Music generation research has grown in popularity over the past decade, thanks to the deep learning revolution that has redefined the landscape of artificial intelligence. In this paper, we propose a novel approach to music generation inspired by musical segment concatenation methods and hash learning algorithms. Given a segment of music, we use a deep recurrent neural network and ranking-based hash learning to assign a forward hash code to the segment to retrieve candidate segments for continuation with matching backward hash codes. The proposed method is thus called Deep Segment Hash Learning (DSHL). To the best of our knowledge, DSHL is the first end-to-end segment hash learning method for music generation, and the first to use pair-wise training with segments of music. We demonstrate that this method is capable of generating music which is both original and enjoyable, and that DSHL offers a promising new direction for music generation research.

In distributed machine learning, where agents collaboratively learn from diverse private data sets, there is a fundamental tension between consensus and optimality. In this paper, we build on recent algorithmic progresses in distributed deep learning to explore various consensus-optimality trade-offs over a fixed communication topology. First, we propose the incremental consensus-based distributed SGD (i-CDSGD) algorithm, which involves multiple consensus steps (where each agent communicates information with its neighbors) within each SGD iteration. Second, we propose the generalized consensus-based distributed SGD (g-CDSGD) algorithm that enables us to navigate the full spectrum from complete consensus (all agents agree) to complete disagreement (each agent converges to individual model parameters). We analytically establish convergence of the proposed algorithms for strongly convex and nonconvex objective functions; we also analyze the momentum variants of the algorithms for the strongly convex case. We support our algorithms via numerical experiments, and demonstrate significant improvements over existing methods for collaborative deep learning.

Corn yield prediction is beneficial as it provides valuable information about production and prices prior the harvest. Publicly available high-quality corn yield prediction can help address emergent information asymmetry problems and in doing so improve price efficiency in futures markets. This paper is the first to employ Long Short-Term Memory (LSTM), a special form of Recurrent Neural Network (RNN) method to predict corn yields. A cross sectional time series of county-level corn yield and hourly weather data made the sample space large enough to use deep learning technics. LSTM is efficient in time series prediction with complex inner relations, which makes it suitable for this task. The empirical results from county level data in Iowa show promising predictive power relative to existing survey based methods.

Adversarial attacks on deep learning models have been demonstrated to be imperceptible to a human, while decreasing the model performance considerably. Attempts to provide invariance against such attacks have denoised adversarial samples to only send cleaned samples to the classifier. In a similar spirit this paper proposes a novel effective strategy that allows to relax adversarial samples onto the underlying manifold of the (unknown) target class distribution. Specifically, given an off-manifold adversarial example, our Metroplis-adjusted Langevin algorithm (Mala) guided through a supervised denoising autoencoder network (sDAE) allows to drive the adversarial samples towards high density regions of the data generating distribution. So, in a nutshell the adversarial example is transformed back from off-manifold onto the data manifold for which the learning model was originally trained and where it can perform well and robustly. Experiments on various benchmark datasets show that our novel Malade method exhibits a high robustness against blackbox and whitebox attacks and outperforms state-of-the-art defense algorithms.

HE FORECASTING of power demand is of crucial importance for the development of modern power systems. The stable and efficient management, scheduling and dispatch in power systems rely heavily on precise forecasting of future loads on various time horizons. In particular, shortterm load forecasting (STLF) focuses on the forecasting of loads from several minutes up to one week into the future [1]. A reliable STLF helps utilities and energy providers deal with the challenges posed by the higher penetration of renewable energies and the development of electricity markets with increasingly complex pricing strategies in future smart grids. Various STLF methods have been proposed by researchers over the years. Some of the models used for STLF include linear or nonparametric regression [2], [3], support vector regression (SVR) [1], [4], autoregressive models [5], fuzzylogic approach [6], etc. Reviews and evaluations of existing methods can be found in [7]-[10]. Building STLF systems with artificial neural networks (ANN) has long been one of the mainstream solutions to this task. As early as 2001, a review paper by Hippert et al. surveyed and examined a collection of papers that had been published between 1991 and 1999, and arrived at the conclusions that most of the proposed models

Understanding the learning dynamics of neural networks is one of the key issues for the improvement of optimization algorithms as well as for the theoretical comprehension of why deep neural nets work so well today. In this paper, we introduce a random matrix-based framework to analyze the learning dynamics of a single-layer linear network on a binary classification problem, for data of simultaneously large dimension and size, trained by gradient descent. Our results provide rich insights into common questions in neural nets, such as overfitting, early stopping and the initialization of training, thereby opening the door for future studies of more elaborate structures and models appearing in today's neural networks.

