoid losing original object details. Extensive experiments on synthetic environments from the Replica dataset, real-world ScanNet sequences and videos captured in our laboratory demonstrate that our approach outperforms state-of-the-art neural implicit models for this task in terms of reconstruction accuracy and completeness.

Improvements in Earth observation by satellites allow for imagery of ever higher temporal and spatial resolution. Leveraging this data for agricultural monitoring is key for addressing environmental and economic challenges. In this paper, we present and compare datasets and methods for both supervised and unsupervised pixel-wise segmentation of agricultural satellite image time series (SITS), and introduce a prototype-based method. Our approach learns deformable prototypes that can be transformed through time warping and spectral adjustments to reconstruct input sequences. The method can be trained with or without supervision and provides interpretable results through visualization of prototypes and their transformations. We demonstrate competitive performance on multiple agricultural SITS datasets while offering insights into crop temporal patterns. Our complete code is available at https://github.com/ElliotVincent/AgriITSC.

A long-standing challenge in AI is to develop agents capable of solving a wide range of physical tasks and generalizing to new, unseen tasks and environments. A popular recent approach involves training a world model from state-action trajectories and subsequently use it with a planning algorithm to solve new tasks. Planning is commonly performed in the input space, but a recent family of methods has introduced planning algorithms that optimize in the learned representation space of the world model, with the promise that abstracting irrelevant details yields more efficient planning. In this work, we characterize models from this family as JEPA-WMs and investigate the technical choices that make algorithms from this class work. We propose a comprehensive study of several key components with the objective of finding the optimal approach within the family. We conducted experiments using both simulated environments and real-world robotic data, and studied how the model architecture, the training objective, and the planning algorithm affect planning success. We combine our findings to propose a model that outperforms two established baselines, DINO-WM and V-JEPA-2-AC, in both navigation and manipulation tasks. Code, data and checkpoints are available at https://github.com/facebookresearch/jepa-wms.

Automated and robust portrait quality assessment (PQA) is of paramount importance in high-impact applications such as smartphone photography. This paper presents FHIQA, a learning-based approach to PQA that introduces a simple but effective quality score rescaling method based on image semantics, to enhance the precision of fine-grained image quality metrics while ensuring robust generalization to various scene settings beyond the training dataset. The proposed approach is validated by extensive experiments on the PIQ23 benchmark and comparisons with the current state of the art. The source code of FHIQA will be made publicly available on the PIQ23 GitHub repository at https://github.com/DXOMARK-Research/PIQ2023.

The search for exoplanets is an active field in astronomy, with direct imaging as one of the most challenging methods due to faint exoplanet signals buried within stronger residual starlight. Successful detection requires advanced image processing to separate the exoplanet signal from this nuisance component. This paper presents a novel statistical model that captures nuisance fluctuations using a multi-scale approach, leveraging problem symmetries and a joint spectral channel representation grounded in physical principles. Our model integrates into an interpretable, end-to end learnable framework for simultaneous exoplanet detection and flux estimation. The proposed algorithm is evaluated against the state of the art using datasets from the SPHERE instrument operating at the Very Large Telescope (VLT). It significantly improves the precision-recall tradeoff notably on challenging datasets that are otherwise unusable by astronomers. The proposed approach is computationally efficient, robust to varying data quality, and well suited for large-scale observational surveys. (1)

Single-molecule localization microscopy (SMLM) allows reconstructing biology-relevant structures beyond the diffraction limit by detecting and localizing individual fluorophores – fluorescent molecules stained onto the observed specimen – over time to reconstruct super-resolved images. Currently, efficient SMLM requires non-overlapping emitting fluorophores, leading to long acquisition times that hinders live-cell imaging. Recent deep-learning approaches can handle denser emissions, but they rely on variants of non-maximum suppression (NMS) layers, which are unfortunately non-differentiable and may discard true positives with their local fusion strategy. In this presentation, we reformulate the SMLM training objective as a set-matching problem, deriving an optimal-transport loss that eliminates the need for NMS during inference and enables end-to-end training. Additionally, we propose an iterative neural network that integrates knowledge of the microscope's optical system inside our model. Experiments on synthetic benchmarks and real biological data show that both our new loss function and architecture surpass the state of the art at moderate and high emitter densities. Code is available at https://github.com/RSLLES/SHOT.

Exoplanet imaging is a major challenge in astrophysics due to the need for high angular resolution and high contrast. We present a multi-scale statistical model for the nuisance component corrupting multivariate image series at high contrast. Integrated into a learnable architecture, it leverages the physics of the problem and enables the fusion of multiple observations of the same star in a way that is optimal in terms of detection signal-to-noise ratio. Applied to data from the VLT/SPHERE instrument, the method significantly improves the detection sensitivity and the accuracy of astrometric and photometric estimation.

This paper explores feature prediction as a stand-alone objective for unsupervised learning from video and introduces V-JEPA, a collection of vision models trained solely using a feature prediction objective, without the use of pretrained image encoders, text, negative examples, reconstruction, or other sources of supervision. The models are trained on 2 million videos collected from public datasets and are evaluated on downstream image and video tasks. Our results show that learning by predicting video features leads to versatile visual representations that perform well on both motion and appearance-based tasks, without adaption of the model's parameters; e.g., using a frozen backbone. Our largest model, a ViT-H/16 trained only on videos, obtains 81.9% on Kinetics-400, 72.2% on Something-Something-v2, and 77.9% on ImageNet1K.

