LLM-wrapper: Black-Box Semantic-Aware Adaptation of Vision-Language Foundation Models

Vision Language Models (VLMs) have shown impressive performances on numerous tasks but their zero-shot capabilities can be limited compared to dedicated or fine-tuned models. Yet, fine-tuning VLMs comes with limitations as it requires `white-box' access to the model's architecture and weights as well as expertise to design the fine-tuning objectives and optimize the hyper-parameters, which are specific to each VLM and downstream task. In this work, we propose LLM-wrapper, a novel approach to adapt VLMs in a `black-box' manner by leveraging large language models (LLMs) so as to reason on their outputs. We demonstrate the effectiveness of LLM-wrapper on Referring Expression Comprehension (REC), a challenging open-vocabulary task that requires spatial and semantic reasoning. Our approach significantly boosts the performance of off-the-shelf models, resulting in competitive results when compared with classic fine-tuning.

PointBeV: A Sparse Approach to BeV Predictions

Bird's-eye View (BeV) representations have emerged as the de-facto shared space in driving applications, offering a unified space for sensor data fusion and supporting various downstream tasks. However, conventional models use grids with fixed resolution and range and face computational inefficiencies due to the uniform allocation of resources across all cells. To address this, we propose PointBeV, a novel sparse BeV segmentation model operating on sparse BeV cells instead of dense grids. This approach offers precise control over memory usage, enabling the use of long temporal contexts and accommodating memory-constrained platforms. PointBeV employs an efficient two-pass strategy for training, enabling focused computation on regions of interest. At inference time, it can be used with various memory/performance trade-offs and flexibly adjusts to new specific use cases. PointBeV achieves state-of-the-art results on the nuScenes dataset for vehicle, pedestrian, and lane segmentation, showcasing superior performance in static and temporal settings despite being trained solely with sparse signals. We will release our code along with two new efficient modules used in the architecture: Sparse Feature Pulling, designed for the effective extraction of features from images to BeV, and Submanifold Attention, which enables efficient temporal modeling. Our code is available at https://github.com/valeoai/PointBeV.

Transductive Zero-Shot Learning using Cross-Modal CycleGAN

In Computer Vision, Zero-Shot Learning (ZSL) aims at classifying unseen classes -- classes for which no matching training image exists. Most of ZSL works learn a cross-modal mapping between images and class labels for seen classes. However, the data distribution of seen and unseen classes might differ, causing a domain shift problem. Following this observation, transductive ZSL (T-ZSL) assumes that unseen classes and their associated images are known during training, but not their correspondence. As current T-ZSL approaches do not scale efficiently when the number of seen classes is high, we tackle this problem with a new model for T-ZSL based upon CycleGAN. Our model jointly (i) projects images on their seen class labels with a supervised objective and (ii) aligns unseen class labels and visual exemplars with adversarial and cycle-consistency objectives. We show the efficiency of our Cross-Modal CycleGAN model (CM-GAN) on the ImageNet T-ZSL task where we obtain state-of-the-art results. We further validate CM-GAN on a language grounding task, and on a new task that we propose: zero-shot sentence-to-image matching on MS COCO.

Incorporating Visual Semantics into Sentence Representations within a Grounded Space

Language grounding is an active field aiming at enriching textual representations with visual information. Generally, textual and visual elements are embedded in the same representation space, which implicitly assumes a one-to-one correspondence between modalities. This hypothesis does not hold when representing words, and becomes problematic when used to learn sentence representations --- the focus of this paper --- as a visual scene can be described by a wide variety of sentences. To overcome this limitation, we propose to transfer visual information to textual representations by learning an intermediate representation space: the grounded space. We further propose two new complementary objectives ensuring that (1) sentences associated with the same visual content are close in the grounded space and (2) similarities between related elements are preserved across modalities. We show that this model outperforms the previous state-of-the-art on classification and semantic relatedness tasks.

Context-Aware Zero-Shot Learning for Object Recognition

Zero-Shot Learning (ZSL) aims at classifying unlabeled objects by leveraging auxiliary knowledge, such as semantic representations. A limitation of previous approaches is that only intrinsic properties of objects, e.g. their visual appearance, are taken into account while their context, e.g. the surrounding objects in the image, is ignored. Following the intuitive principle that objects tend to be found in certain contexts but not others, we propose a new and challenging approach, context-aware ZSL, that leverages semantic representations in a new way to model the conditional likelihood of an object to appear in a given context. Finally, through extensive experiments conducted on Visual Genome, we show that contextual information can substantially improve the standard ZSL approach and is robust to unbalanced classes.