Chen, Dexiong, Jacob, Laurent, Mairal, Julien

Substring kernels are classical tools for representing biological sequences or text. However, when large amounts of annotated data is available, models that allow end-to-end training such as neural networks are often prefered. Links between recurrent neural networks (RNNs) and substring kernels have recently been drawn, by formally showing that RNNs with specific activation functions were points in a reproducing kernel Hilbert space (RKHS). In this paper, we revisit this link by generalizing convolutional kernel networks---originally related to a relaxation of the mismatch kernel---to model gaps in sequences. It results in a new type of recurrent neural network which can be trained end-to-end with backpropagation, or without supervision by using kernel approximation techniques.

Hermans, Michiel, Schrauwen, Benjamin

Time series often have a temporal hierarchy, with information that is spread out over multiple time scales. Common recurrent neural networks, however, do not explicitly accommodate such a hierarchy, and most research on them has been focusing on training algorithms rather than on their basic architecture. In this pa- per we study the effect of a hierarchy of recurrent neural networks on processing time series. Here, each layer is a recurrent network which receives the hidden state of the previous layer as input. This architecture allows us to perform hi- erarchical processing on difficult temporal tasks, and more naturally capture the structure of time series.

Naduvil-Vadukootu, Sajitha (Georgia State University) | Angryk, Rafal A. (Georgia State University) | Riley, Pete (Predictive Science, Inc.)

We propose a novel approach to combine state-of-the-art time series data processing methods, such as symbolic aggregate approximation (SAX), with very recently developed deep neural network architectures, such as deep recurrent neural networks (DRNN), for time series data modeling and prediction. Time series data appear extensively in various scientific domains and industrial applications, yet the challenges in accurate modeling and prediction from such data remain open. Deep recurrent neural networks (DRNN) have been proposed as promising approaches to sequence prediction. We extend this research to the new challenge of the time series prediction space, building a system that effectively combines recurrent neural networks (RNN) with time series specific preprocessing techniques. Our experiments show comparisons of model performance with various data preprocessing techniques. We demonstrate that preprocessed inputs can steer us towards simpler (and therefore more computationally efficient) architectures of neural networks (when compared to original inputs).

She, Yiyuan, He, Yuejia, Wu, Dapeng

Large-scale recurrent networks have drawn increasing attention recently because of their capabilities in modeling a large variety of real-world phenomena and physical mechanisms. This paper studies how to identify all authentic connections and estimate system parameters of a recurrent network, given a sequence of node observations. This task becomes extremely challenging in modern network applications, because the available observations are usually very noisy and limited, and the associated dynamical system is strongly nonlinear. By formulating the problem as multivariate sparse sigmoidal regression, we develop simple-to-implement network learning algorithms, with rigorous convergence guarantee in theory, for a variety of sparsity-promoting penalty forms. A quantile variant of progressive recurrent network screening is proposed for efficient computation and allows for direct cardinality control of network topology in estimation. Moreover, we investigate recurrent network stability conditions in Lyapunov's sense, and integrate such stability constraints into sparse network learning. Experiments show excellent performance of the proposed algorithms in network topology identification and forecasting.

GOYAL, Anirudh Goyal ALIAS PARTH, Sordoni, Alessandro, Côté, Marc-Alexandre, Ke, Nan Rosemary, Bengio, Yoshua

Many efforts have been devoted to training generative latent variable models with autoregressive decoders, such as recurrent neural networks (RNN). Stochastic recurrent models have been successful in capturing the variability observed in natural sequential data such as speech. We unify successful ideas from recently proposed architectures into a stochastic recurrent model: each step in the sequence is associated with a latent variable that is used to condition the recurrent dynamics for future steps. Training is performed with amortised variational inference where the approximate posterior is augmented with a RNN that runs backward through the sequence. In addition to maximizing the variational lower bound, we ease training of the latent variables by adding an auxiliary cost which forces them to reconstruct the state of the backward recurrent network.