High-dimensional data can be converted to low-dimensional codes by training a multilayer neural network with a small central layer to reconstruct high-dimensional input vectors. Gradient descent can be used for fine-tuning the weights in such "autoencoder" networks, but this works well only if the initial weights are close to a good solution. We describe an effective way of initializing the weights that allows deep autoencoder networks to learn low-dimensional codes that work much better than principal components analysis as a tool to reduce the dimensionality of data.

/pdf/reducing-the-dimensionality-of-data-with-neural-networks-51w34pmjfl.pdf

Reducing the Dimensionality of Data with Neural Networks

/pdf/scalable-molecular-dynamics-with-namd-3j7xyc70g4.pdf

Scalable Molecular Dynamics with NAMD

We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems. These networks not only learn the mapping from input image to output image, but also learn a loss function to train this mapping. This makes it possible to apply the same generic approach to problems that traditionally would require very different loss formulations. We demonstrate that this approach is effective at synthesizing photos from label maps, reconstructing objects from edge maps, and colorizing images, among other tasks. Moreover, since the release of the pix2pix software associated with this paper, hundreds of twitter users have posted their own artistic experiments using our system. As a community, we no longer hand-engineer our mapping functions, and this work suggests we can achieve reasonable results without handengineering our loss functions either.

/pdf/image-to-image-translation-with-conditional-adversarial-5058t24et2.pdf

Image-to-Image Translation with Conditional Adversarial Networks

We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems. These networks not only learn the mapping from input image to output image, but also learn a loss function to train this mapping. This makes it possible to apply the same generic approach to problems that traditionally would require very different loss formulations. We demonstrate that this approach is effective at synthesizing photos from label maps, reconstructing objects from edge maps, and colorizing images, among other tasks. Indeed, since the release of the pix2pix software associated with this paper, a large number of internet users (many of them artists) have posted their own experiments with our system, further demonstrating its wide applicability and ease of adoption without the need for parameter tweaking. As a community, we no longer hand-engineer our mapping functions, and this work suggests we can achieve reasonable results without hand-engineering our loss functions either.

Can machine learning deliver AI? Theoretical results, inspiration from the brain and cognition, as well as machine learning experiments suggest that in order to learn the kind of complicated functions that can represent high-level abstractions (e.g. in vision, language, and other AI-level tasks), one would need deep architectures. Deep architectures are composed of multiple levels of non-linear operations, such as in neural nets with many hidden layers, graphical models with many levels of latent variables, or in complicated propositional formulae re-using many sub-formulae. Each level of the architecture represents features at a different level of abstraction, defined as a composition of lower-level features. Searching the parameter space of deep architectures is a difficult task, but new algorithms have been discovered and a new sub-area has emerged in the machine learning community since 2006, following these discoveries. Learning algorithms such as those for Deep Belief Networks and other related unsupervised learning algorithms have recently been proposed to train deep architectures, yielding exciting results and beating the state-of-the-art in certain areas. Learning Deep Architectures for AI discusses the motivations for and principles of learning algorithms for deep architectures. By analyzing and comparing recent results with different learning algorithms for deep architectures, explanations for their success are proposed and discussed, highlighting challenges and suggesting avenues for future explorations in this area.

/pdf/learning-deep-architectures-for-ai-2358znl654.pdf

Learning Deep Architectures for AI

The class of all natural images is an infinitely small fraction of all possible images. The structure of natural images can be statistically modeled, revealing striking regularities. The human visual system appears to be optimized to view natural images, as opposed to any possible image, and therefore expects to interpret images which conform to these statistics. Research has shown that images that do not statistically behave as natural images are harder for the human visual system to interpret. This paper reviews the statistics of natural images and the implications for computer graphics in general are assessed. We argue that these statistics are important for graphics applications and finally, we provide a direct application of these findings to random subdivision terrain modeling.

/pdf/natural-image-statistics-for-computer-graphics-1viqaxu5f2.pdf

Natural image statistics for computer graphics

We present three improvements to Perlin’s gradient noise algorithm. First, a small change in the permutation hash function combined with separate pseudorandom tables yields significantly better axial decorrelation. Second, a modification to the reconstruction kernel approximating a global higher-order differencing operator produces better bandlimitation. Third, the quality of 2D surfaces using solid 3D noise is improved by reconstructing the stencil projected onto a surface normal. These three techniques are mutually orthogonal, generalize to higher dimensions, and are applicable to nearly any gradient noise, including simplex noise. Combining them yields a desirable Fourier spectrum for graphics applications.

https://www.sci.utah.edu/publications/SCITechReports/UUSCI-2008-001.pdf

Better Gradient Noise

A system, method, and computer program product are provided for computing indirect lighting in a cloud network. In operation, one or more scenes for rendering are identified. Further, indirect lighting associated with the one or more scenes is identified. Additionally, computation associated with the indirect lighting is performed in a cloud network utilizing at least one of a voxel-based algorithm, a photon-based algorithm, or an irradiance-map-based algorithm.

System, method, and computer program product for computing indirect lighting in a cloud network

Distribution ray tracing uses multiple samples per pixel to produce antialiased images that include soft shadows, glossy reflection, motion blur, and depth-of-field. The two main potential barriers to making distribution ray tracing interactive are that many rays might be required, and that those rays are not coherent enough to derive efficiency from tracing them in packets. A new interleaved sampling approach based on the Sudoku puzzle is used to minimize the number of rays per pixel. An empirical demonstration is used to show that there is still enough coherence in the rays to allow for a per-ray cost near that of a traditional ray tracer. In addition, a demonstration is provided that participating media can also be handled interactively in a distribution ray tracer.

/pdf/interactive-distribution-ray-tracing-ls76tns3en.pdf

Interactive Distribution Ray Tracing

Ray tracing a three-dimensional scene made up of geometric primitives that are spatially partitioned into a hierarchical data structure. One example embodiment is a method for ray tracing a three-dimensional scene made up of geometric primitives that are spatially partitioned into a hierarchical data structure. In this example embodiment, the hierarchical data structure includes at least a parent node and a corresponding plurality of child nodes. The method includes a first act of determining that a first active ray in the packet hits the parent node and a second act of descending to each of the plurality of child nodes.

Peter Shirley

Papers

Natural image statistics for computer graphics

Better Gradient Noise

System, method, and computer program product for computing indirect lighting in a cloud network

Interactive Distribution Ray Tracing

Ray tracing a three-dimensional scene using a hierarchical data structure