Approximate linear programs (ALPs) are well-known models based on value function approximations (VFAs) to obtain heuristic policies and lower bounds on the optimal policy cost of Markov decision processes (MDPs). The ALP VFA is a linear combination of predefined basis functions that are chosen using domain knowledge and updated heuristically if the ALP optimality gap is large. We side-step the need for such basis function engineering in ALP -- an implementation bottleneck -- by proposing a sequence of ALPs that embed increasing numbers of random basis functions obtained via inexpensive sampling. We provide a sampling guarantee and show that the VFAs from this sequence of models converge to the exact value function. Nevertheless, the performance of the ALP policy can fluctuate significantly as more basis functions are sampled. To mitigate these fluctuations, we "self-guide" our convergent sequence of ALPs using past VFA information such that a worst-case measure of policy performance is improved. We perform numerical experiments on perishable inventory control and generalized joint replenishment applications, which, respectively, give rise to challenging discounted-cost MDPs and average-cost semi-MDPs. We find that self-guided ALPs (i) significantly reduce policy cost fluctuations and improve the optimality gaps from an ALP approach that employs basis functions tailored to the former application, and (ii) deliver optimality gaps that are comparable to a known adaptive basis function generation approach targeting the latter application. More broadly, our methodology provides application-agnostic policies and lower bounds to benchmark approaches that exploit application structure.

We show how to solve a number of problems in numerical linear algebra, such as least squares regression, $\ell_p$-regression for any $p \geq 1$, low rank approximation, and kernel regression, in time $T(A) \poly(\log(nd))$, where for a given input matrix $A \in \mathbb{R}^{n \times d}$, $T(A)$ is the time needed to compute $A\cdot y$ for an arbitrary vector $y \in \mathbb{R}^d$. Since $T(A) \leq O(\nnz(A))$, where $\nnz(A)$ denotes the number of non-zero entries of $A$, the time is no worse, up to polylogarithmic factors, as all of the recent advances for such problems that run in input-sparsity time. However, for many applications, $T(A)$ can be much smaller than $\nnz(A)$, yielding significantly sublinear time algorithms. For example, in the overconstrained $(1+\epsilon)$-approximate polynomial interpolation problem, $A$ is a Vandermonde matrix and $T(A) = O(n \log n)$; in this case our running time is $n \cdot \poly(\log n) + \poly(d/\epsilon)$ and we recover the results of \cite{avron2013sketching} as a special case. For overconstrained autoregression, which is a common problem arising in dynamical systems, $T(A) = O(n \log n)$, and we immediately obtain $n \cdot \poly(\log n) + \poly(d/\epsilon)$ time. For kernel autoregression, we significantly improve the running time of prior algorithms for general kernels. For the important case of autoregression with the polynomial kernel and arbitrary target vector $b\in\mathbb{R}^n$, we obtain even faster algorithms. Our algorithms show that, perhaps surprisingly, most of these optimization problems do not require much more time than that of a polylogarithmic number of matrix-vector multiplications.

We are in the process of writing and adding new material (compact eBooks) exclusively available to our members, and written in simple English, by world leading experts in AI, data science, and machine learning. We invite you to sign up here to not miss these free books. This book is intended for busy professionals working with data of any kind: engineers, BI analysts, statisticians, operations research, AI and machine learning professionals, economists, data scientists, biologists, and quants, ranging from beginners to executives. In about 300 pages and 28 chapters it covers many new topics, offering a fresh perspective on the subject, including rules of thumb and recipes that are easy to automate or integrate in black-box systems, as well as new model-free, data-driven foundations to statistical science and predictive analytics. The approach focuses on robust techniques; it is bottom-up (from applications to theory), in contrast to the traditional top-down approach. The material is accessible to practitioners with a one-year college-level exposure to statistics and probability.

This book provides an excellent pathway for gaining first-class expertise in machine learning. It provides both the technical background that explains why certain approaches, but not others, are best practice in real world problems, and a framework for how to think about and approach new problems. I highly recommend it for people with a signal processing background who are seeking to become an expert in machine learning.

Data privacy and security becomes a major concern in building machine learning models from different data providers. Federated learning shows promise by leaving data at providers locally and exchanging encrypted information. This paper studies the vertical federated learning structure for logistic regression where the data sets at two parties have the same sample IDs but own disjoint subsets of features. Existing frameworks adopt the first-order stochastic gradient descent algorithm, which requires large number of communication rounds. To address the communication challenge, we propose a quasi-Newton method based vertical federated learning framework for logistic regression under the additively homomorphic encryption scheme.

We present two new remarkably simple stochastic second-order methods for minimizing the average of a very large number of sufficiently smooth and strongly convex functions. The first is a stochastic variant of Newton's method (SN), and the second is a stochastic variant of cubically regularized Newton's method (SCN). We establish local linear-quadratic convergence results. Unlike existing stochastic variants of second order methods, which require the evaluation of a large number of gradients and/or Hessians in each iteration to guarantee convergence, our methods do not have this shortcoming. For instance, the simplest variants of our methods in each iteration need to compute the gradient and Hessian of a {\em single} randomly selected function only. In contrast to most existing stochastic Newton and quasi-Newton methods, our approach guarantees local convergence faster than with first-order oracle and adapts to the problem's curvature. Interestingly, our method is not unbiased, so our theory provides new intuition for designing new stochastic methods.

This book is used as the textbook for the course EE103 (Stanford) and EE133A (UCLA), where you will find additional related material. If you find an error not listed in our errata list, please do let us know about it. You're welcome to use the lecture slides posted below, but we'd appreciate it if you acknowledge the source.

Also included is an essay from SIAM News'The Functions of Deep Learning' (December 2018) A second distributor for SIAM members is siam.org We will confirm orders for this new book by email.

Gilbert Strang is a professor of mathematics at MIT and perhaps one of the most famous and impactful teachers of math in the world. His MIT OpenCourseWare lectures on linear algebra have been viewed millions of times. This conversation is part of the Artificial Intelligence podcast.

We live in an interconnected world where network valued data arises in many domains, and, fittingly, statistical network analysis has emerged as an active area in the literature. However, the topic of inference in networks has received relatively less attention. In this work, we consider the paired network inference problem where one is given two networks on the same set of nodes, and the goal is to test whether the given networks are stochastically similar in terms of some notion of similarity. We develop a general inferential framework based on parametric bootstrap to address this problem. Under this setting, we address two specific and important problems: the equality problem, i.e., whether the two networks are generated from the same random graph model, and the scaling problem, i.e., whether the underlying probability matrices of the two random graph models are scaled versions of each other.