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

The success of machine learning algorithms generally depends on data representation, and we hypothesize that this is because different representations can entangle and hide more or less the different explanatory factors of variation behind the data. Although specific domain knowledge can be used to help design representations, learning with generic priors can also be used, and the quest for AI is motivating the design of more powerful representation-learning algorithms implementing such priors. This paper reviews recent work in the area of unsupervised feature learning and deep learning, covering advances in probabilistic models, autoencoders, manifold learning, and deep networks. This motivates longer term unanswered questions about the appropriate objectives for learning good representations, for computing representations (i.e., inference), and the geometrical connections between representation learning, density estimation, and manifold learning.

/pdf/representation-learning-a-review-and-new-perspectives-4sidphwbhg.pdf

Representation Learning: A Review and New Perspectives

Whereas before 2006 it appears that deep multilayer neural networks were not successfully trained, since then several algorithms have been shown to successfully train them, with experimental results showing the superiority of deeper vs less deep architectures. All these experimental results were obtained with new initialization or training mechanisms. Our objective here is to understand better why standard gradient descent from random initialization is doing so poorly with deep neural networks, to better understand these recent relative successes and help design better algorithms in the future. We first observe the influence of the non-linear activations functions. We find that the logistic sigmoid activation is unsuited for deep networks with random initialization because of its mean value, which can drive especially the top hidden layer into saturation. Surprisingly, we find that saturated units can move out of saturation by themselves, albeit slowly, and explaining the plateaus sometimes seen when training neural networks. We find that a new non-linearity that saturates less can often be beneficial. Finally, we study how activations and gradients vary across layers and during training, with the idea that training may be more difficult when the singular values of the Jacobian associated with each layer are far from 1. Based on these considerations, we propose a new initialization scheme that brings substantially faster convergence. 1 Deep Neural Networks Deep learning methods aim at learning feature hierarchies with features from higher levels of the hierarchy formed by the composition of lower level features. They include Appearing in Proceedings of the 13 International Conference on Artificial Intelligence and Statistics (AISTATS) 2010, Chia Laguna Resort, Sardinia, Italy. Volume 9 of JMLR: WC Weston et al., 2008). Much attention has recently been devoted to them (see (Bengio, 2009) for a review), because of their theoretical appeal, inspiration from biology and human cognition, and because of empirical success in vision (Ranzato et al., 2007; Larochelle et al., 2007; Vincent et al., 2008) and natural language processing (NLP) (Collobert & Weston, 2008; Mnih & Hinton, 2009). Theoretical results reviewed and discussed by Bengio (2009), suggest that in order to learn the kind of complicated functions that can represent high-level abstractions (e.g. in vision, language, and other AI-level tasks), one may need deep architectures. Most of the recent experimental results with deep architecture are obtained with models that can be turned into deep supervised neural networks, but with initialization or training schemes different from the classical feedforward neural networks (Rumelhart et al., 1986). Why are these new algorithms working so much better than the standard random initialization and gradient-based optimization of a supervised training criterion? Part of the answer may be found in recent analyses of the effect of unsupervised pretraining (Erhan et al., 2009), showing that it acts as a regularizer that initializes the parameters in a “better” basin of attraction of the optimization procedure, corresponding to an apparent local minimum associated with better generalization. But earlier work (Bengio et al., 2007) had shown that even a purely supervised but greedy layer-wise procedure would give better results. So here instead of focusing on what unsupervised pre-training or semi-supervised criteria bring to deep architectures, we focus on analyzing what may be going wrong with good old (but deep) multilayer neural networks. Our analysis is driven by investigative experiments to monitor activations (watching for saturation of hidden units) and gradients, across layers and across training iterations. We also evaluate the effects on these of choices of activation function (with the idea that it might affect saturation) and initialization procedure (since unsupervised pretraining is a particular form of initialization and it has a drastic impact).

/pdf/understanding-the-difficulty-of-training-deep-feedforward-4q8w9nfz3x.pdf

Understanding the difficulty of training deep feedforward neural networks

Can machine learning deliver AI? Theoretical results, inspiration from the brain and cognition, as well as machine learning experiments suggest that in order to learn the kind of complicated functions that can represent high-level abstractions (e.g. in vision, language, and other AI-level tasks), one would need deep architectures. Deep architectures are composed of multiple levels of non-linear operations, such as in neural nets with many hidden layers, graphical models with many levels of latent variables, or in complicated propositional formulae re-using many sub-formulae. Each level of the architecture represents features at a different level of abstraction, defined as a composition of lower-level features. Searching the parameter space of deep architectures is a difficult task, but new algorithms have been discovered and a new sub-area has emerged in the machine learning community since 2006, following these discoveries. Learning algorithms such as those for Deep Belief Networks and other related unsupervised learning algorithms have recently been proposed to train deep architectures, yielding exciting results and beating the state-of-the-art in certain areas. Learning Deep Architectures for AI discusses the motivations for and principles of learning algorithms for deep architectures. By analyzing and comparing recent results with different learning algorithms for deep architectures, explanations for their success are proposed and discussed, highlighting challenges and suggesting avenues for future explorations in this area.

