In this paper, we study a new task that allows users to edit an input image using language instructions. In this image generation task, the inputs are a reference image and a text instruction that describes desired modifications to the input image. We propose a GAN-based method to tackle this problem. The key idea is to treat language as neural operators to locally modify the image feature. To this end, our model decomposes the generation process into finding where (spatial region) and how (text operators) to apply modifications. We show that the proposed model performs favorably against recent baselines on three datasets.
Image generation from scene description is a cornerstone technique for the controlled generation, which is beneficial to applications such as content creation and image editing. In this work, we aim to synthesize images from scene description with retrieved patches as reference. We propose a differentiable retrieval module. With the differentiable retrieval module, we can (1) make the entire pipeline end-to-end trainable, enabling the learning of better feature embedding for retrieval; (2) encourage the selection of mutually compatible patches with additional objective functions. We conduct extensive quantitative and qualitative experiments to demonstrate that the proposed method can generate realistic and diverse images, where the retrieved patches are reasonable and mutually compatible.
In this paper, we propose a new adversarial augmentation method for Neural Machine Translation (NMT). The main idea is to minimize the vicinal risk over virtual sentences sampled from two vicinity distributions, of which the crucial one is a novel vicinity distribution for adversarial sentences that describes a smooth interpolated embedding space centered around observed training sentence pairs. We then discuss our approach, AdvAug, to train NMT models using the embeddings of virtual sentences in sequence-to-sequence learning. Experiments on Chinese-English, English-French, and English-German translation benchmarks show that AdvAug achieves significant improvements over the Transformer (up to 4.9 BLEU points), and substantially outperforms other data augmentation techniques (e.g. back-translation) without using extra corpora.
This paper focuses on the problem of predicting future trajectories of people in unseen scenarios and camera views. We propose a method to efficiently utilize multi-view 3D simulation data for training. Our approach finds the hardest camera view to mix up with adversarial data from the original camera view in training, thus enabling the model to learn robust representations that can generalize to unseen camera views. We refer to our method as SimAug. We show that SimAug achieves best results on three out-of-domain real-world benchmarks, as well as getting state-of-the-art in the Stanford Drone and the VIRAT/ActEV dataset with in-domain training data. We will release our models and code.
Image extrapolation aims at expanding the narrow field of view of a given image patch. Existing models mainly deal with natural scene images of homogeneous regions and have no control of the content generation process. In this work, we study conditional image extrapolation to synthesize new images guided by the input structured text. The text is represented as a graph to specify the objects and their spatial relation to the unknown regions of the image. Inspired by drawing techniques, we propose a progressive generative model of three stages, i.e., generating a coarse bounding-boxes layout, refining it to a finer segmentation layout, and mapping the layout to a realistic output. Such a multi-stage design is shown to facilitate the training process and generate more controllable results. We validate the effectiveness of the proposed method on the face and human clothing dataset in terms of visual results, quantitative evaluations and flexible controls.
Graphic design is essential for visual communication with layouts being fundamental to composing attractive designs. Layout generation differs from pixel-level image synthesis and is unique in terms of the requirement of mutual relations among the desired components. We propose a method for design layout generation that can satisfy user-specified constraints. The proposed neural design network (NDN) consists of three modules. The first module predicts a graph with complete relations from a graph with user-specified relations. The second module generates a layout from the predicted graph. Finally, the third module fine-tunes the predicted layout. Quantitative and qualitative experiments demonstrate that the generated layouts are visually similar to real design layouts. We also construct real designs based on predicted layouts for a better understanding of the visual quality. Finally, we demonstrate a practical application on layout recommendation.
This paper studies the problem of predicting the distribution over multiple possible future paths of people as they move through various visual scenes. We make two main contributions. The first contribution is a new dataset, created in a realistic 3D simulator, which is based on real world trajectory data, and then extrapolated by human annotators to achieve different latent goals. This provides the first benchmark for quantitative evaluation of the models to predict multi-future trajectories. The second contribution is a new model to generate multiple plausible future trajectories, which contains novel designs of using multi-scale location encodings and convolutional RNNs over graphs. We refer to our model as Multiverse. We show that our model achieves the best results on our dataset, as well as on the real-world VIRAT/ActEV dataset (which just contains one possible future). We will release our data, models and code.
Performing controlled experiments on noisy data is essential in thoroughly understanding deep learning across a spectrum of noise levels. Due to the lack of suitable datasets, previous research have only examined deep learning on controlled synthetic noise, and real-world noise has never been systematically studied in a controlled setting. To this end, this paper establishes a benchmark of real-world noisy labels at 10 controlled noise levels. As real-world noise possesses unique properties, to understand the difference, we conduct a large-scale study across a variety of noise levels and types, architectures, methods, and training settings. Our study shows that: (1) Deep Neural Networks (DNNs) generalize much better on real-world noise. (2) DNNs may not learn patterns first on real-world noisy data. (3) When networks are fine-tuned, ImageNet architectures generalize well on noisy data. (4) Real-world noise appears to be less harmful, yet it is more difficult for robust DNN methods to improve. (5) Robust learning methods that work well on synthetic noise may not work as well on real-world noise, and vice versa. We hope our benchmark, as well as our findings, will facilitate deep learning research on noisy data.