Multimodal learning has been a field of increasing interest, aiming to combine various modalities in a single joint representation. Especially in the area of visiolinguistic (VL) learning multiple models and techniques have been developed, targeting a variety of tasks that involve images and text. VL models have reached unprecedented performances by extending the idea of Transformers, so that both modalities can learn from each other. Massive pre-training procedures enable VL models to acquire a certain level of real-world understanding, although many gaps can be identified: the limited comprehension of commonsense, factual, temporal and other everyday knowledge aspects questions the extendability of VL tasks. Knowledge graphs and other knowledge sources can fill those gaps by explicitly providing missing information, unlocking novel capabilities of VL models. In the same time, knowledge graphs enhance explainability, fairness and validity of decision making, issues of outermost importance for such complex implementations. The current survey aims to unify the fields of VL representation learning and knowledge graphs, and provides a taxonomy and analysis of knowledge-enhanced VL models.
Many high-level skills that are required for computer vision tasks, such as parsing questions, comparing and contrasting semantics, and writing descriptions, are also required in other domains such as natural language processing. In this paper, we ask whether it is possible to learn those skills from text data and then transfer them to vision tasks without ever training on visual training data. Key to our approach is exploiting the joint embedding space of contrastively trained vision and language encoders. In practice, there can be systematic differences between embedding spaces for different modalities in contrastive models, and we analyze how these differences affect our approach and study strategies to mitigate this concern. We produce models using only text training data on four representative tasks: image captioning, visual entailment, visual question answering and visual news captioning, and evaluate them on standard benchmarks using images. We find these models perform close to models trained on images, while surpassing prior work for captioning and visual entailment in this text-only setting by over 9 points, and outperforming all prior work on visual news by over 30 points. We also showcase a variety of stylistic image captioning models that are trained using no image data and no human-curated language data, but instead using readily-available text data from books, the web, or language models.
Visual entailment (VE) is to recognize whether the semantics of a hypothesis text can be inferred from the given premise image, which is one special task among recent emerged vision and language understanding tasks. Currently, most of the existing VE approaches are derived from the methods of visual question answering. They recognize visual entailment by quantifying the similarity between the hypothesis and premise in the content semantic features from multi modalities. Such approaches, however, ignore the VE's unique nature of relation inference between the premise and hypothesis. Therefore, in this paper, a new architecture called AlignVE is proposed to solve the visual entailment problem with a relation interaction method. It models the relation between the premise and hypothesis as an alignment matrix. Then it introduces a pooling operation to get feature vectors with a fixed size. Finally, it goes through the fully-connected layer and normalization layer to complete the classification. Experiments show that our alignment-based architecture reaches 72.45\% accuracy on SNLI-VE dataset, outperforming previous content-based models under the same settings.
Multimodal semantic understanding often has to deal with uncertainty, which means the obtained messages tend to refer to multiple targets. Such uncertainty is problematic for our interpretation, including inter- and intra-modal uncertainty. Little effort has studied the modeling of this uncertainty, particularly in pre-training on unlabeled datasets and fine-tuning in task-specific downstream datasets. In this paper, we project the representations of all modalities as probabilistic distributions via a Probability Distribution Encoder (PDE) by utilizing sequence-level interactions. Compared to the existing deterministic methods, such uncertainty modeling can convey richer multimodal semantic information and more complex relationships. Furthermore, we integrate uncertainty modeling with popular pre-training frameworks and propose suitable pre-training tasks: Distribution-based Vision-Language Contrastive learning (D-VLC), Distribution-based Masked Language Modeling (D-MLM), and Distribution-based Image-Text Matching (D-ITM). The fine-tuned models are applied to challenging downstream tasks, including image-text retrieval, visual question answering, visual reasoning, and visual entailment, and achieve state-of-the-art results.
We present a pre-training approach for vision and language transformer models, which is based on a mixture of diverse tasks. We explore both the use of image-text captioning data in pre-training, which does not need additional supervision, as well as object-aware strategies to pre-train the model. We evaluate the method on a number of textgenerative vision+language tasks, such as Visual Question Answering, visual entailment and captioning, and demonstrate large gains over standard pre-training methods.
