ABSTRACT Deep neural networks (DNNs) have been successfully applied to a wide variety of acoustic modeling tasks in recent years. These include the applications of DNNs either in a discriminative feature extraction or in a hybrid acoustic modeling scenario. Despite the rapid progress in this area, a number of challenges remain in training DNNs. This paper presents an effective way of training DNNs using a manifold learning based regularization framework. In this framework, the parameters of the network are optimized to preserve underlying manifold based relationships between speech feature vectors while minimizing a measure of loss between network outputs and targets. This is achieved by incorporating manifold based locality constraints in the objective criterion of DNNs. Empirical evidence is provided to demonstrate that training a network with manifold constraints preserves structural compactness in the hidden layers of the network. Manifold regularization is applied to train bottleneck DNNs for feature extraction in hidden Markov model (HMM) based speech recognition. The experiments in this work are conducted on the Aurora-2 spoken digits and the Aurora-4 read news large vocabulary continuous speech recognition tasks. The performance is measured in terms of word error rate (WER) on these tasks. It is shown that the manifold regularized DNNs result in up to 37% reduction in WER relative to standard DNNs. Index Terms-- manifold learning, deep neural networks, manifold regularization, manifold regularized deep neural networks, speech recognition 1. INTRODUCTION Recently there has been a resurgence of research in the area of deep neural networks (DNNs) for acoustic modeling in automatic speech recognition (ASR) [1-6]. Much of this research has been concentrated on techniques for regularization of the algorithms used for DNN parameter estimation [7-9]. At the same time, there has also been a great deal of research on graph based techniques that facilitate the preservation of local neighborhood relationships among feature vectors for parameter estimation in a number of application areas [10-13]. Algorithms that preserve these local relationships are often referred to as having the effect of applying manifold based constraints.

Structure Of Intelligent Agents 12. Types of agent program 13. Goal based Agents 14. Utility-based agents 15.

We show how to estimate a model's test error from unlabeled data, on distributions very different from the training distribution, while assuming only that certain conditional independencies are preserved between train and test. We do not need to assume that the optimal predictor is the same between train and test, or that the true distribution lies in any parametric family. We can also efficiently differentiate the error estimate to perform unsupervised discriminative learning. Our technical tool is the method of moments, which allows us to exploit conditional independencies in the absence of a fully-specified model. Our framework encompasses a large family of losses including the log and exponential loss, and extends to structured output settings such as hidden Markov models.

The task of dialog management is commonly decomposed into two sequential subtasks: dialog state tracking and dialog policy learning. In an end-to-end dialog system, the aim of dialog state tracking is to accurately estimate the true dialog state from noisy observations produced by the speech recognition and the natural language understanding modules. The state tracking task is primarily meant to support a dialog policy. From a probabilistic perspective, this is achieved by maintaining a posterior distribution over hidden dialog states composed of a set of context dependent variables. Once a dialog policy is learned, it strives to select an optimal dialog act given the estimated dialog state and a defined reward function. This paper introduces a novel method of dialog state tracking based on a bilinear algebric decomposition model that provides an efficient inference schema through collective matrix factorization. We evaluate the proposed approach on the second Dialog State Tracking Challenge (DSTC-2) dataset and we show that the proposed tracker gives encouraging results compared to the state-of-the-art trackers that participated in this standard benchmark. Finally, we show that the prediction schema is computationally efficient in comparison to the previous approaches.

L. Schwaller · S. Robin Abstract We consider the problem of change-point detection in multivariate time-series. The multivariate distribution of the observations is supposed to follow a graphical model, whose graph and parameters are affected by abrupt changes throughout time. We demonstrate that it is possible to perform exact Bayesian inference whenever one considers a simple class of undirected graphs called spanning trees as possible structures. We are then able to integrate on the graph and segmentation spaces at the same time by combining classical dynamic programming with algebraic results pertaining to spanning trees. In particular, we show that quantities such as posterior distributions for change-points or posterior edge probabilities over time can efficiently be obtained. We illustrate our results on both synthetic and experimental data arising from biology and neuroscience. Keywords change-point detection, exact Bayesian inference, graphical model, multivariate time-series, spanning tree 1 Introduction We are interested in time-series data where several variables are observed throughout time. An assumption often made in multivariate settings is that there exists an underlying network describing the dependences between the different variables. When modelling time-series data, one is faced with a choice: shall this network be considered stationary or not? Taking the example of genomic data, it might for instance be un-L. This network might slowly evolve, or undergo abrupt changes leading to the initialisation of new morphological development stages in the organism of interest. Here, we focus our interest on the second scenario. The inference of the dependence structure ruling a multivariate time-series was first performed under the assumption that this structure was stationary ( e.g.

We consider interactive algorithms in the pool-based setting, and in the stream-based setting. Interactive algorithms observe suggested elements (representing actions or queries), and interactively select some of them and receive responses. Pool-based algorithms can select elements at any order, while stream-based algorithms observe elements in sequence, and can only select elements immediately after observing them. We assume that the suggested elements are generated independently from some source distribution, and ask what is the stream size required for emulating a pool algorithm with a given pool size. We provide algorithms and matching lower bounds for general pool algorithms, and for utility-based pool algorithms. We further show that a maximal gap between the two settings exists also in the special case of active learning for binary classification.

In this work, we consider compressed sensing reconstruction from $M$ measurements of $K$-sparse structured signals which do not possess a writable correlation model. Assuming that a generative statistical model, such as a Boltzmann machine, can be trained in an unsupervised manner on example signals, we demonstrate how this signal model can be used within a Bayesian framework of signal reconstruction. By deriving a message-passing inference for general distribution restricted Boltzmann machines, we are able to integrate these inferred signal models into approximate message passing for compressed sensing reconstruction. Finally, we show for the MNIST dataset that this approach can be very effective, even for $M < K$.

The environment is partially observable and will never be fully observable, due to a lack of information. Does anyone know of any models suitable for learning such a value function?

Gibbs sampling is a Markov Chain Monte Carlo sampling technique that iteratively samples variables from their conditional distributions. There are two common scan orders for the variables: random scan and systematic scan. Due to the benefits of locality in hardware, systematic scan is commonly used, even though most statistical guarantees are only for random scan. While it has been conjectured that the mixing times of random scan and systematic scan do not differ by more than a logarithmic factor, we show by counterexample that this is not the case, and we prove that that the mixing times do not differ by more than a polynomial factor under mild conditions. To prove these relative bounds, we introduce a method of augmenting the state space to study systematic scan using conductance.

Recently a variety of LSTM-based conditional language models (LM) have been applied across a range of language generation tasks. In this work we study various model architectures and different ways to represent and aggregate the source information in an end-to-end neural dialogue system framework. A method called snapshot learning is also proposed to facilitate learning from supervised sequential signals by applying a companion cross-entropy objective function to the conditioning vector. The experimental and analytical results demonstrate firstly that competition occurs between the conditioning vector and the LM, and the differing architectures provide different trade-offs between the two. Secondly, the discriminative power and transparency of the conditioning vector is key to providing both model interpretability and better performance. Thirdly, snapshot learning leads to consistent performance improvements independent of which architecture is used.