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

From the Publisher:
A valuable reference for the novice as well as for the expert who needs a wider scope of coverage within the area of cryptography, this book provides easy and rapid access of information and includes more than 200 algorithms and protocols; more than 200 tables and figures; more than 1,000 numbered definitions, facts, examples, notes, and remarks; and over 1,250 significant references, including brief comments on each paper.

Handbook of Applied Cryptography

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

This survey provides an overview of higher-order tensor decompositions, their applications, and available software. A tensor is a multidimensional or $N$-way array. Decompositions of higher-order tensors (i.e., $N$-way arrays with $N \geq 3$) have applications in psycho-metrics, chemometrics, signal processing, numerical linear algebra, computer vision, numerical analysis, data mining, neuroscience, graph analysis, and elsewhere. Two particular tensor decompositions can be considered to be higher-order extensions of the matrix singular value decomposition: CANDECOMP/PARAFAC (CP) decomposes a tensor as a sum of rank-one tensors, and the Tucker decomposition is a higher-order form of principal component analysis. There are many other tensor decompositions, including INDSCAL, PARAFAC2, CANDELINC, DEDICOM, and PARATUCK2 as well as nonnegative variants of all of the above. The N-way Toolbox, Tensor Toolbox, and Multilinear Engine are examples of software packages for working with tensors.

/pdf/tensor-decompositions-and-applications-46vl2ih5gn.pdf

Tensor Decompositions and Applications

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

We discuss some results about the efficient approximability of constraint satisfaction problems. in particular we focus on the question on an efficient algorithm can perform significantly better than the algorithm that picks a solution uniformly at random.

/pdf/on-the-efficient-approximability-of-constraint-satisfaction-53q8ngkr15.pdf

On the efficient approximability of constraint satisfaction problems

We present a reformulation of the 2n+o(n) lower bound of Bent and John [Proceedings of the 17th Annual ACM Symposium on Theory of Computing, 1985, pp. 213--216] for the number of comparisons needed for selecting the median of n elements. Our reformulation uses a weight function. Apart from giving a more intuitive proof for the lower bound, the new formulation opens up possibilities for improving it. We use the new formulation to show that any pair-forming median finding algorithm, i.e., a median finding algorithm that starts by comparing $\lfloor n/2\rfloor$ disjoint pairs of elements must perform, in the worst case, at least 2.01 n + o(n) comparisons. This provides strong evidence that selecting the median requires at least cn+o(n) comparisons for some c> 2.

On Lower Bounds for Selecting the Median

We present a top-down lower bound method for depth 3 AND-OR-NOT circuits which is simpler than the previous methods and in some cases gives better lower bounds. In particular we prove that depth 3 AND-OR-NOT circuits that compute PARITY resp. MAJORITY require size at least 2/sup 0.618/ .../spl radic/n/ resp. 2/sup 0.849/.../spl radic/n/. This is the first simple proof of a strong lower bound by a top-down argument for non-monotone circuits. >

Top-down lower bounds for depth 3 circuits

A well-characterized approximation problem

We introduce a new method to construct approximation algorithms for combinatorial optimization problems using semidefinite programming. It consists of expressing each combinatorial object in the original problem as a constellation of vectors in the semidefinite program. When we apply this technique to systems of linear equations mod p with at most two variables in each equation, we can show that the problem is approximable within (1 − κ(p))p, where κ(p) > 0 for all p. Using standard techniques, we also show that it is NP-hard to approximate the problem within a constant ratio, independent of p. Warning: Essentially this paper has been published in Journal of Algorithms and is subject to copyright restrictions. In particular it is for personal use only. New Use of Semidefinite Programming 4 List of symbols used: α lower case Greek letter alpha. δ lower case Greek letter delta. e lower case Greek letter epsilon. κ lower case Greek letter kappa. λ lower case Greek letter lambda. μ lower case Greek letter mu. π lower case Greek letter pi. σ lower case Greek letter sigma. θ lower case Greek letter theta. Θ upper case Greek letter Theta. φ lower case Greek letter phi. Φ upper case Greek letter Phi. ζ lower case Greek letter zeta. ξ lower case Greek letter xi. lower case Latin letter ell. ∫ integral sign. ∑ summation sign. R bold face upper case Roman letter ar (denotes the set of reals). Z bold face upper case Roman letter zed (denotes the set of integers). ⊥ perpendicular sign. ← left arrow. → right arrow with small bar (denotes “maps to”). =⇒ double right arrow (denotes implication). ⇐⇒ double left-right arrow (denotes equivalence). ∅ the empty set. ∪ set union. ∩ set intersection. ∈ belongs to sign. | · | absolute value. ‖ · ‖ Euclidean norm. 〈·, ·〉 angular brackets (denotes inner product) ∀ universal quantifier. ∃ existential quantifier. √· square root sign. × multiplication sign. New Use of Semidefinite Programming 5

/pdf/a-new-way-to-use-semidefinite-programming-with-applications-52cn7sjfx6.pdf

Johan Håstad

Papers

On the efficient approximability of constraint satisfaction problems

On Lower Bounds for Selecting the Median

Top-down lower bounds for depth 3 circuits

A well-characterized approximation problem

A new way to use semidefinite programming with applications to linear equations mod p