Contrastive Language-Image Pretraining (CLIP) has achieved remarkable success in cross-modal tasks such as zero-shot image classification and text-image retrieval by effectively aligning visual and textual representations. However, the theoretical foundations underlying CLIP's strong generalization remain unclear. In this work, we address this gap by proposing the Cross-modal Information Bottleneck (CIB) framework. CIB offers a principled interpretation of CLIP's contrastive learning objective as an implicit Information Bottleneck optimization. Under this view, the model maximizes shared cross-modal information while discarding modality-specific redundancies, thereby preserving essential semantic alignment across modalities. Building on this insight, we introduce a Cross-modal Information Bottleneck Regularization (CIBR) method that explicitly enforces these IB principles during training. CIBR introduces a penalty term to discourage modality-specific redundancy, thereby enhancing semantic alignment between image and text features. We validate CIBR on extensive vision-language benchmarks, including zero-shot classification across seven diverse image datasets and text-image retrieval on MSCOCO and Flickr30K. The results show consistent performance gains over standard CLIP. These findings provide the first theoretical understanding of CLIP's generalization through the IB lens. They also demonstrate practical improvements, offering guidance for future cross-modal representation learning.
This study addresses the technical bottlenecks in handling long text and the "hallucination" issue caused by insufficient short text information in remote sensing vision-language foundation models (VLFM). We propose a novel vision-language foundation model, LRSCLIP, and a multimodal dataset, LRS2M. The main contributions are as follows: (1) By integrating multi-source remote sensing data and adopting a large language model labeling strategy, we construct the LRS2M dataset, which contains 2 million image-text pairs, providing both short and long texts for the first time, thus solving the problem of semantic granularity limitations in existing datasets; (2) The design of the LRSCLIP architecture based on Long-CLIP's KPS module, which extends CLIP's text processing capacity and achieves fine-grained cross-modal feature alignment through a dual-text loss weighting mechanism. Experimental results show that LRSCLIP improves retrieval accuracy by 10\%-20\% over the Long-CLIP baseline in the zero-shot long-text cross-modal retrieval task. For the zero-shot short-text cross-modal retrieval task, LRSCLIP achieves improvements over the current best model, GeoRSCLIP, with increases of 0.17\%, 0.67\%, and 0.92\% in Text to Image R@1, Image to Text R@1, and mR on RSITMD, respectively, and 0.04\%, 2.93\%, and 1.28\% on RSICD. In the zero-shot image classification task (average accuracy=75.75\%) and semantic localization task (Rmi=0.7653), LRSCLIP achieves state-of-the-art performance. These results validate the dual advantages of fine-grained semantic understanding and global feature matching in LRSCLIP. This work provides a new benchmark model and data support for remote sensing multimodal learning. The related code has been open source and is available at https://github.com/MitsuiChen14/LRSCLIP.
Vision-language models for Earth observation (EO) typically rely on the visual spectrum of data as the only model input, thus failing to leverage the rich spectral information available in the multispectral channels recorded by satellites. Therefore, in this paper, we introduce Llama3-MS-CLIP, the first vision-language model pre-trained with contrastive learning on a large-scale multispectral dataset and report on the performance gains due to the extended spectral range. Furthermore, we present the largest-to-date image-caption dataset for multispectral data, consisting of one million Sentinel-2 samples and corresponding textual descriptions generated with Llama3-LLaVA-Next and Overture Maps data. We develop a scalable captioning pipeline, which is validated by domain experts. We evaluate Llama3-MS-CLIP on multispectral zero-shot image classification and retrieval using three datasets of varying complexity. Our results demonstrate that Llama3-MS-CLIP significantly outperforms other RGB-based approaches, improving classification accuracy by 6.77% on average and retrieval performance by 4.63% mAP compared to the second-best model. Our results emphasize the relevance of multispectral vision-language learning. We release the image-caption dataset, code, and model weights under an open-source license.
Test-time adaptation with pre-trained vision-language models, such as CLIP, aims to adapt the model to new, potentially out-of-distribution test data. Existing methods calculate the similarity between visual embedding and learnable class embeddings, which are initialized by text embeddings, for zero-shot image classification. In this work, we first analyze this process based on Bayes theorem, and observe that the core factors influencing the final prediction are the likelihood and the prior. However, existing methods essentially focus on adapting class embeddings to adapt likelihood, but they often ignore the importance of prior. To address this gap, we propose a novel approach, \textbf{B}ayesian \textbf{C}lass \textbf{A}daptation (BCA), which in addition to continuously updating class embeddings to adapt likelihood, also uses the posterior of incoming samples to continuously update the prior for each class embedding. This dual updating mechanism allows the model to better adapt to distribution shifts and achieve higher prediction accuracy. Our method not only surpasses existing approaches in terms of performance metrics but also maintains superior inference rates and memory usage, making it highly efficient and practical for real-world applications.
