Understanding the representational power of Restricted Boltzmann Machines (RBMs) with multiple layers is an ill-understood problem and is an area of active research. Motivated from the approach of \emph{Inherent Structure formalism} (Stillinger & Weber, 1982), extensively used in analysing Spin Glasses, we propose a novel measure called \emph{Inherent Structure Capacity} (ISC), which characterizes the representation capacity of a fixed architecture RBM by the expected number of modes of distributions emanating from the RBM with parameters drawn from a prior distribution. Though ISC is intractable, we show that for a single layer RBM architecture ISC approaches a finite constant as number of hidden units are increased and to further improve the ISC, one needs to add a second layer. Furthermore, we introduce \emph{Lean} RBMs, which are multi-layer RBMs where each layer can have at-most $O(n)$ units with the number of visible units being n. We show that for every single layer RBM with $\Omega(n^{2+r}), r \ge 0$, hidden units there exists a two-layered \emph{lean} RBM with $\Theta(n^2)$ parameters with the same ISC, establishing that 2 layer RBMs can achieve the same representational power as single-layer RBMs but using far fewer number of parameters. To the best of our knowledge, this is the first result which quantitatively establishes the need for layering.

We show empirically that the optimal strategy of parameter averaging in a minmax convex-concave game setting is also strikingly effective in the non convex-concave GAN setting, specifically alleviating the convergence issues associated with cycling behavior observed in GANs. We show that averaging over generator parameters outside of the trainig loop consistently improves inception and FID scores on different architectures and for different GAN objectives. We provide comprehensive experimental results across a range of datasets, bilinear games, mixture of Gaussians, CIFAR-10, STL-10, CelebA and ImageNet, to demonstrate its effectiveness. We achieve state-of-the-art results on CIFAR-10 and produce clean CelebA face images, demonstrating that averaging is one of the most effective techniques for training highly performant GANs.

Stochastic gradient Markov chain Monte Carlo (SG-MCMC) has become increasingly popular for simulating posterior samples in large-scale Bayesian modeling. However, existing SG-MCMC schemes are not tailored to any specific probabilistic model, even a simple modification of the underlying dynamical system requires significant physical intuition. This paper presents the first meta-learning algorithm that allows automated design for the underlying continuous dynamics of an SG-MCMC sampler. The learned sampler generalizes Hamiltonian dynamics with state-dependent drift and diffusion, enabling fast traversal and efficient exploration of neural network energy landscapes. Experiments validate the proposed approach on both Bayesian fully connected neural network and Bayesian recurrent neural network tasks, showing that the learned sampler out-performs generic, hand-designed SG-MCMC algorithms, and generalizes to different datasets and larger architectures.

We propose a method for maximizing a partial area under a receiver operating characteristic (ROC) curve (pAUC) for binary classification tasks. In binary classification tasks, accuracy is the most commonly used as a measure of classifier performance. In some applications such as anomaly detection and diagnostic testing, accuracy is not an appropriate measure since prior probabilties are often greatly biased. Although in such cases the pAUC has been utilized as a performance measure, few methods have been proposed for directly maximizing the pAUC. This optimization is achieved by using a scoring function. The conventional approach utilizes a linear function as the scoring function. In contrast we newly introduce nonlinear scoring functions for this purpose. Specifically, we present two types of nonlinear scoring functions based on generative models and deep neural networks. We show experimentally that nonlinear scoring fucntions improve the conventional methods through the application of a binary classification of real and bogus objects obtained with the Hyper Suprime-Cam on the Subaru telescope.

Stream constraint programming is a recent addition to the family of constraint programming frameworks, where variable domains are sets of infinite streams over finite alphabets. Previous works showed promising results for its applicability to real-world planning and control problems. In this paper, motivated by the modelling of planning applications, we improve the expressiveness of the framework by introducing 1) the "until" constraint, a new construct that is adapted from Linear Temporal Logic and 2) the @ operator on streams, a syntactic sugar for which we provide a more efficient solving algorithm over simple desugaring. For both constructs, we propose corresponding novel solving algorithms and prove their correctness.

