We present VILLA, the first known effort on large-scale adversarial training for vision-and-language (V+L) representation learning. VILLA consists of two training stages: (i) task-agnostic adversarial pre-training; followed by (ii) task-specific adversarial finetuning. Instead of adding adversarial perturbations on image pixels and textual tokens, we propose to perform adversarial training in the embedding space of each modality. To enable large-scale training, we adopt the "free" adversarial training strategy, and combine it with KL-divergence-based regularization to promote higher invariance in the embedding space. We apply VILLA to current best-performing V+L models, and achieve new state of the art on a wide range of tasks, including Visual Question Answering, Visual Commonsense Reasoning, Image-Text Retrieval, Referring Expression Comprehension, Visual Entailment, and NLVR2.
We present a new method to learn video representations from large-scale unlabeled video data. Ideally, this representation will be generic and transferable, directly usable for new tasks such as action recognition and zero or few-shot learning. We formulate unsupervised representation learning as a multi-modal, multi-task learning problem, where the representations are shared across different modalities via distillation. Further, we introduce the concept of loss function evolution by using an evolutionary search algorithm to automatically find optimal combination of loss functions capturing many (self-supervised) tasks and modalities. Thirdly, we propose an unsupervised representation evaluation metric using distribution matching to a large unlabeled dataset as a prior constraint, based on Zipf's law. This unsupervised constraint, which is not guided by any labeling, produces similar results to weakly-supervised, task-specific ones. The proposed unsupervised representation learning results in a single RGB network and outperforms previous methods. Notably, it is also more effective than several label-based methods (e.g., ImageNet), with the exception of large, fully labeled video datasets.
This paper presents SimCLR: a simple framework for contrastive learning of visual representations. We simplify recently proposed contrastive self-supervised learning algorithms without requiring specialized architectures or a memory bank. In order to understand what enables the contrastive prediction tasks to learn useful representations, we systematically study the major components of our framework. We show that (1) composition of data augmentations plays a critical role in defining effective predictive tasks, (2) introducing a learnable nonlinear transformation between the representation and the contrastive loss substantially improves the quality of the learned representations, and (3) contrastive learning benefits from larger batch sizes and more training steps compared to supervised learning. By combining these findings, we are able to considerably outperform previous methods for self-supervised and semi-supervised learning on ImageNet. A linear classifier trained on self-supervised representations learned by SimCLR achieves 76.5% top-1 accuracy, which is a 7% relative improvement over previous state-of-the-art, matching the performance of a supervised ResNet-50. When fine-tuned on only 1% of the labels, we achieve 85.8% top-5 accuracy, outperforming AlexNet with 100X fewer labels.
This paper shows that pretraining multilingual language models at scale leads to significant performance gains for a wide range of cross-lingual transfer tasks. We train a Transformer-based masked language model on one hundred languages, using more than two terabytes of filtered CommonCrawl data. Our model, dubbed XLM-R, significantly outperforms multilingual BERT (mBERT) on a variety of cross-lingual benchmarks, including +13.8% average accuracy on XNLI, +12.3% average F1 score on MLQA, and +2.1% average F1 score on NER. XLM-R performs particularly well on low-resource languages, improving 11.8% in XNLI accuracy for Swahili and 9.2% for Urdu over the previous XLM model. We also present a detailed empirical evaluation of the key factors that are required to achieve these gains, including the trade-offs between (1) positive transfer and capacity dilution and (2) the performance of high and low resource languages at scale. Finally, we show, for the first time, the possibility of multilingual modeling without sacrificing per-language performance; XLM-Ris very competitive with strong monolingual models on the GLUE and XNLI benchmarks. We will make XLM-R code, data, and models publicly available.
Joint image-text embedding is the bedrock for most Vision-and-Language (V+L) tasks, where multimodality inputs are jointly processed for visual and textual understanding. In this paper, we introduce UNITER, a UNiversal Image-TExt Representation, learned through large-scale pre-training over four image-text datasets (COCO, Visual Genome, Conceptual Captions, and SBU Captions), which can power heterogeneous downstream V+L tasks with joint multimodal embeddings. We design three pre-training tasks: Masked Language Modeling (MLM), Image-Text Matching (ITM), and Masked Region Modeling (MRM, with three variants). Different from concurrent work on multimodal pre-training that apply joint random masking to both modalities, we use conditioned masking on pre-training tasks (i.e., masked language/region modeling is conditioned on full observation of image/text). Comprehensive analysis shows that conditioned masking yields better performance than unconditioned masking. We also conduct a thorough ablation study to find an optimal setting for the combination of pre-training tasks. Extensive experiments show that UNITER achieves new state of the art across six V+L tasks (over nine datasets), including Visual Question Answering, Image-Text Retrieval, Referring Expression Comprehension, Visual Commonsense Reasoning, Visual Entailment, and NLVR2.
