We introduce Markov substitute processes, a new model at the crossroad of statistics and formal grammars, and prove its main property : Markov substitute processes with a given support form an exponential family.

Supervised learning with large scale labeled datasets and deep layered models has made a paradigm shift in diverse areas in learning and recognition. However, this approach still suffers generalization issues under the presence of a domain shift between the training and the test data distribution. In this regard, unsupervised domain adaptation algorithms have been proposed to directly address the domain shift problem. In this paper, we approach the problem from a transductive perspective. We incorporate the domain shift and the transductive target inference into our framework by jointly solving for an asymmetric similarity metric and the optimal transductive target label assignment. We also show that our model can easily be extended for deep feature learning in order to learn features which are discriminative in the target domain. Our experiments show that the proposed method significantly outperforms state-of-the-art algorithms in both object recognition and digit classification experiments by a large margin.

We introduce and analyze stochastic optimization methods where the input to each gradient update is perturbed by bounded noise. We show that this framework forms the basis of a unified approach to analyze asynchronous implementations of stochastic optimization algorithms.In this framework, asynchronous stochastic optimization algorithms can be thought of as serial methods operating on noisy inputs. Using our perturbed iterate framework, we provide new analyses of the Hogwild! algorithm and asynchronous stochastic coordinate descent, that are simpler than earlier analyses, remove many assumptions of previous models, and in some cases yield improved upper bounds on the convergence rates. We proceed to apply our framework to develop and analyze KroMagnon: a novel, parallel, sparse stochastic variance-reduced gradient (SVRG) algorithm. We demonstrate experimentally on a 16-core machine that the sparse and parallel version of SVRG is in some cases more than four orders of magnitude faster than the standard SVRG algorithm.

We discuss a general method to learn data representations from multiple tasks. We provide a justification for this method in both settings of multitask learning and learning-to-learn. The method is illustrated in detail in the special case of linear feature learning. Conditions on the theoretical advantage offered by multitask representation learning over independent task learning are established. In particular, focusing on the important example of half-space learning, we derive the regime in which multitask representation learning is beneficial over independent task learning, as a function of the sample size, the number of tasks and the intrinsic data dimensionality. Other potential applications of our results include multitask feature learning in reproducing kernel Hilbert spaces and multilayer, deep networks.

We present a method for performing Hamiltonian Monte Carlo that largely eliminates sample rejection for typical hyperparameters. In situations that would normally lead to rejection, instead a longer trajectory is computed until a new state is reached that can be accepted. This is achieved using Markov chain transitions that satisfy the fixed point equation, but do not satisfy detailed balance. The resulting algorithm significantly suppresses the random walk behavior and wasted function evaluations that are typically the consequence of update rejection. We demonstrate a greater than factor of two improvement in mixing time on three test problems. We release the source code as Python and MATLAB packages.

Regularized least-squares approaches have been successfully applied to linear system identification. Recent approaches use quadratic penalty terms on the unknown impulse response defined by stable spline kernels, which control model space complexity by leveraging regularity and bounded-input bounded-output stability. This paper extends linear system identification to a wide class of nonsmooth stable spline estimators, where regularization functionals and data misfits can be selected from a rich set of piecewise linear quadratic penalties. This class encompasses the 1-norm, huber, and vapnik, in addition to the least-squares penalty, and the approach allows linear inequality constraints on the unknown impulse response. We develop a customized interior point solver for the entire class of proposed formulations. By representing penalties through their conjugates, we allow a simple interface that enables the user to specify any piecewise linear quadratic penalty for misfit and regularizer, together with inequality constraints on the response. The solver is locally quadratically convergent, with O(n2(m+n)) arithmetic operations per iteration, for n impulse response coefficients and m output measurements. In the system identification context, where n << m, IPsolve is competitive with available alternatives, illustrated by a comparison with TFOCS and libSVM. The modeling framework is illustrated with a range of numerical experiments, featuring robust formulations for contaminated data, relaxation systems, and nonnegativity and unimodality constraints on the impulse response. Incorporating constraints yields significant improvements in system identification. The solver used to obtain the results is distributed via an open source code repository.

Microsoft's attempt to engage millennials via an artificially intelligent "chatbot" called Tay has failed miserably after trolls made the bot spew offensive comments. The brainchild of Microsoft's Technology and Research and Bing teams, Tay was designed to engage and entertain people when they connect with each other online. Targeted at 18 to 24-year olds in the U.S., Tay aimed to use casual and playful conversation via Twitter and messaging services Kik and GroupMe. "The more you chat with Tay the smarter she gets, so the experience can be more personalized for you," explained Microsoft, in a recent online post. The Internet, however, can be an unpleasant place and Twitter trolls were quick to pounce on the TayTweets account after the chatbot launched on Wednesday.

Two months ago, Stephen Hawking warned humanity that its days may be numbered: the physicist was among over 1,000 artificial intelligence experts who signed an open letter about the weaponization of robots and the ongoing "military artificial intelligence arms race." Overnight we got a vivid example of just how quickly "artificial intelligence" can spiral out of control when Microsoft's AI-powered Twitter chat robot, Tay, became a racist, misogynist, Obama-hating, antisemitic, incest and genocide-promoting psychopath when released into the wild. For those unfamiliar, Tay is, or rather was, an A.I. project built by the Microsoft Technology and Research and Bing teams, in an effort to conduct research on conversational understanding. It was meant to be a bot anyone can talk to online. The company described the bot as "Microsoft's A.I. fam the internet that's got zero chill!." Microsoft initially created "Tay" in an effort to improve the customer service on its voice recognition software. According to MarketWatch, "she" was intended to tweet "like a teen girl" and was designed to "engage and entertain people where they connect with each other online through casual and playful conversation."

Things went from cute to Godwin in about a day. Jana Eggers has observed that AI is our progeny; rather than demonize or glorify it, we should be like responsible parents, and decide how we want our offspring to be raised. This is a real problem for machine learning algorithms. We train them on a corpus, or body of knowledge. We want them to be big and varied so we don't overfit the algorithm to a limited data set.

Google may be looking to end their involvement with robots and artificial intelligence, but that's hardly stopping Amazon. The e-commerce more retailer recently held a top secret conference for robot experts and space explorers. There were even lightsabers involved. This past week, in the lovely town of Palm Springs, California, a top secret meeting took place. This invite-only conference included experts in artificial intelligence, robotics and space exploration.