Vision and Language Pretraining has become the prevalent approach for tackling multimodal downstream tasks. The current trend is to move towards ever larger models and pretraining datasets. This computational headlong rush does not seem reasonable in the long term to move toward sustainable solutions, and de facto excludes academic laboratories with limited resources. In this work, we propose a new framework, dubbed ViCHA, that efficiently exploits the input data to boost the learning by: (a) a new hierarchical cross-modal alignment loss, (b) new self-supervised scheme based on masked image modeling, (c) leveraging image-level annotations, called Visual Concepts, obtained with existing foundation models such as CLIP to boost the performance of the image encoder. Although pretrained on four times less data, our ViCHA strategy outperforms other approaches on several downstream tasks such as Image-Text Retrieval, VQA, Visual Reasoning, Visual Entailment and Visual Grounding. The code will be made publicly available here: https://github.com/mshukor/ViCHA
Prompt tuning has become a new paradigm for model tuning and it has demonstrated success in natural language pretraining and even vision pretraining. In this work, we explore the transfer of prompt tuning to multimodal pretraining, with a focus on generative multimodal pretrained models, instead of contrastive ones. Specifically, we implement prompt tuning on the unified sequence-to-sequence pretrained model adaptive to both understanding and generation tasks. Experimental results demonstrate that the light-weight prompt tuning can achieve comparable performance with finetuning and surpass other light-weight tuning methods. Besides, in comparison with finetuned models, the prompt-tuned models demonstrate improved robustness against adversarial attacks. We further figure out that experimental factors, including the prompt length, prompt depth, and reparameteratization, have great impacts on the model performance, and thus we empirically provide a recommendation for the setups of prompt tuning. Despite the observed advantages, we still find some limitations in prompt tuning, and we correspondingly point out the directions for future studies. Codes are available at \url{https://github.com/OFA-Sys/OFA}
Visual Entailment with natural language explanations aims to infer the relationship between a text-image pair and generate a sentence to explain the decision-making process. Previous methods rely mainly on a pre-trained vision-language model to perform the relation inference and a language model to generate the corresponding explanation. However, the pre-trained vision-language models mainly build token-level alignment between text and image yet ignore the high-level semantic alignment between the phrases (chunks) and visual contents, which is critical for vision-language reasoning. Moreover, the explanation generator based only on the encoded joint representation does not explicitly consider the critical decision-making points of relation inference. Thus the generated explanations are less faithful to visual-language reasoning. To mitigate these problems, we propose a unified Chunk-aware Alignment and Lexical Constraint based method, dubbed as CALeC. It contains a Chunk-aware Semantic Interactor (arr. CSI), a relation inferrer, and a Lexical Constraint-aware Generator (arr. LeCG). Specifically, CSI exploits the sentence structure inherent in language and various image regions to build chunk-aware semantic alignment. Relation inferrer uses an attention-based reasoning network to incorporate the token-level and chunk-level vision-language representations. LeCG utilizes lexical constraints to expressly incorporate the words or chunks focused by the relation inferrer into explanation generation, improving the faithfulness and informativeness of the explanations. We conduct extensive experiments on three datasets, and experimental results indicate that CALeC significantly outperforms other competitor models on inference accuracy and quality of generated explanations.
Data augmentation is a necessity to enhance data efficiency in deep learning. For vision-language pre-training, data is only augmented either for images or for text in previous works. In this paper, we present MixGen: a joint data augmentation for vision-language representation learning to further improve data efficiency. It generates new image-text pairs with semantic relationships preserved by interpolating images and concatenating text. It's simple, and can be plug-and-played into existing pipelines. We evaluate MixGen on four architectures, including CLIP, ViLT, ALBEF and TCL, across five downstream vision-language tasks to show its versatility and effectiveness. For example, adding MixGen in ALBEF pre-training leads to absolute performance improvements on downstream tasks: image-text retrieval (+6.2% on COCO fine-tuned and +5.3% on Flicker30K zero-shot), visual grounding (+0.9% on RefCOCO+), visual reasoning (+$0.9% on NLVR2), visual question answering (+0.3% on VQA2.0), and visual entailment (+0.4% on SNLI-VE).