/pdf/learning-deep-architectures-for-ai-2358znl654.pdf

Learning Deep Architectures for AI

https://web.cs.wpi.edu/~rail/files/ArgallEtAlRAS09.pdf

A survey of robot learning from demonstration

The DARPA PerceptOR program has implemented a rigorous evaluative test program which fosters the development of field relevant outdoor mobile robots. Autonomous ground vehicles were deployed on diverse test courses throughout the USA and quantitatively evaluated on such factors as autonomy level, waypoint acquisition, failure rate, speed, and communications bandwidth. Our efforts over the three year program have produced new approaches in planning, perception, localization, and control which have been driven by the quest for reliable operation in challenging environments. This paper focuses on some of the most unique aspects of the systems developed by the CMU PerceptOR team, the lessons learned during the effort, and the most immediate challenges that remain to be addressed.

Toward Reliable Off Road Autonomous Vehicles Operating in Challenging Environments

The Maximum Margin Planning (MMP) (Ratliff et al., 2006) algorithm solves imitation learning problems by learning linear mappings from features to cost functions in a planning domain. The learned policy is the result of minimum-cost planning using these cost functions. These mappings are chosen so that example policies (or trajectories) given by a teacher appear to be lower cost (with a loss-scaled margin) than any other policy for a given planning domain. We provide a novel approach, MMPBOOST, based on the functional gradient descent view of boosting (Mason et al., 1999; Friedman, 1999a) that extends MMP by "boosting" in new features. This approach uses simple binary classification or regression to improve performance of MMP imitation learning, and naturally extends to the class of structured maximum margin prediction problems. (Taskar et al., 2005) Our technique is applied to navigation and planning problems for outdoor mobile robots and robotic legged locomotion.

/pdf/boosting-structured-prediction-for-imitation-learning-22onchuige.pdf

Boosting Structured Prediction for Imitation Learning

Prior work has shown that features which appear to be biologically plausible as well as empirically useful can be found by sparse coding with a prior such as a laplacian (L1) that promotes sparsity. We show how smoother priors can preserve the benefits of these sparse priors while adding stability to the Maximum A-Posteriori (MAP) estimate that makes it more useful for prediction problems. Additionally, we show how to calculate the derivative of the MAP estimate efficiently with implicit differentiation. One prior that can be differentiated this way is KL-regularization. We demonstrate its effectiveness on a wide variety of applications, and find that online optimization of the parameters of the KL-regularized model can significantly improve prediction performance.

/pdf/differentiable-sparse-coding-22l7psw0wk.pdf

Differentiable Sparse Coding

Autonomous navigation by a mobile robot through L natural, unstructured terrain is one of the premier k challenges in field robotics. Tremendous advances V in autonomous navigation have been made recently in field robotics. Machine learning has played an increasingly important role in these advances. The Defense Advanced Research Projects Agency (DARPA) UGCV-Perceptor Integration (UPI) program was conceived to take a fresh approach to all aspects of autonomous outdoor mobile robot design, from vehicle design to the design of perception and control systems with the goal of achieving a leap in performance to enable the next generation of robotic applications in commercial, industrial, and military applications. The essential problem addressed by the UPI program is to enable safe autonomous traverse of a robot from Point A to Point B in the least time possible given a series of waypoints in complex, unstructured terrain separated by 0.2-2 km. To accomplish this goal, machine learning techniques were heavily used to provide robust and adaptive performance, while simultaneously reducing the required development and deployment time. This article describes the autonomous system, Crusher, developed for the UPI program and the learning approaches that aided in its successful performance.

Learning for Autonomous Navigation

A key challenge for autonomous navigation in cluttered outdoor environments is the reliable discrimination between obstacles that must be avoided at all costs, and lesser obstacles which the robot can drive over if necessary. Chlorophyll-rich vegetation in particular is often not an obstacle to a capable off-road vehicle, and it has long been recognized in the satellite imaging community that a simple comparison of the red and near-infrared (NIR) reflectance of a material provides a reliable technique for measuring chlorophyll content in natural scenes. This paper evaluates the effectiveness of using this chlorophyll-detection technique to improve autonomous navigation in natural, off-road environments. We demonstrate through extensive experiments that this feature has properties complementary to the color and shape descriptors traditionally used for point cloud analysis, and show significant improvement in classification performance for tasks relevant to outdoor navigation. Results are shown from field testing onboard a robot operating in off-road terrain.

/pdf/vegetation-detection-for-driving-in-complex-environments-12k1nk7vv2.pdf

David McAllister Bradley

Papers

Toward Reliable Off Road Autonomous Vehicles Operating in Challenging Environments

Boosting Structured Prediction for Imitation Learning

Differentiable Sparse Coding

Learning for Autonomous Navigation

Vegetation Detection for Driving in Complex Environments