Vision-language models (VLMs), e.g., CLIP, have shown remarkable potential in zero-shot image classification. However, adapting these models to new domains remains challenging, especially in unsupervised settings where labeled data is unavailable. Recent research has proposed pseudo-labeling approaches to adapt CLIP in an unsupervised manner using unlabeled target data. Nonetheless, these methods struggle due to noisy pseudo-labels resulting from the misalignment between CLIP's visual and textual representations. This study introduces DPA, an unsupervised domain adaptation method for VLMs. DPA introduces the concept of dual prototypes, acting as distinct classifiers, along with the convex combination of their outputs, thereby leading to accurate pseudo-label construction. Next, it ranks pseudo-labels to facilitate robust self-training, particularly during early training. Finally, it addresses visual-textual misalignment by aligning textual prototypes with image prototypes to further improve the adaptation performance. Experiments on 13 downstream vision tasks demonstrate that DPA significantly outperforms zero-shot CLIP and the state-of-the-art unsupervised adaptation baselines.
Recent advances in vision-language foundational models have enabled development of systems that can perform visual understanding and reasoning tasks. However, it is unclear if these models are robust to distribution shifts, and how their performance and generalization capabilities vary under changes in data distribution. In this project we strive to answer the question "Are vision-language foundational models robust to distribution shifts like human perception?" Specifically, we consider a diverse range of vision-language models and compare how the performance of these systems is affected by corruption based distribution shifts (such as \textit{motion blur, fog, snow, gaussian noise}) commonly found in practical real-world scenarios. We analyse the generalization capabilities qualitatively and quantitatively on zero-shot image classification task under aforementioned distribution shifts. Our code will be avaible at \url{https://github.com/shivam-chandhok/CPSC-540-Project}
We propose a novel prompt tuning method called CoAPT(Context Attribute words in Prompt Tuning) for few/zero-shot image classification. The core motivation is that attributes are descriptive words with rich information about a given concept. Thus, we aim to enrich text queries of existing prompt tuning methods, improving alignment between text and image embeddings in CLIP embedding space. To do so, CoAPT integrates attribute words as additional prompts within learnable prompt tuning and can be easily incorporated into various existing prompt tuning methods. To facilitate the incorporation of attributes into text embeddings and the alignment with image embeddings, soft prompts are trained together with an additional meta-network that generates input-image-wise feature biases from the concatenated feature encodings of the image-text combined queries. Our experiments demonstrate that CoAPT leads to considerable improvements for existing baseline methods on several few/zero-shot image classification tasks, including base-to-novel generalization, cross-dataset transfer, and domain generalization. Our findings highlight the importance of combining hard and soft prompts and pave the way for future research on the interplay between text and image latent spaces in pre-trained models.
We introduce NOVIC, an innovative real-time uNconstrained Open Vocabulary Image Classifier that uses an autoregressive transformer to generatively output classification labels as language. Leveraging the extensive knowledge of CLIP models, NOVIC harnesses the embedding space to enable zero-shot transfer from pure text to images. Traditional CLIP models, despite their ability for open vocabulary classification, require an exhaustive prompt of potential class labels, restricting their application to images of known content or context. To address this, we propose an "object decoder" model that is trained on a large-scale 92M-target dataset of templated object noun sets and LLM-generated captions to always output the object noun in question. This effectively inverts the CLIP text encoder and allows textual object labels from essentially the entire English language to be generated directly from image-derived embedding vectors, without requiring any a priori knowledge of the potential content of an image, and without any label biases. The trained decoders are tested on a mix of manually and web-curated datasets, as well as standard image classification benchmarks, and achieve fine-grained prompt-free prediction scores of up to 87.5%, a strong result considering the model must work for any conceivable image and without any contextual clues.
In the field of vision-language contrastive learning, models such as CLIP capitalize on matched image-caption pairs as positive examples and leverage within-batch non-matching pairs as negatives. This approach has led to remarkable outcomes in zero-shot image classification, cross-modal retrieval, and linear evaluation tasks. We show that the zero-shot classification and retrieval capabilities of CLIP-like models can be improved significantly through the introduction of semantically composite examples during pretraining. Inspired by CutMix in vision categorization, we create semantically composite image-caption pairs by merging elements from two distinct instances in the dataset via a novel procedure. Our method fuses the captions and blends 50% of each image to form a new composite sample. This simple technique (termed CLIP-C for CLIP Compositions), devoid of any additional computational overhead or increase in model parameters, significantly improves zero-shot image classification and cross-modal retrieval. The benefits of CLIP-C are particularly pronounced in settings with relatively limited pretraining data.