Exploring large-scale pretrained foundation models is of significant interest in computer vision because these models can be quickly transferred to many downstream tasks. This paper presents Contrastive Captioner (CoCa), a minimalist design to pretrain an image-text encoder-decoder foundation model jointly with contrastive loss and captioning loss, thereby subsuming model capabilities from contrastive approaches like CLIP and generative methods like SimVLM. In contrast to standard encoder-decoder transformers where all decoder layers attend to encoder outputs, CoCa omits cross-attention in the first half of decoder layers to encode unimodal text representations, and cascades the remaining decoder layers which cross-attend to the image encoder for multimodal image-text representations. We apply a contrastive loss between unimodal image and text embeddings, in addition to a captioning loss on the multimodal decoder outputs which predicts text tokens autoregressively. By sharing the same computational graph, the two training objectives are computed efficiently with minimal overhead. CoCa is pretrained end-to-end and from scratch on both web-scale alt-text data and annotated images by treating all labels simply as text, seamlessly unifying natural language supervision for representation learning. Empirically, CoCa achieves state-of-the-art performance with zero-shot transfer or minimal task-specific adaptation on a broad range of downstream tasks, spanning visual recognition (ImageNet, Kinetics-400/600/700, Moments-in-Time), crossmodal retrieval (MSCOCO, Flickr30K, MSR-VTT), multimodal understanding (VQA, SNLI-VE, NLVR2), and image captioning (MSCOCO, NoCaps). Notably on ImageNet classification, CoCa obtains 86.3% zero-shot top-1 accuracy, 90.6% with a frozen encoder and learned classification head, and new state-of-the-art 91.0% top-1 accuracy on ImageNet with a finetuned encoder.
We present Answer-Me, a task-aware multi-task framework which unifies a variety of question answering tasks, such as, visual question answering, visual entailment, visual reasoning. In contrast to previous works using contrastive or generative captioning training, we propose a novel and simple recipe to pre-train a vision-language joint model, which is multi-task as well. The pre-training uses only noisy image captioning data, and is formulated to use the entire architecture end-to-end with both a strong language encoder and decoder. Our results show state-of-the-art performance, zero-shot generalization, robustness to forgetting, and competitive single-task results across a variety of question answering tasks. Our multi-task mixture training learns from tasks of various question intents and thus generalizes better, including on zero-shot vision-language tasks. We conduct experiments in the challenging multi-task and open-vocabulary settings and across a variety of datasets and tasks, such as VQA2.0, SNLI-VE, NLVR2, GQA. We observe that the proposed approach is able to generalize to unseen tasks and that more diverse mixtures lead to higher accuracy in both known and novel tasks.
Spatial relations are a basic part of human cognition. However, they are expressed in natural language in a variety of ways, and previous work has suggested that current vision-and-language models (VLMs) struggle to capture relational information. In this paper, we present Visual Spatial Reasoning (VSR), a dataset containing more than 10k natural text-image pairs with 66 types of spatial relations in English (such as: under, in front of, and facing). While using a seemingly simple annotation format, we show how the dataset includes challenging linguistic phenomena, such as varying reference frames. We demonstrate a large gap between human and model performance: the human ceiling is above 95%, while state-of-the-art models only achieve around 70%. We observe that VLMs' by-relation performances have little correlation with the number of training examples and the tested models are in general incapable of recognising relations concerning the orientations of objects.
Cross-modal encoders for vision-language (VL) tasks are often pretrained with carefully curated vision-language datasets. While these datasets reach an order of 10 million samples, the labor cost is prohibitive to scale further. Conversely, unimodal encoders are pretrained with simpler annotations that are less cost-prohibitive, achieving scales of hundreds of millions to billions. As a result, unimodal encoders have achieved state-of-art (SOTA) on many downstream tasks. However, challenges remain when applying to VL tasks. The pretraining data is not optimal for cross-modal architectures and requires heavy computational resources. In addition, unimodal architectures lack cross-modal interactions that have demonstrated significant benefits for VL tasks. Therefore, how to best leverage pretrained unimodal encoders for VL tasks is still an area of active research. In this work, we propose a method to leverage unimodal vision and text encoders for VL tasks that augment existing VL approaches while conserving computational complexity. Specifically, we propose Multimodal Adaptive Distillation (MAD), which adaptively distills useful knowledge from pretrained encoders to cross-modal VL encoders. Second, to better capture nuanced impacts on VL task performance, we introduce an evaluation protocol that includes Visual Commonsense Reasoning (VCR), Visual Entailment (SNLI-VE), and Visual Question Answering (VQA), across a variety of data constraints and conditions of domain shift. Experiments demonstrate that MAD leads to consistent gains in the low-shot, domain-shifted, and fully-supervised conditions on VCR, SNLI-VE, and VQA, achieving SOTA performance on VCR compared to other single models pretrained with image-text data. Finally, MAD outperforms concurrent works utilizing pretrained vision encoder from CLIP. Code will be made available.
