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

Reinforcement learning, one of the most active research areas in artificial intelligence, is a computational approach to learning whereby an agent tries to maximize the total amount of reward it receives when interacting with a complex, uncertain environment. In Reinforcement Learning, Richard Sutton and Andrew Barto provide a clear and simple account of the key ideas and algorithms of reinforcement learning. Their discussion ranges from the history of the field's intellectual foundations to the most recent developments and applications. The only necessary mathematical background is familiarity with elementary concepts of probability. The book is divided into three parts. Part I defines the reinforcement learning problem in terms of Markov decision processes. Part II provides basic solution methods: dynamic programming, Monte Carlo methods, and temporal-difference learning. Part III presents a unified view of the solution methods and incorporates artificial neural networks, eligibility traces, and planning; the two final chapters present case studies and consider the future of reinforcement learning.

/pdf/reinforcement-learning-an-introduction-rzxgej9p17.pdf

Reinforcement Learning: An Introduction

In the first part of the paper we consider the problem of dynamically apportioning resources among a set of options in a worst-case on-line framework. The model we study can be interpreted as a broad, abstract extension of the well-studied on-line prediction model to a general decision-theoretic setting. We show that the multiplicative weight-update Littlestone?Warmuth rule can be adapted to this model, yielding bounds that are slightly weaker in some cases, but applicable to a considerably more general class of learning problems. We show how the resulting learning algorithm can be applied to a variety of problems, including gambling, multiple-outcome prediction, repeated games, and prediction of points in Rn. In the second part of the paper we apply the multiplicative weight-update technique to derive a new boosting algorithm. This boosting algorithm does not require any prior knowledge about the performance of the weak learning algorithm. We also study generalizations of the new boosting algorithm to the problem of learning functions whose range, rather than being binary, is an arbitrary finite set or a bounded segment of the real line.

/pdf/a-decision-theoretic-generalization-of-on-line-learning-and-3ykfx1u7er.pdf

A Decision-Theoretic Generalization of On-Line Learning and an Application to Boosting

https://www2.econ.iastate.edu/tesfatsi/DeepLearningInNeuralNetworksOverview.JSchmidhuber2015.pdf

Deep learning in neural networks

The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses ‘value networks’ to evaluate board positions and ‘policy networks’ to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of stateof-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.

/pdf/mastering-the-game-of-go-with-deep-neural-networks-and-tree-3viuxo6uxm.pdf

Mastering the game of Go with deep neural networks and tree search

We consider online learning in episodic loop-free Markov decision processes (MDPs), where the loss function can change arbitrarily between episodes. The transition function is fixed but unknown to the learner, and the learner only observes bandit feedback (not the entire loss function). For this problem we develop no-regret algorithms that perform asymptotically as well as the best stationary policy in hindsight. Assuming that all states are reachable with probability $\beta > 0$ under any policy, we give a regret bound of $\tilde{O} ( L|X|\sqrt{|A|T} / \beta )$, where $T$ is the number of episodes, $X$ is the state space, $A$ is the action space, and $L$ is the length of each episode. When this assumption is removed we give a regret bound of $\tilde{O} ( L^{3/2} |X| |A|^{1/4} T^{3/4})$, that holds for an arbitrary transition function. To our knowledge these are the first algorithms that in our setting handle both bandit feedback and an unknown transition function.

/pdf/online-stochastic-shortest-path-with-bandit-feedback-and-3pql4n25tn.pdf

Online Stochastic Shortest Path with Bandit Feedback and Unknown Transition Function

We consider incorporating action elimination procedures in reinforcement learning algorithms. We suggest a framework that is based on learning an upper and a lower estimates of the value function or the Q-function and eliminating actions that are not optimal. We provide a model-based and a model-free variants of the elimination method. We further derive stopping conditions that guarantee that the learned policy is approximately optimal with high probability. Simulations demonstrate a considerable speedup and added robustness.

/pdf/action-elimination-and-stopping-conditions-for-reinforcement-1lyndkyd32.pdf

Action elimination and stopping conditions for reinforcement learning

We propose a general method for converting online algorithms to local computation algorithms by selecting a random permutation of the input, and simulating running the online algorithm. We bound the number of steps of the algorithm using a query tree, which models the dependencies between queries. We improve previous analyses of query trees on graphs of bounded degree, and extend the analysis to the cases where the degrees are distributed binomially, and to a special case of bipartite graphs. 
Using this method, we give a local computation algorithm for maximal matching in graphs of bounded degree, which runs in time and space O(log^3 n). 
We also show how to convert a large family of load balancing algorithms (related to balls and bins problems) to local computation algorithms. This gives several local load balancing algorithms which achieve the same approximation ratios as the online algorithms, but run in O(log n) time and space. 
Finally, we modify existing local computation algorithms for hypergraph 2-coloring and k-CNF and use our improved analysis to obtain better time and space bounds, of O(log^4 n), removing the dependency on the maximal degree of the graph from the exponent.

Converting online algorithms to local computation algorithms

External regret compares the performance of an online algorithm, selecting among N actions, to the performance of the best of those actions in hindsight. Internal regret compares the loss of an online algorithm to the loss of a modified online algorithm, which consistently replaces one action by another.

In this paper, we give a simple generic reduction that, given an algorithm for the external regret problem, converts it to an efficient online algorithm for the internal regret problem. We provide methods that work both in the full information model, in which the loss of every action is observed at each time step, and the partial information (bandit) model, where at each time step only the loss of the selected action is observed. The importance of internal regret in game theory is due to the fact that in a general game, if each player has sublinear internal regret, then the empirical frequencies converge to a correlated equilibrium.

For external regret we also derive a quantitative regret bound for a very general setting of regret, which includes an arbitrary set of modification rules (that possibly modify the online algorithm) and an arbitrary set of time selection functions (each giving different weight to each time step). The regret for a given time selection and modification rule is the difference between the cost of the online algorithm and the cost of the modified online algorithm, where the costs are weighted by the time selection function. This can be viewed as a generalization of the previously-studied sleeping experts setting.

/pdf/from-external-to-internal-regret-ols745ahu6.pdf

From external to internal regret

Bandwidth allocation is a fundamental problem in the design of networks where bandwidth has to be reserved for connections in advance. The problem is intensified when the overall requested bandwidth exceeds the capacity and not all requests can be served. Furthermore, acceptance/rejection decisions regarding connections have to be made online, without knowledge of future requests. We show that the ability to preempt (i.e., abort) connections while in service in order to schedule "more valuable" connections substantially improves the throughput of some networks. We present bandwidth allocation strategies that use preemption and show that they achieve constant competitiveness with respect to the throughput, given that any single call requests at most a constant fraction of the bandwidth. Our results should be contrasted with recent works showing that nonpreemptive strategies have at most inverse logarithmic competitiveness.

Yishay Mansour

Papers

Online Stochastic Shortest Path with Bandit Feedback and Unknown Transition Function

Action elimination and stopping conditions for reinforcement learning

Converting online algorithms to local computation algorithms

From external to internal regret

Bandwidth Allocation with Preemption