Vision Language Models (VLMs) like CLIP have attracted substantial attention in pathology, serving as backbones for applications such as zero-shot image classification and Whole Slide Image (WSI) analysis. Additionally, they can function as vision encoders when combined with large language models (LLMs) to support broader capabilities. Current efforts to train pathology VLMs rely on pathology image-text pairs from platforms like PubMed, YouTube, and Twitter, which provide limited, unscalable data with generally suboptimal image quality. In this work, we leverage large-scale WSI datasets like TCGA to extract numerous high-quality image patches. We then train a large multimodal model to generate captions for these images, creating PathGen-1.6M, a dataset containing 1.6 million high-quality image-caption pairs. Our approach involves multiple agent models collaborating to extract representative WSI patches, generating and refining captions to obtain high-quality image-text pairs. Extensive experiments show that integrating these generated pairs with existing datasets to train a pathology-specific CLIP model, PathGen-CLIP, significantly enhances its ability to analyze pathological images, with substantial improvements across nine pathology-related zero-shot image classification tasks and three whole-slide image tasks. Furthermore, we construct 200K instruction-tuning data based on PathGen-1.6M and integrate PathGen-CLIP with the Vicuna LLM to create more powerful multimodal models through instruction tuning. Overall, we provide a scalable pathway for high-quality data generation in pathology, paving the way for next-generation general pathology models.
Decoding images from non-invasive electroencephalographic (EEG) signals has been a grand challenge in understanding how the human brain process visual information in real-world scenarios. To cope with the issues of signal-to-noise ratio and nonstationarity, this paper introduces a MUltimodal Similarity-keeping contrastivE learning (MUSE) framework for zero-shot EEG-based image classification. We develop a series of multivariate time-series encoders tailored for EEG signals and assess the efficacy of regularized contrastive EEG-Image pretraining using an extensive visual EEG dataset. Our method achieves state-of-the-art performance, with a top-1 accuracy of 19.3% and a top-5 accuracy of 48.8% in 200-way zero-shot image classification. Furthermore, we visualize neural patterns via model interpretation, shedding light on the visual processing dynamics in the human brain. The code repository for this work is available at: https://github.com/ChiShengChen/MUSE_EEG.
Contrastive Language--Image Pre-training (CLIP) has manifested remarkable improvements in zero-shot classification and cross-modal vision-language tasks. Yet, from a geometrical point of view, the CLIP embedding space has been found to have a pronounced modality gap. This gap renders the embedding space overly sparse and disconnected, with different modalities being densely distributed in distinct subregions of the hypersphere. In this work, we aim at answering three main questions: 1. Does sharing the parameter space between the multi-modal encoders reduce the modality gap? 2. Can the gap be mitigated by pushing apart the uni-modal embeddings via intra-modality separation? 3. How do these gap reduction approaches affect the downstream performance? We design AlignCLIP, in order to answer these questions and through extensive experiments, we show that AlignCLIP achieves noticeable enhancements in the cross-modal alignment of the embeddings, and thereby, reduces the modality gap, while improving the performance across several zero-shot and fine-tuning downstream evaluations.
Vision-Language Models (VLMs) such as CLIP have yielded unprecedented performance for zero-shot image classification, yet their generalization capability may still be seriously challenged when confronted to domain shifts. In response, we present Weight Average Test-Time Adaptation (WATT) of CLIP, a pioneering approach facilitating full test-time adaptation (TTA) of this VLM. Our method employs a diverse set of templates for text prompts, augmenting the existing framework of CLIP. Predictions are utilized as pseudo labels for model updates, followed by weight averaging to consolidate the learned information globally. Furthermore, we introduce a text ensemble strategy, enhancing overall test performance by aggregating diverse textual cues. Our findings underscore the efficacy of WATT in enhancing performance across diverse datasets, including CIFAR-10-C, CIFAR-10.1, CIFAR-100-C, VisDA-C, and several other challenging datasets, effectively covering a wide range of domain shifts. Notably, these enhancements are achieved without necessitating additional model transformations or trainable modules. Moreover, compared to other Test-Time Adaptation methods, our approach can operate effectively with just a single image. Highlighting the potential of innovative test-time strategies, this research emphasizes their role in fortifying the adaptability of VLMs. The implementation is available at: \url{https://github.com/Mehrdad-Noori/WATT.git}.