Zero-shot learning (ZSL) aims to train a model on seen classes and recognize unseen classes by knowledge transfer through shared auxiliary information. Recent studies reveal that documents from encyclopedias provide helpful auxiliary information. However, existing methods align noisy documents, entangled in visual and non-visual descriptions, with image regions, yet solely depend on implicit learning. These models fail to filter non-visual noise reliably and incorrectly align non-visual words to image regions, which is harmful to knowledge transfer. In this work, we propose a novel multi-attribute document supervision framework to remove noises at both document collection and model learning stages. With the help of large language models, we introduce a novel prompt algorithm that automatically removes non-visual descriptions and enriches less-described documents in multiple attribute views. Our proposed model, MADS, extracts multi-view transferable knowledge with information decoupling and semantic interactions for semantic alignment at local and global levels. Besides, we introduce a model-agnostic focus loss to explicitly enhance attention to visually discriminative information during training, also improving existing methods without additional parameters. With comparable computation costs, MADS consistently outperforms the SOTA by 7.2% and 8.2% on average in three benchmarks for document-based ZSL and GZSL settings, respectively. Moreover, we qualitatively offer interpretable predictions from multiple attribute views.
Despite significant progress in Vision-Language Pre-training (VLP), current approaches predominantly emphasize feature extraction and cross-modal comprehension, with limited attention to generating or transforming visual content. This gap hinders the model's ability to synthesize coherent and novel visual representations from textual prompts, thereby reducing the effectiveness of multi-modal learning. In this work, we propose MedUnifier, a unified VLP framework tailored for medical data. MedUnifier seamlessly integrates text-grounded image generation capabilities with multi-modal learning strategies, including image-text contrastive alignment, image-text matching and image-grounded text generation. Unlike traditional methods that reply on continuous visual representations, our approach employs visual vector quantization, which not only facilitates a more cohesive learning strategy for cross-modal understanding but also enhances multi-modal generation quality by effectively leveraging discrete representations. Our framework's effectiveness is evidenced by the experiments on established benchmarks, including uni-modal tasks (supervised fine-tuning), cross-modal tasks (image-text retrieval and zero-shot image classification), and multi-modal tasks (medical report generation, image synthesis), where it achieves state-of-the-art performance across various tasks. MedUnifier also offers a highly adaptable tool for a wide range of language and vision tasks in healthcare, marking advancement toward the development of a generalizable AI model for medical applications.
Pre-trained multi-modal Vision-Language Models like CLIP are widely used off-the-shelf for a variety of applications. In this paper, we show that the common practice of individually exploiting the text or image encoders of these powerful multi-modal models is highly suboptimal for intra-modal tasks like image-to-image retrieval. We argue that this is inherently due to the CLIP-style inter-modal contrastive loss that does not enforce any intra-modal constraints, leading to what we call intra-modal misalignment. To demonstrate this, we leverage two optimization-based modality inversion techniques that map representations from their input modality to the complementary one without any need for auxiliary data or additional trained adapters. We empirically show that, in the intra-modal tasks of image-to-image and text-to-text retrieval, approaching these tasks inter-modally significantly improves performance with respect to intra-modal baselines on more than fifteen datasets. Additionally, we demonstrate that approaching a native inter-modal task (e.g. zero-shot image classification) intra-modally decreases performance, further validating our findings. Finally, we show that incorporating an intra-modal term in the pre-training objective or narrowing the modality gap between the text and image feature embedding spaces helps reduce the intra-modal misalignment. The code is publicly available at: https://github.com/miccunifi/Cross-the-Gap.