For many computer vision applications such as image captioning, visual question answering, and person search, learning discriminative feature representations at both image and text level is an essential yet challenging problem. Its challenges originate from the large word variance in the text domain as well as the difficulty of accurately measuring the distance between the features of the two modalities. Most prior work focuses on the latter challenge, by introducing loss functions that help the network learn better feature representations but fail to account for the complexity of the textual input. With that in mind, we introduce TIMAM: a Text-Image Modality Adversarial Matching approach that learns modality-invariant feature representations using adversarial and cross-modal matching objectives. In addition, we demonstrate that BERT, a publicly-available language model that extracts word embeddings, can successfully be applied in the text-to-image matching domain. The proposed approach achieves state-of-the-art cross-modal matching performance on four widely-used publicly-available datasets resulting in absolute improvements ranging from 2% to 5% in terms of rank-1 accuracy.
In existing visual representation learning tasks, deep convolutional neural networks (CNNs) are often trained on images annotated with single tags, such as ImageNet. However, a single tag cannot describe all important contents of one image, and some useful visual information may be wasted during training. In this work, we propose to train CNNs from images annotated with multiple tags, to enhance the quality of visual representation of the trained CNN model. To this end, we build a large-scale multi-label image database with 18M images and 11K categories, dubbed Tencent ML-Images. We efficiently train the ResNet-101 model with multi-label outputs on Tencent ML-Images, taking 90 hours for 60 epochs, based on a large-scale distributed deep learning framework,i.e.,TFplus. The good quality of the visual representation of the Tencent ML-Images checkpoint is verified through three transfer learning tasks, including single-label image classification on ImageNet and Caltech-256, object detection on PASCAL VOC 2007, and semantic segmentation on PASCAL VOC 2012. The Tencent ML-Images database, the checkpoints of ResNet-101, and all the training codehave been released at https://github.com/Tencent/tencent-ml-images. It is expected to promote other vision tasks in the research and industry community.
Despite recent progress in generative image modeling, successfully generating high-resolution, diverse samples from complex datasets such as ImageNet remains an elusive goal. To this end, we train Generative Adversarial Networks at the largest scale yet attempted, and study the instabilities specific to such scale. We find that applying orthogonal regularization to the generator renders it amenable to a simple "truncation trick", allowing fine control over the trade-off between sample fidelity and variety by truncating the latent space. Our modifications lead to models which set the new state of the art in class-conditional image synthesis. When trained on ImageNet at 128x128 resolution, our models (BigGANs) achieve an Inception Score (IS) of 166.3 and Frechet Inception Distance (FID) of 9.6, improving over the previous best IS of 52.52 and FID of 18.65.
In this paper, we propose the Self-Attention Generative Adversarial Network (SAGAN) which allows attention-driven, long-range dependency modeling for image generation tasks. Traditional convolutional GANs generate high-resolution details as a function of only spatially local points in lower-resolution feature maps. In SAGAN, details can be generated using cues from all feature locations. Moreover, the discriminator can check that highly detailed features in distant portions of the image are consistent with each other. Furthermore, recent work has shown that generator conditioning affects GAN performance. Leveraging this insight, we apply spectral normalization to the GAN generator and find that this improves training dynamics. The proposed SAGAN achieves the state-of-the-art results, boosting the best published Inception score from 36.8 to 52.52 and reducing Frechet Inception distance from 27.62 to 18.65 on the challenging ImageNet dataset. Visualization of the attention layers shows that the generator leverages neighborhoods that correspond to object shapes rather than local regions of fixed shape.
We propose a flipped-Adversarial AutoEncoder (FAAE) that simultaneously trains a generative model G that maps an arbitrary latent code distribution to a data distribution and an encoder E that embodies an "inverse mapping" that encodes a data sample into a latent code vector. Unlike previous hybrid approaches that leverage adversarial training criterion in constructing autoencoders, FAAE minimizes re-encoding errors in the latent space and exploits adversarial criterion in the data space. Experimental evaluations demonstrate that the proposed framework produces sharper reconstructed images while at the same time enabling inference that captures rich semantic representation of data.