Large-scale pretrained vision-language models like CLIP have demonstrated remarkable zero-shot image classification capabilities across diverse domains. To enhance CLIP's performance while preserving the zero-shot paradigm, various test-time prompt tuning methods have been introduced to refine class embeddings through unsupervised learning objectives during inference. However, these methods often encounter challenges in selecting appropriate learning rates to prevent collapsed training in the absence of validation data during test-time adaptation. In this study, we propose a novel backpropagation-free algorithm BaFTA for test-time adaptation of vision-language models. Instead of fine-tuning text prompts to refine class embeddings, our approach directly estimates class centroids using online clustering within a projected embedding space that aligns text and visual embeddings. We dynamically aggregate predictions from both estimated and original class embeddings, as well as from distinct augmented views, by assessing the reliability of each prediction using R\'enyi Entropy. Through extensive experiments, we demonstrate that BaFTA consistently outperforms state-of-the-art test-time adaptation methods in both effectiveness and efficiency.
Multi-modal contrastive models such as CLIP achieve state-of-the-art performance in zero-shot classification by embedding input images and texts on a joint representational space. Recently, a modality gap has been reported in two-encoder contrastive models like CLIP, meaning that the image and text embeddings reside in disjoint areas of the latent space. Previous studies suggest that this gap exists due to 1) the cone effect, 2) mismatched pairs in the dataset, and 3) insufficient training. We show that, even when accounting for all these factors, and even when using the same modality, the contrastive loss actually creates a gap during training. As a result, We propose that the modality gap is inherent to the two-encoder contrastive loss and rename it the contrastive gap. We present evidence that attributes this contrastive gap to low uniformity in CLIP space, resulting in embeddings that occupy only a small portion of the latent space. To close the gap, we adapt the uniformity and alignment properties of unimodal contrastive loss to the multi-modal setting and show that simply adding these terms to the CLIP loss distributes the embeddings more uniformly in the representational space, closing the gap. In our experiments, we show that the modified representational space achieves better performance than default CLIP loss in downstream tasks such as zero-shot image classification and multi-modal arithmetic.
Pre-trained vision-language (VL) models are highly vulnerable to adversarial attacks. However, existing defense methods primarily focus on image classification, overlooking two key aspects of VL tasks: multimodal attacks, where both image and text can be perturbed, and the one-to-many relationship of images and texts, where a single image can correspond to multiple textual descriptions and vice versa (1:N and N:1). This work is the first to explore defense strategies against multimodal attacks in VL tasks, whereas prior VL defense methods focus on vision robustness. We propose multimodal adversarial training (MAT), which incorporates adversarial perturbations in both image and text modalities during training, significantly outperforming existing unimodal defenses. Furthermore, we discover that MAT is limited by deterministic one-to-one (1:1) image-text pairs in VL training data. To address this, we conduct a comprehensive study on leveraging one-to-many relationships to enhance robustness, investigating diverse augmentation techniques. Our analysis shows that, for a more effective defense, augmented image-text pairs should be well-aligned, diverse, yet avoid distribution shift -- conditions overlooked by prior research. Our experiments show that MAT can effectively be applied to different VL models and tasks to improve adversarial robustness, outperforming previous efforts. Our code will be made public upon acceptance.
Large language models (LLMs) have been effectively used for many computer vision tasks, including image classification. In this paper, we present a simple yet effective approach for zero-shot image classification using multimodal LLMs. Using multimodal LLMs, we generate comprehensive textual representations from input images. These textual representations are then utilized to generate fixed-dimensional features in a cross-modal embedding space. Subsequently, these features are fused together to perform zero-shot classification using a linear classifier. Our method does not require prompt engineering for each dataset; instead, we use a single, straightforward set of prompts across all datasets. We evaluated our method on several datasets and our results demonstrate its remarkable effectiveness, surpassing benchmark accuracy on multiple datasets. On average, for ten benchmarks, our method achieved an accuracy gain of 6.2 percentage points, with an increase of 6.8 percentage points on the ImageNet dataset, compared to prior methods re-evaluated with the same setup. Our findings highlight the potential of multimodal LLMs to enhance computer vision tasks such as zero-shot image classification, offering a significant improvement over traditional methods.