Visual Language Models such as CLIP excel in image recognition due to extensive image-text pre-training. However, applying the CLIP inference in zero-shot classification, particularly for medical image diagnosis, faces challenges due to: 1) the inadequacy of representing image classes solely with single category names; 2) the modal gap between the visual and text spaces generated by CLIP encoders. Despite attempts to enrich disease descriptions with large language models, the lack of class-specific knowledge often leads to poor performance. In addition, empirical evidence suggests that existing proxy learning methods for zero-shot image classification on natural image datasets exhibit instability when applied to medical datasets. To tackle these challenges, we introduce the Knowledge Proxy Learning (KPL) to mine knowledge from CLIP. KPL is designed to leverage CLIP's multimodal understandings for medical image classification through Text Proxy Optimization and Multimodal Proxy Learning. Specifically, KPL retrieves image-relevant knowledge descriptions from the constructed knowledge-enhanced base to enrich semantic text proxies. It then harnesses input images and these descriptions, encoded via CLIP, to stably generate multimodal proxies that boost the zero-shot classification performance. Extensive experiments conducted on both medical and natural image datasets demonstrate that KPL enables effective zero-shot image classification, outperforming all baselines. These findings highlight the great potential in this paradigm of mining knowledge from CLIP for medical image classification and broader areas.
Vision-language models (VLMs), such as CLIP and SigLIP, have found remarkable success in classification, retrieval, and generative tasks. For this, VLMs deterministically map images and text descriptions to a joint latent space in which their similarity is assessed using the cosine similarity. However, a deterministic mapping of inputs fails to capture uncertainties over concepts arising from domain shifts when used in downstream tasks. In this work, we propose post-hoc uncertainty estimation in VLMs that does not require additional training. Our method leverages a Bayesian posterior approximation over the last layers in VLMs and analytically quantifies uncertainties over cosine similarities. We demonstrate its effectiveness for uncertainty quantification and support set selection in active learning. Compared to baselines, we obtain improved and well-calibrated predictive uncertainties, interpretable uncertainty estimates, and sample-efficient active learning. Our results show promise for safety-critical applications of large-scale models.
Beyond the success of Contrastive Language-Image Pre-training (CLIP), recent trends mark a shift toward exploring the applicability of lightweight vision-language models for resource-constrained scenarios. These models often deliver suboptimal performance when relying solely on a single image-text contrastive learning objective, spotlighting the need for more effective training mechanisms that guarantee robust cross-modal feature alignment. In this work, we propose CLIP-PING: Contrastive Language-Image Pre-training with Proximus Intrinsic Neighbors Guidance, a novel yet simple and efficient training paradigm designed to boost the performance of lightweight vision-language models with minimal computational overhead and lower data demands. CLIP-PING bootstraps unimodal features extracted from arbitrary pre-trained encoders to obtain intrinsic guidance of proximus neighbor samples, i.e., nearest-neighbor (NN) and cross nearest-neighbor (XNN). We find that extra contrastive supervision from these neighbors substantially boosts cross-modal alignment, enabling lightweight models to learn more generic features with rich semantic diversity. Extensive experiments reveal that CLIP-PING notably surpasses its peers in zero-shot generalization and cross-modal retrieval tasks. Specifically, a 5.5% gain on zero-shot ImageNet1K classification with 10.7% (I2T) and 5.7% (T2I) on Flickr30K retrieval, compared to the original CLIP when using ViT-XS image encoder trained on 3 million (image, text) pairs. Moreover, CLIP-PING showcases a strong transferability under the linear evaluation protocol across several downstream tasks.
Contrastive Language-Image Pretraining (CLIP) models maximize the mutual information between text and visual modalities to learn representations. This makes the nature of the training data a significant factor in the efficacy of CLIP for downstream tasks. However, the lack of compositional diversity in contemporary image-text datasets limits the compositional reasoning ability of CLIP. We show that generating ``hard'' negative captions via in-context learning and synthesizing corresponding negative images with text-to-image generators offers a solution. We introduce a novel contrastive pre-training strategy that leverages these hard negative captions and images in an alternating fashion to train CLIP. We demonstrate that our method, named TripletCLIP, when applied to existing datasets such as CC3M and CC12M, enhances the compositional capabilities of CLIP, resulting in an absolute improvement of over 9% on the SugarCrepe benchmark on an equal computational budget, as well as improvements in zero-shot image classification and image retrieval. Our code, models, and data are available at: https://tripletclip.github.io
Low-resource domains, characterized by scarce data and annotations, present significant challenges for language and visual understanding tasks, with the latter much under-explored in the literature. Recent advancements in Vision-Language Models (VLM) have shown promising results in high-resource domains but fall short in low-resource concepts that are under-represented (e.g. only a handful of images per category) in the pre-training set. We tackle the challenging task of zero-shot low-resource image classification from a novel perspective. By leveraging a retrieval-based strategy, we achieve this in a training-free fashion. Specifically, our method, named CoRE (Combination of Retrieval Enrichment), enriches the representation of both query images and class prototypes by retrieving relevant textual information from large web-crawled databases. This retrieval-based enrichment significantly boosts classification performance by incorporating the broader contextual information relevant to the specific class. We validate our method on a newly established benchmark covering diverse low-resource domains, including medical imaging, rare plants, and circuits. Our experiments demonstrate that CORE outperforms existing state-of-the-art methods that rely on synthetic data generation and model fine-tuning.