Abstract--In this paper, we train a recurrent neural network to learn dynamics of a chaotic road environment and to project the future of the environment on an image. Future projection can be used to anticipate an unseen environment for example, in autonomous driving. Road environment is highly dynamic and complex due to the interaction among traffic participants such as vehicles and pedestrians. Even in this complex environment, a human driver is efficacious to safely drive on chaotic roads irrespective of the number of traffic participants. The proliferation of deep learning research has shown the efficacy of neural networks in learning this human behavior . In the same direction, we investigate recurrent neural networks to understand the chaotic road environment which is shared by pedestrians, vehicles (cars, trucks, bicycles etc.), and sometimes animals as well. We propose Foresee, a unidirectional gated recurrent units (GRUs) network with attention to project future of the environment in the form of images. We have collected several videos on Delhi roads consisting of various traffic participants, background and infrastructure differences (like 3D pedestrian crossing) at various times on various days. We train Foresee in an unsupervised way and we use online training to project frames up to 0 . We show that our proposed model performs better than state of the art methods (prednet [20], Enc. Dec. LSTM [28]) and finally, we show that our trained model generalizes to a public dataset for future projections. Environment anticipation is an important task for situation awareness and decision making.

Long short-term memory (LSTM) has been widely used for sequential data modeling. Researchers have increased LSTM depth by stacking LSTM cells to improve performance. This incurs model redundancy, increases run-time delay, and makes the LSTMs more prone to overfitting. To address these problems, we propose a hidden-layer LSTM (H-LSTM) that adds hidden layers to LSTM's original one level non-linear control gates. H-LSTM increases accuracy while employing fewer external stacked layers, thus reducing the number of parameters and run-time latency significantly. We employ grow-and-prune (GP) training to iteratively adjust the hidden layers through gradient-based growth and magnitude-based pruning of connections. This learns both the weights and the compact architecture of H-LSTM control gates. We have GP-trained H-LSTMs for image captioning and speech recognition applications. For the NeuralTalk architecture on the MSCOCO dataset, our three models reduce the number of parameters by 38.7x [floating-point operations (FLOPs) by 45.5x], run-time latency by 4.5x, and improve the CIDEr score by 2.6. For the DeepSpeech2 architecture on the AN4 dataset, our two models reduce the number of parameters by 19.4x (FLOPs by 23.5x), run-time latency by 15.7%, and the word error rate from 12.9% to 8.7%. Thus, GP-trained H-LSTMs can be seen to be compact, fast, and accurate.

While variational dropout approaches have been shown to be effective for network sparsification, they are still suboptimal in the sense that they set the dropout rate for each neuron without consideration of the input data. With such input-independent dropout, each neuron is evolved to be generic across inputs, which makes it difficult to sparsify networks without accuracy loss. To overcome this limitation, we propose adaptive variational dropout whose probabilities are drawn from sparsity-inducing beta-Bernoulli prior. It allows each neuron to be evolved either to be generic or specific for certain inputs, or dropped altogether. Such input-adaptive sparsity- inducing dropout allows the resulting network to tolerate larger degree of sparsity without losing its expressive power by removing redundancies among features. We validate our dependent variational beta-Bernoulli dropout on multiple public datasets, on which it obtains significantly more compact networks than baseline methods, with consistent accuracy improvements over the base networks.

In the past decade, we have seen many successful applications of recurrent neural networks (RNNs) on electronic medical records (EMRs), which contain histories of patients' diagnoses, medications, and other various events, in order to predict the current and future states of patients. Despite the strong performance of RNNs, it is often very challenging for users to understand why the model makes a particular prediction. Such black box nature of RNNs can impede its wide adoption in clinical practice. Furthermore, we have no established method to interactively leverage users' domain expertise and prior knowledge as inputs for steering the model. Therefore, our design study aims to provide a visual analytics solution to increase interpretability and interactivity of RNNs via a joint effort of medical experts, artificial intelligence scientists, and visual analytics researchers. Following the iterative design process between the experts, we design, implement, and evaluate a visual analytics tool called RetainVis, which couples a recently proposed, interpretable RNN-based model called RETAIN and visualizations for users' exploration of EMR data in the context of prediction tasks. Our study shows the effective use of RetainVis for gaining insights into how RNN models EMR data, using real medical records of patients with heart failure, cataract, or dermatological symptoms. Our study also demonstrates how we made substantial changes to the state-of-the-art RNN model called RETAIN in order to make use of temporal information and increase interactivity. This study will provide a useful guideline for researchers who aim to design more interpretable and interactive visual analytics tool for RNNs.