The detection of gravitational waves with LIGO and Virgo requires a detailed understanding of the response of these instruments in the presence of environmental and instrumental noise. Of particular interest is the study of anomalous non-Gaussian noise transients known as glitches, since their high occurrence rate in LIGO/Virgo data can obscure or even mimic true gravitational wave signals. Therefore, successfully identifying and excising glitches is of utmost importance to detect and characterize gravitational waves. In this article, we present the first application of Deep Learning combined with Transfer Learning for glitch classification, using real data from LIGO's first discovery campaign labeled by Gravity Spy, showing that knowledge from pre-trained models for real-world object recognition can be transferred for classifying spectrograms of glitches. We demonstrate that this method enables the optimal use of very deep convolutional neural networks for glitch classification given small unbalanced training datasets, significantly reduces the training time, and achieves state-of-the-art accuracy above 98.8%. Once trained via transfer learning, we show that the networks can be truncated and used as feature extractors for unsupervised clustering to automatically group together new classes of glitches and anomalies. This novel capability is of critical importance to identify and remove new types of glitches which will occur as the LIGO/Virgo detectors gradually attain design sensitivity.

In this paper, we tackle the real-world problem of predicting Yelp star-review rating based on business features (such as images, descriptions), user features (average previous ratings), and, of particular interest, network properties (which businesses has a user rated before). We compare multiple models on different sets of features -- from simple linear regression on network features only to deep learning models on network and item features. In recent years, breakthroughs in deep learning have led to increased accuracy in common supervised learning tasks, such as image classification, captioning, and language understanding. However, the idea of combining deep learning with network feature and structure appears to be novel. While the problem of predicting future interactions in a network has been studied at length, these approaches have often ignored either node-specific data or global structure. We demonstrate that taking a mixed approach combining both node-level features and network information can effectively be used to predict Yelp-review star ratings. We evaluate on the Yelp dataset by splitting our data along the time dimension (as would naturally occur in the real-world) and comparing our model against others which do no take advantage of the network structure and/or deep learning.

Directed latent variable models that formulate the joint distribution as $p(x,z) = p(z) p(x \mid z)$ have the advantage of fast and exact sampling. However, these models have the weakness of needing to specify $p(z)$, often with a simple fixed prior that limits the expressiveness of the model. Undirected latent variable models discard the requirement that $p(z)$ be specified with a prior, yet sampling from them generally requires an iterative procedure such as blocked Gibbs-sampling that may require many steps to draw samples from the joint distribution $p(x, z)$. We propose a novel approach to learning the joint distribution between the data and a latent code which uses an adversarially learned iterative procedure to gradually refine the joint distribution, $p(x, z)$, to better match with the data distribution on each step. GibbsNet is the best of both worlds both in theory and in practice. Achieving the speed and simplicity of a directed latent variable model, it is guaranteed (assuming the adversarial game reaches the virtual training criteria global minimum) to produce samples from $p(x, z)$ with only a few sampling iterations. Achieving the expressiveness and flexibility of an undirected latent variable model, GibbsNet does away with the need for an explicit $p(z)$ and has the ability to do attribute prediction, class-conditional generation, and joint image-attribute modeling in a single model which is not trained for any of these specific tasks. We show empirically that GibbsNet is able to learn a more complex $p(z)$ and show that this leads to improved inpainting and iterative refinement of $p(x, z)$ for dozens of steps and stable generation without collapse for thousands of steps, despite being trained on only a few steps.