Visual entailment is a recently proposed multimodal reasoning task where the goal is to predict the logical relationship of a piece of text to an image. In this paper, we propose an extension of this task, where the goal is to predict the logical relationship of fine-grained knowledge elements within a piece of text to an image. Unlike prior work, our method is inherently explainable and makes logical predictions at different levels of granularity. Because we lack fine-grained labels to train our method, we propose a novel multi-instance learning approach which learns a fine-grained labeling using only sample-level supervision. We also impose novel semantic structural constraints which ensure that fine-grained predictions are internally semantically consistent. We evaluate our method on a new dataset of manually annotated knowledge elements and show that our method achieves 68.18\% accuracy at this challenging task while significantly outperforming several strong baselines. Finally, we present extensive qualitative results illustrating our method's predictions and the visual evidence our method relied on. Our code and annotated dataset can be found here: https://github.com/SkrighYZ/FGVE.
CLIP has shown a remarkable zero-shot capability on a wide range of vision tasks. Previously, CLIP is only regarded as a powerful visual encoder. However, after being pre-trained by language supervision from a large amount of image-caption pairs, CLIP itself should also have acquired some few-shot abilities for vision-language tasks. In this work, we empirically show that CLIP can be a strong vision-language few-shot learner by leveraging the power of language. We first evaluate CLIP's zero-shot performance on a typical visual question answering task and demonstrate a zero-shot cross-modality transfer capability of CLIP on the visual entailment task. Then we propose a parameter-efficient fine-tuning strategy to boost the few-shot performance on the vqa task. We achieve competitive zero/few-shot results on the visual question answering and visual entailment tasks without introducing any additional pre-training procedure.
Natural language explanation (NLE) models aim at explaining the decision-making process of a black box system via generating natural language sentences which are human-friendly, high-level and fine-grained. Current NLE models explain the decision-making process of a vision or vision-language model (a.k.a., task model), e.g., a VQA model, via a language model (a.k.a., explanation model), e.g., GPT. Other than the additional memory resources and inference time required by the task model, the task and explanation models are completely independent, which disassociates the explanation from the reasoning process made to predict the answer. We introduce NLX-GPT, a general, compact and faithful language model that can simultaneously predict an answer and explain it. We first conduct pre-training on large scale data of image-caption pairs for general understanding of images, and then formulate the answer as a text prediction task along with the explanation. Without region proposals nor a task model, our resulting overall framework attains better evaluation scores, contains much less parameters and is 15$\times$ faster than the current SoA model. We then address the problem of evaluating the explanations which can be in many times generic, data-biased and can come in several forms. We therefore design 2 new evaluation measures: (1) explain-predict and (2) retrieval-based attack, a self-evaluation framework that requires no labels. Code is at: https://github.com/fawazsammani/nlxgpt.
Vision-and-Language (V+L) pre-training models have achieved tremendous success in recent years on various multi-modal benchmarks. However, the majority of existing models require pre-training on a large set of parallel image-text data, which is costly to collect, compared to image-only or text-only data. In this paper, we explore unsupervised Vision-and-Language pre-training (UVLP) to learn the cross-modal representation from non-parallel image and text datasets. We found two key factors that lead to good unsupervised V+L pre-training without parallel data: (i) joint image-and-text input (ii) overall image-text alignment (even for non-parallel data). Accordingly, we propose a novel unsupervised V+L pre-training curriculum for non-parallel texts and images. We first construct a weakly aligned image-text corpus via a retrieval-based approach, then apply a set of multi-granular alignment pre-training tasks, including region-to-tag, region-to-phrase, and image-to-sentence alignment, to bridge the gap between the two modalities. A comprehensive ablation study shows each granularity is helpful to learn a stronger pre-trained model. We adapt our pre-trained model to a set of V+L downstream tasks, including VQA, NLVR2, Visual Entailment, and RefCOCO+. Our model achieves the state-of-art performance in all these tasks under the unsupervised setting.