As training datasets become increasingly drawn from unstructured, uncontrolled environments such as the web, researchers and industry practitioners have increasingly relied upon data filtering techniques to "filter out the noise" of web-scraped data. While datasets have been widely shown to reflect the biases and values of their creators, in this paper we contribute to an emerging body of research that assesses the filters used to create these datasets. We show that image-text data filtering also has biases and is value-laden, encoding specific notions of what is counted as "high-quality" data. In our work, we audit a standard approach of image-text CLIP-filtering on the academic benchmark DataComp's CommonPool by analyzing discrepancies of filtering through various annotation techniques across multiple modalities of image, text, and website source. We find that data relating to several imputed demographic groups -- such as LGBTQ+ people, older women, and younger men -- are associated with higher rates of exclusion. Moreover, we demonstrate cases of exclusion amplification: not only are certain marginalized groups already underrepresented in the unfiltered data, but CLIP-filtering excludes data from these groups at higher rates. The data-filtering step in the machine learning pipeline can therefore exacerbate representation disparities already present in the data-gathering step, especially when existing filters are designed to optimize a specifically-chosen downstream performance metric like zero-shot image classification accuracy. Finally, we show that the NSFW filter fails to remove sexually-explicit content from CommonPool, and that CLIP-filtering includes several categories of copyrighted content at high rates. Our conclusions point to a need for fundamental changes in dataset creation and filtering practices.
This paper introduces a novel framework for zero-shot learning (ZSL), i.e., to recognize new categories that are unseen during training, by using a multi-model and multi-alignment integration method. Specifically, we propose three strategies to enhance the model's performance to handle ZSL: 1) Utilizing the extensive knowledge of ChatGPT and the powerful image generation capabilities of DALL-E to create reference images that can precisely describe unseen categories and classification boundaries, thereby alleviating the information bottleneck issue; 2) Integrating the results of text-image alignment and image-image alignment from CLIP, along with the image-image alignment results from DINO, to achieve more accurate predictions; 3) Introducing an adaptive weighting mechanism based on confidence levels to aggregate the outcomes from different prediction methods. Experimental results on multiple datasets, including CIFAR-10, CIFAR-100, and TinyImageNet, demonstrate that our model can significantly improve classification accuracy compared to single-model approaches, achieving AUROC scores above 96% across all test datasets, and notably surpassing 99% on the CIFAR-10 dataset.
The success of contrastive language-image pretraining (CLIP) relies on the supervision from the pairing between images and captions, which tends to be noisy in web-crawled data. We present Mixture of Data Experts (MoDE) and learn a system of CLIP data experts via clustering. Each data expert is trained on one data cluster, being less sensitive to false negative noises in other clusters. At inference time, we ensemble their outputs by applying weights determined through the correlation between task metadata and cluster conditions. To estimate the correlation precisely, the samples in one cluster should be semantically similar, but the number of data experts should still be reasonable for training and inference. As such, we consider the ontology in human language and propose to use fine-grained cluster centers to represent each data expert at a coarse-grained level. Experimental studies show that four CLIP data experts on ViT-B/16 outperform the ViT-L/14 by OpenAI CLIP and OpenCLIP on zero-shot image classification but with less ($<$35\%) training cost. Meanwhile, MoDE can train all data expert asynchronously and can flexibly include new data experts. The code is available at https://github.com/facebookresearch/MetaCLIP/tree/main/mode.
Pre-trained vision-language (V-L) models such as CLIP have shown excellent performance in many downstream cross-modal tasks. However, most of them are only applicable to the English context. Subsequent research has focused on this problem and proposed improved models, such as CN-CLIP and AltCLIP, to facilitate their applicability to Chinese and even other languages. Nevertheless, these models suffer from high latency and a large memory footprint in inference, which limits their further deployment on resource-constrained edge devices. In this work, we propose a conceptually simple yet effective multilingual CLIP Compression framework and train a lightweight multilingual vision-language model, called DC-CLIP, for both Chinese and English context. In this framework, we collect high-quality Chinese and English text-image pairs and design two training stages, including multilingual vision-language feature distillation and alignment. During the first stage, lightweight image/text student models are designed to learn robust visual/multilingual textual feature representation ability from corresponding teacher models, respectively. Subsequently, the multilingual vision-language alignment stage enables effective alignment of visual and multilingual textual features to further improve the model's multilingual performance. Comprehensive experiments in zero-shot image classification, conducted based on the ELEVATER benchmark, showcase that DC-CLIP achieves superior performance in the English context and competitive performance in the Chinese context, even with less training data, when compared to existing models of similar parameter magnitude. The evaluation demonstrates the effectiveness of our designed training mechanism.
