Deep learning is a form of machine learning that enables computers to learn from experience and understand the world in terms of a hierarchy of concepts. Because the computer gathers knowledge from experience, there is no need for a human computer operator to formally specify all the knowledge that the computer needs. The hierarchy of concepts allows the computer to learn complicated concepts by building them out of simpler ones; a graph of these hierarchies would be many layers deep. This book introduces a broad range of topics in deep learning. The text offers mathematical and conceptual background, covering relevant concepts in linear algebra, probability theory and information theory, numerical computation, and machine learning. It describes deep learning techniques used by practitioners in industry, including deep feedforward networks, regularization, optimization algorithms, convolutional networks, sequence modeling, and practical methodology; and it surveys such applications as natural language processing, speech recognition, computer vision, online recommendation systems, bioinformatics, and videogames. Finally, the book offers research perspectives, covering such theoretical topics as linear factor models, autoencoders, representation learning, structured probabilistic models, Monte Carlo methods, the partition function, approximate inference, and deep generative models. Deep Learning can be used by undergraduate or graduate students planning careers in either industry or research, and by software engineers who want to begin using deep learning in their products or platforms. A website offers supplementary material for both readers and instructors.

Deep Learning

We describe latent Dirichlet allocation (LDA), a generative probabilistic model for collections of discrete data such as text corpora. LDA is a three-level hierarchical Bayesian model, in which each item of a collection is modeled as a finite mixture over an underlying set of topics. Each topic is, in turn, modeled as an infinite mixture over an underlying set of topic probabilities. In the context of text modeling, the topic probabilities provide an explicit representation of a document. We present efficient approximate inference techniques based on variational methods and an EM algorithm for empirical Bayes parameter estimation. We report results in document modeling, text classification, and collaborative filtering, comparing to a mixture of unigrams model and the probabilistic LSI model.

/pdf/latent-dirichlet-allocation-4gmscb06en.pdf

Latent dirichlet allocation

We propose a generative model for text and other collections of discrete data that generalizes or improves on several previous models including naive Bayes/unigram, mixture of unigrams [6], and Hof-mann's aspect model, also known as probabilistic latent semantic indexing (pLSI) [3]. In the context of text modeling, our model posits that each document is generated as a mixture of topics, where the continuous-valued mixture proportions are distributed as a latent Dirichlet random variable. Inference and learning are carried out efficiently via variational algorithms. We present empirical results on applications of this model to problems in text modeling, collaborative filtering, and text classification.

/pdf/latent-dirichlet-allocation-rkwbqihm3p.pdf

Latent Dirichlet Allocation

Data Mining: Practical Machine Learning Tools and Techniques offers a thorough grounding in machine learning concepts as well as practical advice on applying machine learning tools and techniques in real-world data mining situations. This highly anticipated third edition of the most acclaimed work on data mining and machine learning will teach you everything you need to know about preparing inputs, interpreting outputs, evaluating results, and the algorithmic methods at the heart of successful data mining. Thorough updates reflect the technical changes and modernizations that have taken place in the field since the last edition, including new material on Data Transformations, Ensemble Learning, Massive Data Sets, Multi-instance Learning, plus a new version of the popular Weka machine learning software developed by the authors. Witten, Frank, and Hall include both tried-and-true techniques of today as well as methods at the leading edge of contemporary research. *Provides a thorough grounding in machine learning concepts as well as practical advice on applying the tools and techniques to your data mining projects *Offers concrete tips and techniques for performance improvement that work by transforming the input or output in machine learning methods *Includes downloadable Weka software toolkit, a collection of machine learning algorithms for data mining tasks-in an updated, interactive interface. Algorithms in toolkit cover: data pre-processing, classification, regression, clustering, association rules, visualization

Data Mining: Practical Machine Learning Tools and Techniques

抗原变异可使得多种致病微生物易于逃避宿主免疫应答。表达在感染红细胞表面的恶性疟原虫红细胞表面蛋白1（PfPMP1）与感染红细胞、内皮细胞、树突状细胞以及胎盘的单个或多个受体作用，在黏附及免疫逃避中起关键的作用。每个单倍体基因组var基因家族编码约60种成员，通过启动转录不同的var基因变异体为抗原变异提供了分子基础。