Meta-learning is a powerful tool that builds on multi-task learning to learn how to quickly adapt a model to new tasks. In the context of reinforcement learning, meta-learning algorithms can acquire reinforcement learning procedures to solve new problems more efficiently by meta-learning prior tasks. The performance of meta-learning algorithms critically depends on the tasks available for meta-training: in the same way that supervised learning algorithms generalize best to test points drawn from the same distribution as the training points, meta-learning methods generalize best to tasks from the same distribution as the meta-training tasks. In effect, meta-reinforcement learning offloads the design burden from algorithm design to task design. If we can automate the process of task design as well, we can devise a meta-learning algorithm that is truly automated. In this work, we take a step in this direction, proposing a family of unsupervised meta-learning algorithms for reinforcement learning. We describe a general recipe for unsupervised meta-reinforcement learning, and describe an effective instantiation of this approach based on a recently proposed unsupervised exploration technique and model-agnostic meta-learning. We also discuss practical and conceptual considerations for developing unsupervised meta-learning methods. Our experimental results demonstrate that unsupervised meta-reinforcement learning effectively acquires accelerated reinforcement learning procedures without the need for manual task design, significantly exceeds the performance of learning from scratch, and even matches performance of meta-learning methods that use hand-specified task distributions.

Deep learning has enabled traditional reinforcement learning methods to deal with high-dimensional problems. However, one of the disadvantages of deep reinforcement learning methods is the limited exploration capacity of learning agents. In this paper, we introduce an approach that integrates human strategies to increase the exploration capacity of multiple deep reinforcement learning agents. We also report the development of our own multi-agent environment called Multiple Tank Defence to simulate the proposed approach. The results show the significant performance improvement of multiple agents that have learned cooperatively with human strategies. This implies that there is a critical need for human intellect teamed with machines to solve complex problems. In addition, the success of this simulation indicates that our developed multi-agent environment can be used as a testbed platform to develop and validate other multi-agent control algorithms. Details of the environment implementation can be referred to http://www.deakin.edu.au/~thanhthi/madrl_human.htm

And hey, more CG footage of Beyond Good & Evil 2. With Ubisoft's E3 press conferences, you always know exactly what you're going to get, and yet it's also impressive (to me at least) to watch the machine at work, to watch Ubisoft trot out such a full lineup of experiences every year, without fail. We've rounded up all the trailers from Ubisoft's E3 2018 press conference below, and it's exactly what you'd expect. And damn, it might not be inspiring but on some level I can respect the craft. After a fever dream of an introduction for Just Dance 2019 (featuring a dancing panda), Ubisoft finally moved into something we could care about: Beyond Good & Evil 2. First up, just a stunning CG trailer. Like, good enough that I wish Ubisoft would make a Beyond Good & Evil film. It has some fantastic shots of the world itself--including a faux X-Wing flying through empty space, a ship AI being channeled through a jewel-encrusted skull, and the return of the original game's protagonist Jade.

Yes. First, artificial neural networks as whole were inspired--as their name suggests--by the emerging biology of neurons being developed in the mid-20th century. Artificial neurons were designed to mimic the basic characteristics of how neurons take in and transform information. Second, the main features and computations done by convolutional networks were directly inspired by some of the early findings about the visual system. In 1962 Hubel and Wiesel discovered that neurons in primary visual cortex respond to specific, simple features in the visual environment (particularly, oriented edges). Furthermore, they noticed two different kinds of cells: simple cells--which responded most strongly to their preferred orientation only at a very particular spatial location--and complex cells--which had more spatial invariance in their response.

From the spears hurled by Romans to the missiles launched by fighter pilots, the weapons humans use to kill each other have always been subject to improvement. Militaries seek to make each one ever-more lethal and, in doing so, better protect the soldier who wields it. But in the next evolution of combat, the U.S. Army is heading down a path that may lead humans off the battlefield entirely. Over the next few years, the Pentagon is poised to spend almost $1 billion for a range of robots designed to complement combat troops. Beyond scouting and explosives disposal, these new machines will sniff out hazardous chemicals or other agents, perform complex reconnaissance and even carry a soldier's gear.