State-of-the-art methods in the unpaired image-to-image translation are capable of learning a mapping from a source domain to a target domain with unpaired image data. Though the existing methods have achieved promising results, they still produce unsatisfied artifacts, being able to convert low-level information while limited in transforming high-level semantics of input images. One possible reason is that generators do not have the ability to perceive the most discriminative semantic parts between the source and target domains, thus making the generated images low quality. In this paper, we propose a new Attention-Guided Generative Adversarial Networks (AttentionGAN) for the unpaired image-to-image translation task. AttentionGAN can identify the most discriminative semantic objects and minimize changes of unwanted parts for semantic manipulation problems without using extra data and models. The attention-guided generators in AttentionGAN are able to produce attention masks via a built-in attention mechanism, and then fuse the generation output with the attention masks to obtain high-quality target images. Accordingly, we also design a novel attention-guided discriminator which only considers attended regions. Extensive experiments are conducted on several generative tasks, demonstrating that the proposed model is effective to generate sharper and more realistic images compared with existing competitive models. The source code for the proposed AttentionGAN is available at https://github.com/Ha0Tang/AttentionGAN.
It is well known that deep generative models have a rich latent space, and that it is possible to smoothly manipulate their outputs by traversing this latent space. Recently, architectures have emerged that allow for more complex manipulations, such as making an image look as though it were from a different class, or painted in a certain style. These methods typically require large amounts of training in order to learn a single class of manipulations. We present Transflow Learning, a method for transforming a pre-trained generative model so that its outputs more closely resemble data that we provide afterwards. In contrast to previous methods, Transflow Learning does not require any training at all, and instead warps the probability distribution from which we sample latent vectors using Bayesian inference. Transflow Learning can be used to solve a wide variety of tasks, such as neural style transfer and few-shot classification.
Existing approaches to amortized inference in probabilistic programs with unbounded loops can produce estimators with infinite variance. An instance of this is importance sampling inference in programs that explicitly include rejection sampling as part of the user-programmed generative procedure. In this paper we develop a new and efficient amortized importance sampling estimator. We prove finite variance of our estimator and empirically demonstrate our method's correctness and efficiency compared to existing alternatives on generative programs containing rejection sampling loops and discuss how to implement our method in a generic probabilistic programming framework.
We present Siam R-CNN, a Siamese re-detection architecture which unleashes the full power of two-stage object detection approaches for visual object tracking. We combine this with a novel tracklet-based dynamic programming algorithm, which takes advantage of re-detections of both the first-frame template and previous-frame predictions, to model the full history of both the object to be tracked and potential distractor objects. This enables our approach to make better tracking decisions, as well as to re-detect tracked objects after long occlusion. Finally, we propose a novel hard example mining strategy to improve Siam R-CNN's robustness to similar looking objects. The proposed tracker achieves the current best performance on ten tracking benchmarks, with especially strong results for long-term tracking.
Learning generative models that span multiple data modalities, such as vision and language, is often motivated by the desire to learn more useful, generalisable representations that faithfully capture common underlying factors between the modalities. In this work, we characterise successful learning of such models as the fulfillment of four criteria: i) implicit latent decomposition into shared and private subspaces, ii) coherent joint generation over all modalities, iii) coherent cross-generation across individual modalities, and iv) improved model learning for individual modalities through multi-modal integration. Here, we propose a mixture-of-experts multimodal variational autoencoder (MMVAE) to learn generative models on different sets of modalities, including a challenging image-language dataset, and demonstrate its ability to satisfy all four criteria, both qualitatively and quantitatively.
Unsupervised video object segmentation has often been tackled by methods based on recurrent neural networks and optical flow. Despite their complexity, these kinds of approaches tend to favour short-term temporal dependencies and are thus prone to accumulating inaccuracies, which cause drift over time. Moreover, simple (static) image segmentation models, alone, can perform competitively against these methods, which further suggests that the way temporal dependencies are modelled should be reconsidered. Motivated by these observations, in this paper we explore simple yet effective strategies to model long-term temporal dependencies. Inspired by the non-local operators of [70], we introduce a technique to establish dense correspondences between pixel embeddings of a reference "anchor" frame and the current one. This allows the learning of pairwise dependencies at arbitrarily long distances without conditioning on intermediate frames. Without online supervision, our approach can suppress the background and precisely segment the foreground object even in challenging scenarios, while maintaining consistent performance over time. With a mean IoU of $81.7\%$, our method ranks first on the DAVIS-2016 leaderboard of unsupervised methods, while still being competitive against state-of-the-art online semi-supervised approaches. We further evaluate our method on the FBMS dataset and the ViSal video saliency dataset, showing results competitive with the state of the art.
Quantizing large Neural Networks (NN) while maintaining the performance is highly desirable for resource-limited devices due to reduced memory and time complexity. NN quantization is usually formulated as a constrained optimization problem and optimized via a modified version of gradient descent. In this work, by interpreting the continuous parameters (unconstrained) as the dual of the quantized ones, we introduce a Mirror Descent (MD) framework (Bubeck (2015)) for NN quantization. Specifically, we provide conditions on the projections (i.e., mapping from continuous to quantized ones) which would enable us to derive valid mirror maps and in turn the respective MD updates. Furthermore, we discuss a numerically stable implementation of MD by storing an additional set of auxiliary dual variables (continuous). This update is strikingly analogous to the popular Straight Through Estimator (STE) based method which is typically viewed as a "trick" to avoid vanishing gradients issue but here we show that it is an implementation method for MD for certain projections. Our experiments on standard classification datasets (CIFAR-10/100, TinyImageNet) with convolutional and residual architectures show that our MD variants obtain fully-quantized networks with accuracies very close to the floating-point networks.