疟原虫var基因转换速率变化导致抗原变异[英]／Paul H, Robert P, Christodoulou Z, et al//Proc Natl Acad Sci U S A

Methods and devices for providing a user-interface are disclosed. In one aspect, a head-mounted-device system includes a processor data storage comprising user-interface logic executable by the at least one processor to receive data corresponding to first position of a head-mounted display (HMD) and responsively cause the HMD to display a user-interface comprising a view region, at least one content region located above the view region, and a history region located below the view region. The user-interface logic is further executable to receive data corresponding to an left or right movement of the HMD and responsively cause the HMD to move the field of view such that the at least one content region becomes more visible, for example, scrolling an item in a user interface. The scrolling may have a non-linear relationship with the head movement speed.

Input detection for a head mounted device

This paper describes a robot controller which uses probabilistic decision-making techniques at the highest-level of behavior control. The POMDP-based robot controller has the ability to incorporate noisy and partial sensor information, and can arbitrate between information gathering and performance-related actions. The complexity of the robot control domain requires a POMDP model that is beyond the capability of current exact POMDP solvers, therefore we present a hierarchical variant of the POMDP model which exploits structure in the problem domain to accelerate planning. This POMDP controller is implemented and tested onboard a mobile robot in the context of an interactive service task. During the course of experiments conducted in an assisted living facility, the robot successfully demonstrated that it could autonomously provide guidance and information to elderly residents with mild physical and cognitive disabilities.

/pdf/high-level-robot-behavior-control-using-pomdps-1yi8ifwrmu.pdf

High-level robot behavior control using POMDPs

This paper summarizes a probabilistic approach for localizing people through the signal strengths of a wireless IEEE 802.1 lb network. Our approach uses data labeled by ground truth position to learn a probabilistic mapping from locations to wireless signals, represented by piecewise linear Gaussians. It then uses sequences of wireless signal data (without position labels) to acquire motion models of individual people, which further improves the localization accuracy. The approach has been implemented and evaluated in an office environment.

/pdf/a-learning-algorithm-for-localizing-people-based-on-wireless-ypyeif8vnt.pdf

A learning algorithm for localizing people based on wireless signal strength that uses labeled and unlabeled data

In this paper we present a method of computing the posterior probability of conditional independence of two or more continuous variables from data, examined at several resolutions Our approach is motivated by the observation that the appearance of continuous data varies widely at various resolutions, producing very different independence estimates between the variables involved Therefore, it is difficult to ascertain independence without examining discretized data at several carefully selected resolutions In our paper, we accomplish this using the exact computation of the posterior probability of independence, calculated analytically given a resolution At each examined resolution and boundary placement, we assume a multinomial distribution with Dirichlet priors for the discretized table parameters, and compute the posterior using Bayesian integration Across resolutions, we use a search procedure to approximate the Bayesian integral of probability over an exponential number of possible boundary placements Our method generalizes to an arbitrary number variables in a straightforward manner The test is suitable for Bayesian network learning algorithms that use independence tests to infer the network structure, in domains that contain any mix of continuous, ordinal discrete, and categorical variables

/pdf/a-bayesian-multiresolution-independence-test-for-continuous-2wo65tbeit.pdf

A Bayesian multiresolution independence test for continuous variables

This chapter introduces the major learning approach studied in this book: the explanation-based neural network learning algorithm (EBNN) EBNN approaches the meta-level learning problem by learning a theory of the domain This domain theory is domain-specific It characterizes, for example, the relevance of individual features, their cross-dependencies, or certain invariant properties of the domain that apply to all learning tasks within the domain Obviously, when the learner has a model of such regularities, there is an opportunity to generalize more accurately or, alternatively, learn from less training data This is because without knowledge about these regularities the learner has to learn them from scratch, which necessarily requires more training data EBNN transfers previously learned knowledge by explaining and analyzing training examples in terms of the domain theory The result of this analysis is a set of domain-specific shape constraints for the function to be learned at the base-level Thus, these constraints guide the base-level learning of new functions in a knowledgeable, domain-specific way

Sebastian Thrun

Papers

Input detection for a head mounted device

High-level robot behavior control using POMDPs

A learning algorithm for localizing people based on wireless signal strength that uses labeled and unlabeled data

A Bayesian multiresolution independence test for continuous variables

Explanation-based neural network learning