We present TaxaBind, a unified embedding space for characterizing any species of interest. TaxaBind is a multimodal embedding space across six modalities: ground-level images of species, geographic location, satellite image, text, audio, and environmental features, useful for solving ecological problems. To learn this joint embedding space, we leverage ground-level images of species as a binding modality. We propose multimodal patching, a technique for effectively distilling the knowledge from various modalities into the binding modality. We construct two large datasets for pretraining: iSatNat with species images and satellite images, and iSoundNat with species images and audio. Additionally, we introduce TaxaBench-8k, a diverse multimodal dataset with six paired modalities for evaluating deep learning models on ecological tasks. Experiments with TaxaBind demonstrate its strong zero-shot and emergent capabilities on a range of tasks including species classification, cross-model retrieval, and audio classification. The datasets and models are made available at https://github.com/mvrl/TaxaBind.
Methods based on Contrastive Language-Image Pre-training (CLIP) are nowadays extensively used in support of vision-and-language tasks involving remote sensing data, such as cross-modal retrieval. The adaptation of CLIP to this specific domain has relied on model fine-tuning with the standard contrastive objective, using existing human-labeled image-caption datasets, or using synthetic data corresponding to image-caption pairs derived from other annotations over remote sensing images (e.g., object classes). The use of different pre-training mechanisms has received less attention, and only a few exceptions have considered multilingual inputs. This work proposes a novel vision-and-language model for the remote sensing domain, exploring the fine-tuning of a multilingual CLIP model and testing the use of a self-supervised method based on aligning local and global representations from individual input images, together with the standard CLIP objective. Model training relied on assembling pre-existing datasets of remote sensing images paired with English captions, followed by the use of automated machine translation into nine additional languages. We show that translated data is indeed helpful, e.g. improving performance also on English. Our resulting model, which we named Remote Sensing Multilingual CLIP (RS-M-CLIP), obtains state-of-the-art results in a variety of vision-and-language tasks, including cross-modal and multilingual image-text retrieval, or zero-shot image classification.
This paper focuses on creating synthetic data to improve the quality of image captions. Existing works typically have two shortcomings. First, they caption images from scratch, ignoring existing alt-text metadata, and second, lack transparency if the captioners' training data (e.g. GPT) is unknown. In this paper, we study a principled approach Altogether based on the key idea to edit and re-align existing alt-texts associated with the images. To generate training data, we perform human annotation where annotators start with the existing alt-text and re-align it to the image content in multiple rounds, consequently constructing captions with rich visual concepts. This differs from prior work that carries out human annotation as a one-time description task solely based on images and annotator knowledge. We train a captioner on this data that generalizes the process of re-aligning alt-texts at scale. Our results show our Altogether approach leads to richer image captions that also improve text-to-image generation and zero-shot image classification tasks.
Open-vocabulary object detection (OVD), detecting specific classes of objects using only their linguistic descriptions (e.g., class names) without any image samples, has garnered significant attention. However, in real-world applications, the target class concepts is often hard to describe in text and the only way to specify target objects is to provide their image examples, yet it is often challenging to obtain a good number of samples. Thus, there is a high demand from practitioners for few-shot object detection (FSOD). A natural question arises: Can the benefits of OVD extend to FSOD for object classes that are difficult to describe in text? Compared to traditional methods that learn only predefined classes (referred to in this paper as closed-set object detection, COD), can the extra cost of OVD be justified? To answer these questions, we propose a method to quantify the ``text-describability'' of object detection datasets using the zero-shot image classification accuracy with CLIP. This allows us to categorize various OD datasets with different text-describability and emprically evaluate the FSOD performance of OVD and COD methods within each category. Our findings reveal that: i) there is little difference between OVD and COD for object classes with low text-describability under equal conditions in OD pretraining; and ii) although OVD can learn from more diverse data than OD-specific data, thereby increasing the volume of training data, it can be counterproductive for classes with low-text-describability. These findings provide practitioners with valuable guidance amidst the recent advancements of OVD methods.
