We first demonstrate the capability of our system to learn a continuous latent texture space when trained on a dataset consisting of diverse textures (Section 1). In Section 5, we tabulate the network architectures of the convolutional neural networks used in our experiments. First and last square in each strip correspond to resynthesized exemplars. Note how our result is closer to the exemplar texture. See Figure 7 (second row) and Figure 8 (third row) for exemplar images.

Data hiding with deep neural networks (DNNs) has experienced impressive successes in recent years. A prevailing scheme is to train an autoencoder, consisting of an encoding network to embed (or transform) secret messages in (or into) a carrier, and a decoding network to extract the hidden messages. This scheme may suffer from several limitations regarding practicability, security, and embedding capacity. In this work, we describe a different computational framework to hide images in deep probabilistic models. Specifically, we use a DNN to model the probability density of cover images, and hide a secret image in one particular location of the learned distribution.

We first demonstrate the capability of our system to learn a continuous latent texture space when trained on a dataset consisting of diverse textures (Section 1). In Section 5, we tabulate the network architectures of the convolutional neural networks used in our experiments. First and last square in each strip correspond to resynthesized exemplars. Note how our result is closer to the exemplar texture. See Figure 7 (second row) and Figure 8 (third row) for exemplar images.

Data hiding with deep neural networks (DNNs) has experienced impressive successes in recent years. A prevailing scheme is to train an autoencoder, consisting of an encoding network to embed (or transform) secret messages in (or into) a carrier, and a decoding network to extract the hidden messages. This scheme may suffer from several limitations regarding practicability, security, and embedding capacity. In this work, we describe a different computational framework to hide images in deep probabilistic models. Specifically, we use a DNN to model the probability density of cover images, and hide a secret image in one particular location of the learned distribution.

Typical deep visual recognition models are capable of performing the one task they were trained on. In this paper, we tackle the extremely difficult problem of combining completely distinct models with different initializations, each solving a separate task, into one multi-task model without any additional training. Prior work in model merging permutes one model to the space of the other then adds them together. While this works for models trained on the same task, we find that this fails to account for the differences in models trained on disjoint tasks. Thus, we introduce "ZipIt!", a general method for merging two arbitrary models of the same architecture that incorporates two simple strategies. First, in order to account for features that aren't shared between models, we expand the model merging problem to additionally allow for merging features within each model by defining a general "zip" operation. Second, we add support for partially zipping the models up until a specified layer, naturally creating a multi-head model. We find that these two changes combined account for a staggering 20-60% improvement over prior work, making the merging of models trained on disjoint tasks feasible.

In recent years, multi-scale generative adversarial networks (GANs) have been proposed to build generalized image processing models based on single sample. Constraining on the sample size, multi-scale GANs have much difficulty converging to the global optimum, which ultimately leads to limitations in their capabilities. In this paper, we pioneered the introduction of PAC-Bayes generalized bound theory into the training analysis of specific models under different adversarial training methods, which can obtain a non-vacuous upper bound on the generalization error for the specified multi-scale GAN structure. Based on the drastic changes we found of the generalization error bound under different adversarial attacks and different training states, we proposed an adaptive training method which can greatly improve the image manipulation ability of multi-scale GANs. The final experimental results show that our adaptive training method in this paper has greatly contributed to the improvement of the quality of the images generated by multi-scale GANs on several image manipulation tasks. In particular, for the image super-resolution restoration task, the multi-scale GAN model trained by the proposed method achieves a 100% reduction in natural image quality evaluator (NIQE) and a 60% reduction in root mean squared error (RMSE), which is better than many models trained on large-scale datasets.

The various levels, quests and characters in modern video games play a major role in these games' engagement and entertainment values. One way to keep things fresh is procedural content generation (PCG), the algorithmic generation of game content using a random process that can produce an unpredictable range of possible gameplay spaces, freeing human game designers from the laborious task of manual content generation. Recent improvements in machine learning (ML) have spurred interest in applying such techniques to PCG, but research on level generation in 3D games remains limited. In the popular 3D Minecraft game, for example, humans still play a central role in content generation -- structures have to be placed manually in a fixed world because the Minecraft World Generator can't generate new structures on its own. To fill the gap between the Minecraft World Generator's PCG and manually created custom structures, a research team from Leibniz University Hannover recently introduced World-GAN, a 3D generative adversarial network (GAN) that can learn and generate structures directly in the Minecraft 3D voxel space.

The extensive augmentations significantly increase the in-sample distribution for the upsampling network enabling the upscaling of highly variable inputs. A compact latent space is jointly learned allowing for controlled image synthesis. Differently from Single Image GAN, our approach does not require GAN training and takes place in an end-to-end fashion allowing fast and stable training. We experimentally evaluate our method and show that it obtains compelling novel animations of single-image, as well as, state-of-the-art performance on conditional generation tasks e.g.