In this work, we pursue a unified paradigm for multimodal pretraining to break the scaffolds of complex task/modality-specific customization. We propose OFA, a Task-Agnostic and Modality-Agnostic framework that supports Task Comprehensiveness. OFA unifies a diverse set of cross-modal and unimodal tasks, including image generation, visual grounding, image captioning, image classification, language modeling, etc., in a simple sequence-to-sequence learning framework. OFA follows the instruction-based learning in both pretraining and finetuning stages, requiring no extra task-specific layers for downstream tasks. In comparison with the recent state-of-the-art vision & language models that rely on extremely large cross-modal datasets, OFA is pretrained on only 20M publicly available image-text pairs. Despite its simplicity and relatively small-scale training data, OFA achieves new SOTAs in a series of cross-modal tasks while attaining highly competitive performances on uni-modal tasks. Our further analysis indicates that OFA can also effectively transfer to unseen tasks and unseen domains. Our code and models are publicly available at https://github.com/OFA-Sys/OFA.
Contrastive language-image pretraining (CLIP) links vision and language modalities into a unified embedding space, yielding the tremendous potential for vision-language (VL) tasks. While early concurrent works have begun to study this potential on a subset of tasks, important questions remain: 1) What is the benefit of CLIP on unstudied VL tasks? 2) Does CLIP provide benefit in low-shot or domain-shifted scenarios? 3) Can CLIP improve existing approaches without impacting inference or pretraining complexity? In this work, we seek to answer these questions through two key contributions. First, we introduce an evaluation protocol that includes Visual Commonsense Reasoning (VCR), Visual Entailment (SNLI-VE), and Visual Question Answering (VQA), across a variety of data availability constraints and conditions of domain shift. Second, we propose an approach, named CLIP Targeted Distillation (CLIP-TD), to intelligently distill knowledge from CLIP into existing architectures using a dynamically weighted objective applied to adaptively selected tokens per instance. Experiments demonstrate that our proposed CLIP-TD leads to exceptional gains in the low-shot (up to 51.9%) and domain-shifted (up to 71.3%) conditions of VCR, while simultaneously improving performance under standard fully-supervised conditions (up to 2%), achieving state-of-art performance on VCR compared to other single models that are pretrained with image-text data only. On SNLI-VE, CLIP-TD produces significant gains in low-shot conditions (up to 6.6%) as well as fully supervised (up to 3%). On VQA, CLIP-TD provides improvement in low-shot (up to 9%), and in fully-supervised (up to 1.3%). Finally, CLIP-TD outperforms concurrent works utilizing CLIP for finetuning, as well as baseline naive distillation approaches. Code will be made available.
This paper describes our participant system for the multi-modal fact verification (Factify) challenge at AAAI 2022. Despite the recent advance in text based verification techniques and large pre-trained multimodal models cross vision and language, very limited work has been done in applying multimodal techniques to automate fact checking process, particularly considering the increasing prevalence of claims and fake news about images and videos on social media. In our work, the challenge is treated as multimodal entailment task and framed as multi-class classification. Two baseline approaches are proposed and explored including an ensemble model (combining two uni-modal models) and a multi-modal attention network (modeling the interaction between image and text pair from claim and evidence document). We conduct several experiments investigating and benchmarking different SoTA pre-trained transformers and vision models in this work. Our best model is ranked first in leaderboard which obtains a weighted average F-measure of 0.77 on both validation and test set. Exploratory analysis of dataset is also carried out on the Factify data set and uncovers salient patterns and issues (e.g., word overlapping, visual entailment correlation, source bias) that motivates our hypothesis. Finally, we highlight challenges of the task and multimodal dataset for future research.