Generative adversarial networks (GANs) are innovative techniques for learning generative models of complex data distributions from samples. Despite remarkable recent improvements in generating realistic images, one of their major shortcomings is the fact that in practice, they tend to produce samples with little diversity, even when trained on diverse datasets. This phenomenon, known as mode collapse, has been the main focus of several recent advances in GANs. Yet there is little understanding of why mode collapse happens and why existing approaches are able to mitigate mode collapse. We propose a principled approach to handling mode collapse, which we call packing. The main idea is to modify the discriminator to make decisions based on multiple samples from the same class, either real or artificially generated. We borrow analysis tools from binary hypothesis testing---in particular the seminal result of Blackwell [Bla53]---to prove a fundamental connection between packing and mode collapse. We show that packing naturally penalizes generators with mode collapse, thereby favoring generator distributions with less mode collapse during the training process. Numerical experiments on benchmark datasets suggests that packing provides significant improvements in practice as well.

The paper approaches the problem of image-to-text with attention-based encoder-decoder networks that are trained to handle sequences of characters rather than words. We experiment on lines of text from a popular handwriting database with different attention mechanisms for the decoder. The model trained with softmax attention achieves the lowest test error, outperforming several other RNN-based models. Our results show that softmax attention is able to learn a linear alignment whereas the alignment generated by sigmoid attention is linear but much less precise.

In this paper we propose new algorithm to reduce autocorrelation in Markov chain Monte-Carlo algorithms for euclidean field theories on the lattice. Our proposing algorithm is the Hybrid Monte-Carlo algorithm (HMC) with restricted Boltzmann machine. We examine the validity of the algorithm by employing the phi-fourth theory in three dimension. We observe reduction of the autocorrelation both in symmetric and broken phase as well. Our proposing algorithm provides consistent central values of expectation values of the action density and one-point Green's function with ones from the original HMC in both the symmetric phase and broken phase within the statistical error. On the other hand, two-point Green's functions have slight difference between one calculated by the HMC and one by our proposing algorithm in the symmetric phase. Furthermore, near the criticality, the distribution of the one-point Green's function differs from the one from HMC. We discuss the origin of discrepancies and its improvement.

Many radiological studies can reveal the presence of several co-existing abnormalities, each one represented by a distinct visual pattern. In this article we address the problem of learning a distance metric for plain radiographs that captures a notion of "radiological similarity": two chest radiographs are considered to be similar if they share similar abnormalities. Deep convolutional neural networks (DCNs) are used to learn a low-dimensional embedding for the radiographs that is equipped with the desired metric. Two loss functions are proposed to deal with multi-labelled images and potentially noisy labels. We report on a large-scale study involving over 745,000 chest radiographs whose labels were automatically extracted from free-text radiological reports through a natural language processing system. Using 4,500 validated exams, we demonstrate that the methodology performs satisfactorily on clustering and image retrieval tasks. Remarkably, the learned metric separates normal exams from those having radiological abnormalities.

Attempts to train a comprehensive artificial intelligence capable of solving multiple tasks have been impeded by a chronic problem called catastrophic forgetting. Although simply replaying all previous data alleviates the problem, it requires large memory and even worse, often infeasible in real world applications where the access to past data is limited. Inspired by the generative nature of the hippocampus as a short-term memory system in primate brain, we propose the Deep Generative Replay, a novel framework with a cooperative dual model architecture consisting of a deep generative model ("generator") and a task solving model ("solver"). With only these two models, training data for previous tasks can easily be sampled and interleaved with those for a new task. We test our methods in several sequential learning settings involving image classification tasks.

What are neural networks, but more important, how are they trained in practice? How can data scientists design an optimal neural network when a single training run can take 2 weeks? In this Data Science Central webinar we will start from the foundation of what deep learning is then fast forward through what it takes to train a production quality neural network. You won't be able to train a network when this talk is over, but you'll understand enough basics to start smart conversations about our customers' practice of deep learning.

This is a graphics card created for the PC. VentureBeat's Blair Frank said "The new Titan V card will provide customers with a Nvidia Volta chip that they can plug into a desktop computer." Thursday marked its debut, positioned as "the world's most powerful GPU for the PC." CEO Jensen Huang did the introduction. The announcement took place at the annual AI gathering, the NIPS (Neural Information Processing Systems) conference. It can carry massive amounts of power and speed AI computation.