This paper addresses the Object Goal Navigation problem, where a robot must efficiently find a target object in an unknown environment. Existing implicit memory-based methods struggle with long-term memory retention and planning, while explicit map-based approaches lack rich semantic information. To address these challenges, we propose FOM-Nav, a modular framework that enhances exploration efficiency through Frontier-Object Maps and vision-language models. Our Frontier-Object Maps are built online and jointly encode spatial frontiers and fine-grained object information. Using this representation, a vision-language model performs multimodal scene understanding and high-level goal prediction, which is executed by a low-level planner for efficient trajectory generation. To train FOM-Nav, we automatically construct large-scale navigation datasets from real-world scanned environments. Extensive experiments validate the effectiveness of our model design and constructed dataset. FOM-Nav achieves state-of-the-art performance on the MP3D and HM3D benchmarks, particularly in navigation efficiency metric SPL, and yields promising results on a real robot.

The composed image retrieval (CIR) task is to retrieve target images given a reference image and a modification text. Recent methods for CIR leverage large pretrained vision-language models (VLMs) and achieve good performance on general-domain concepts like color and texture. However, they still struggle with application domains like fashion, because the rich and diverse vocabulary used in fashion requires specific fine-grained vision and language understanding. An additional difficulty is the lack of large-scale fashion datasets with detailed and relevant annotations, due to the expensive cost of manual annotation by specialists. To address these challenges, we introduce FACap, a large-scale, automatically constructed fashion-domain CIR dataset. It leverages web-sourced fashion images and a two-stage annotation pipeline powered by a VLM and a large language model (LLM) to generate accurate and detailed modification texts. Then, we propose a new CIR model FashionBLIP-2, which fine-tunes the general-domain BLIP-2 model on FACap with lightweight adapters and multi-head query-candidate matching to better account for fine-grained fashion-specific information. FashionBLIP-2 is evaluated with and without additional fine-tuning on the Fashion IQ benchmark and the enhanced evaluation dataset enhFashionIQ, leveraging our pipeline to obtain higher-quality annotations. Experimental results show that the combination of FashionBLIP-2 and pretraining with FACap significantly improves the model's performance in fashion CIR especially for retrieval with fine-grained modification texts, demonstrating the value of our dataset and approach in a highly demanding environment such as e-commerce websites. Code is available at https://fgxaos.github.io/facap-paper-website/.

Archaeological sites are the physical remains of past human activity and one of the main sources of information about past societies and cultures. However, they are also the target of malevolent human actions, especially in countries having experienced inner turmoil and conflicts. Because monitoring these sites from space is a key step towards their preservation, we introduce the DAFA Looted Sites dataset, \datasetname, a labeled multi-temporal remote sensing dataset containing 55,480 images acquired monthly over 8 years across 675 Afghan archaeological sites, including 135 sites looted during the acquisition period. \datasetname~is particularly challenging because of the limited number of training samples, the class imbalance, the weak binary annotations only available at the level of the time series, and the subtlety of relevant changes coupled with important irrelevant ones over a long time period. It is also an interesting playground to assess the performance of satellite image time series (SITS) classification methods on a real and important use case. We evaluate a large set of baselines, outline the substantial benefits of using foundation models and show the additional boost that can be provided by using complete time series instead of using a single image.

Identifying the underlying dynamics of actuated physical systems from sensory input is of high interest in control, robotics, and engineering in general. In the context of control problems, existing approaches decouple the construction of the feature space where the dynamics identification process occurs from the target control tasks, potentially leading to a mismatch between feature and state spaces: the systems may not be controllable in feature space, and synthesized controls may not be applicable in state space. Borrowing from the Koopman formalism, we propose instead to learn an embedding of both the states and controls into a feature space where the dynamics are linear, and include the target control task in the learning objective in the form of a differentiable and robust optimal control problem. We validate the proposed approach with simulation experiments using systems with non-linear dynamics, demonstrating that the controls obtained in feature space can be used to drive the corresponding physical systems and that the learned model can serve for future state prediction.

Satellite imagery plays a crucial role in monitoring changes happening on Earth's surface and aiding in climate analysis, ecosystem assessment, and disaster response. In this paper, we tackle semantic change detection with satellite image time series (SITS-SCD) which encompasses both change detection and semantic segmentation tasks. We propose a new architecture that improves over the state of the art, scales better with the number of parameters, and leverages long-term temporal information. However, for practical use cases, models need to adapt to spatial and temporal shifts, which remains a challenge. We investigate the impact of temporal and spatial shifts separately on global, multi-year SITS datasets using DynamicEarthNet and MUDS. We show that the spatial domain shift represents the most complex setting and that the impact of temporal shift on performance is more pronounced on change detection than on semantic segmentation, highlighting that it is a specific issue deserving further attention.

Blind image quality assessment (BIQA) approaches, while promising for automating image quality evaluation, often fall short in real-world scenarios due to their reliance on a generic quality standard applied uniformly across diverse images. This one-size-fits-all approach overlooks the crucial perceptual relationship between image content and quality, leading to a 'domain shift' challenge where a single quality metric inadequately represents various content types. Furthermore, BIQA techniques typically overlook the inherent differences in the human visual system among different observers. In response to these challenges, this paper introduces PICNIQ, an innovative pairwise comparison framework designed to bypass the limitations of conventional BIQA by emphasizing relative, rather than absolute, quality assessment. PICNIQ is specifically designed to assess the quality differences between image pairs. The proposed framework implements a carefully crafted deep learning architecture, a specialized loss function, and a training strategy optimized for sparse comparison settings. By employing psychometric scaling algorithms like TrueSkill, PICNIQ transforms pairwise comparisons into just-objectionable-difference (JOD) quality scores, offering a granular and interpretable measure of image quality. We conduct our research using comparison matrices from the PIQ23 dataset, which are published in this paper. Our extensive experimental analysis showcases PICNIQ's broad applicability and superior performance over existing models, highlighting its potential to set new standards in the field of BIQA.