We propose a novel architecture and method of explainable classification with Concept Bottleneck Models (CBMs). While SOTA approaches to Image Classification task work as a black box, there is a growing demand for models that would provide interpreted results. Such a models often learn to predict the distribution over class labels using additional description of this target instances, called concepts. However, existing Bottleneck methods have a number of limitations: their accuracy is lower than that of a standard model and CBMs require an additional set of concepts to leverage. We provide a framework for creating Concept Bottleneck Model from pre-trained multi-modal encoder and new CLIP-like architectures. By introducing a new type of layers known as Concept Bottleneck Layers, we outline three methods for training them: with $\ell_1$-loss, contrastive loss and loss function based on Gumbel-Softmax distribution (Sparse-CBM), while final FC layer is still trained with Cross-Entropy. We show a significant increase in accuracy using sparse hidden layers in CLIP-based bottleneck models. Which means that sparse representation of concepts activation vector is meaningful in Concept Bottleneck Models. Moreover, with our Concept Matrix Search algorithm we can improve CLIP predictions on complex datasets without any additional training or fine-tuning. The code is available at: https://github.com/Andron00e/SparseCBM.
Existing vision-language models (VLMs) treat text descriptions as a unit, confusing individual concepts in a prompt and impairing visual semantic matching and reasoning. An important aspect of reasoning in logic and language is negations. This paper highlights the limitations of popular VLMs such as CLIP, at understanding the implications of negations, i.e., the effect of the word "not" in a given prompt. To enable evaluation of VLMs on fluent prompts with negations, we present CC-Neg, a dataset containing 228,246 images, true captions and their corresponding negated captions. Using CC-Neg along with modifications to the contrastive loss of CLIP, our proposed CoN-CLIP framework, has an improved understanding of negations. This training paradigm improves CoN-CLIP's ability to encode semantics reliably, resulting in 3.85% average gain in top-1 accuracy for zero-shot image classification across 8 datasets. Further, CoN-CLIP outperforms CLIP on challenging compositionality benchmarks such as SugarCREPE by 4.4%, showcasing emergent compositional understanding of objects, relations, and attributes in text. Overall, our work addresses a crucial limitation of VLMs by introducing a dataset and framework that strengthens semantic associations between images and text, demonstrating improved large-scale foundation models with significantly reduced computational cost, promoting efficiency and accessibility.
Vision Language Models (VLMs) excel in zero-shot image classification by pairing images with textual category names. The expanding variety of Pre-Trained VLMs enhances the likelihood of identifying a suitable VLM for specific tasks. To better reuse the VLM resource and fully leverage its potential on different zero-shot image classification tasks, a promising strategy is selecting appropriate Pre-Trained VLMs from the VLM Zoo, relying solely on the text data of the target dataset without access to the dataset's images. In this paper, we analyze two inherent challenges in assessing the ability of a VLM in this Language-Only VLM selection: the "Modality Gap" - the disparity in VLM's embeddings across two different modalities, making text a less reliable substitute for images; and the "Capability Gap" - the discrepancy between the VLM's overall ranking and its ranking for target dataset, hindering direct prediction of a model's dataset-specific performance from its general performance. We propose VLM Selection With gAp Bridging (SWAB) to mitigate the negative impact of two gaps. SWAB first adopts optimal transport to capture the relevance between open-source and target datasets with a transportation matrix. It then uses this matrix to transfer useful statistics of VLMs from open-source datasets to the target dataset for bridging two gaps. By bridging two gaps to obtain better substitutes for test images, SWAB can accurately predict the performance ranking of different VLMs on the target task without the need for the dataset's images. Experiments across various VLMs and image classification datasets validate SWAB's effectiveness.