Contrastive Language-Image Pretraining (CLIP) performs zero-shot image classification by mapping images and textual class representation into a shared embedding space, then retrieving the class closest to the image. This work provides a new approach for interpreting CLIP models for image classification from the lens of mutual knowledge between the two modalities. Specifically, we ask: what concepts do both vision and language CLIP encoders learn in common that influence the joint embedding space, causing points to be closer or further apart? We answer this question via an approach of textual concept-based explanations, showing their effectiveness, and perform an analysis encompassing a pool of 13 CLIP models varying in architecture, size and pretraining datasets. We explore those different aspects in relation to mutual knowledge, and analyze zero-shot predictions. Our approach demonstrates an effective and human-friendly way of understanding zero-shot classification decisions with CLIP.
Multimodal encoders like CLIP excel in tasks such as zero-shot image classification and cross-modal retrieval. However, they require excessive training data. We propose canonical similarity analysis (CSA), which uses two unimodal encoders to replicate multimodal encoders using limited data. CSA maps unimodal features into a multimodal space, using a new similarity score to retain only the multimodal information. CSA only involves the inference of unimodal encoders and a cubic-complexity matrix decomposition, eliminating the need for extensive GPU-based model training. Experiments show that CSA outperforms CLIP while requiring $50,000\times$ fewer multimodal data pairs to bridge the modalities given pre-trained unimodal encoders on ImageNet classification and misinformative news caption detection. CSA surpasses the state-of-the-art method to map unimodal features to multimodal features. We also demonstrate the ability of CSA with modalities beyond image and text, paving the way for future modality pairs with limited paired multimodal data but abundant unpaired unimodal data, such as lidar and text.
State-of-the-art medical multi-modal large language models (med-MLLM), like LLaVA-Med or BioMedGPT, leverage instruction-following data in pre-training. However, those models primarily focus on scaling the model size and data volume to boost performance while mainly relying on the autoregressive learning objectives. Surprisingly, we reveal that such learning schemes might result in a weak alignment between vision and language modalities, making these models highly reliant on extensive pre-training datasets - a significant challenge in medical domains due to the expensive and time-consuming nature of curating high-quality instruction-following instances. We address this with LoGra-Med, a new multi-graph alignment algorithm that enforces triplet correlations across image modalities, conversation-based descriptions, and extended captions. This helps the model capture contextual meaning, handle linguistic variability, and build cross-modal associations between visuals and text. To scale our approach, we designed an efficient end-to-end learning scheme using black-box gradient estimation, enabling faster LLaMa 7B training. Our results show LoGra-Med matches LLAVA-Med performance on 600K image-text pairs for Medical VQA and significantly outperforms it when trained on 10% of the data. For example, on VQA-RAD, we exceed LLAVA-Med by 20.13% and nearly match the 100% pre-training score (72.52% vs. 72.64%). We also surpass SOTA methods like BiomedGPT on visual chatbots and RadFM on zero-shot image classification with VQA, highlighting the effectiveness of multi-graph alignment.
In recent years, Contrastive Language-Image Pre-training (CLIP) has become a cornerstone in multimodal intelligence. However, recent studies have identified that the information loss in the CLIP encoding process is substantial, and CLIP tends to capture only coarse-grained features from the input. This deficiency significantly limits the ability of a single CLIP model to handle images rich in visual detail. In this work, we propose a simple yet effective model-agnostic strategy, Diversified Multiplet Upcycling (DMU), for CLIP. DMU efficiently fine-tunes a series of CLIP models that capture different feature spaces, from a dense pre-trained CLIP checkpoint, sharing parameters except for the Feed-Forward Network (FFN). These models can then be transformed into a CLIP-MoE with a larger model capacity, leading to significantly enhanced performance with minimal computational overhead. To the best of our knowledge, Diversified Multiplet Upcycling is the first approach to introduce sparsely activated MoE into CLIP foundation models. Extensive experiments demonstrate the significant performance of CLIP-MoE across various zero-shot retrieval, zero-shot image classification tasks, and downstream Multimodal Large Language Model (MLLM) benchmarks by serving as a vision encoder. Furthermore, Diversified Multiplet Upcycling enables the conversion of any dense CLIP model into CLIP-MoEs, which can seamlessly replace CLIP in a plug-and-play manner without requiring further adaptation in downstream frameworks. Through Diversified Multiplet Upcycling, we aim to provide valuable insights for future research on developing more efficient and effective multimodal learning systems.