We propose a novel, efficient approach for distributed sparse learning in high-dimensions, where observations are randomly partitioned across machines. Computationally, at each round our method only requires the master machine to solve a shifted ell_1 regularized M-estimation problem, and other workers to compute the gradient. In respect of communication, the proposed approach provably matches the estimation error bound of centralized methods within constant rounds of communications (ignoring logarithmic factors). We conduct extensive experiments on both simulated and real world datasets, and demonstrate encouraging performances on high-dimensional regression and classification tasks.

In many machine learning applications, crowdsourcing has become the primary means for label collection. In this paper, we study the optimal error rate for aggregating labels provided by a set of non-expert workers. Under the classic Dawid-Skene model, we establish matching upper and lower bounds with an exact exponent $mI(\pi)$ in which $m$ is the number of workers and $I(\pi)$ the average Chernoff information that characterizes the workers' collective ability. Such an exact characterization of the error exponent allows us to state a precise sample size requirement $m>\frac{1}{I(\pi)}\log\frac{1}{\epsilon}$ in order to achieve an $\epsilon$ misclassification error. In addition, our results imply the optimality of various EM algorithms for crowdsourcing initialized by consistent estimators.

Humans have an impressive ability to reason about new concepts and experiences from just a single example. In particular, humans have an ability for one-shot generalization: an ability to encounter a new concept, understand its structure, and then be able to generate compelling alternative variations of the concept. We develop machine learning systems with this important capacity by developing new deep generative models, models that combine the representational power of deep learning with the inferential power of Bayesian reasoning. We develop a class of sequential generative models that are built on the principles of feedback and attention. These two characteristics lead to generative models that are among the state-of-the art in density estimation and image generation. We demonstrate the one-shot generalization ability of our models using three tasks: unconditional sampling, generating new exemplars of a given concept, and generating new exemplars of a family of concepts. In all cases our models are able to generate compelling and diverse samples---having seen new examples just once---providing an important class of general-purpose models for one-shot machine learning.

Partition functions of probability distributions are important quantities for model evaluation and comparisons. We present a new method to compute partition functions of complex and multimodal distributions. Such distributions are often sampled using simulated tempering, which augments the target space with an auxiliary inverse temperature variable. Our method exploits the multinomial probability law of the inverse temperatures, and provides estimates of the partition function in terms of a simple quotient of Rao-Blackwellized marginal inverse temperature probability estimates, which are updated while sampling. We show that the method has interesting connections with several alternative popular methods, and offers some significant advantages. In particular, we empirically find that the new method provides more accurate estimates than Annealed Importance Sampling when calculating partition functions of large Restricted Boltzmann Machines (RBM); moreover, the method is sufficiently accurate to track training and validation log-likelihoods during learning of RBMs, at minimal computational cost.

Zoubin Ghahramani We show that a neural network with arbitrary depth and non-linearities, with dropout applied before every weight layer, is mathematically equivalent to an approximation to a well known Bayesian model. This interpretation might offer an explanation to some of dropout's key properties, such as its robustness to overfitting. Our interpretation allows us to reason about uncertainty in deep learning, and allows the introduction of the Bayesian machinery into existing deep learning frameworks in a principled way. This document is an appendix for the main paper "Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning" by Gal and Ghahramani, 2015 (http://arxiv.org/abs/1506.02142).

Machines, not humans, are the world's dominant knowledge accumulators but humans remain the dominant decision makers. Interpreting and disseminating the knowledge accumulated by machines requires expertise, time, and is prone to failure. The problem of how best to convey accumulated knowledge from computers to humans is a critical bottleneck in the broader application of machine learning. We propose an approach based on human teaching where the problem is formalized as selecting a small subset of the data that will, with high probability, lead the human user to the correct inference. This approach, though successful for modeling human learning in simple laboratory experiments, has failed to achieve broader relevance due to challenges in formulating general and scalable algorithms. We propose general-purpose teaching via pseudo-marginal sampling and demonstrate the algorithm by teaching topic models. Simulation results show our sampling-based approach: effectively approximates the probability where ground-truth is possible via enumeration, results in data that are markedly different from those expected by random sampling, and speeds learning especially for small amounts of data. Application to movie synopsis data illustrates differences between teaching and random sampling for teaching distributions and specific topics, and demonstrates gains in scalability and applicability to real-world problems.

Recurrent neural networks (RNNs) are notoriously difficult to train. When the eigenvalues of the hidden to hidden weight matrix deviate from absolute value 1, optimization becomes difficult due to the well studied issue of vanishing and exploding gradients, especially when trying to learn long-term dependencies. To circumvent this problem, we propose a new architecture that learns a unitary weight matrix, with eigenvalues of absolute value exactly 1. The challenge we address is that of parametrizing unitary matrices in a way that does not require expensive computations (such as eigendecomposition) after each weight update. We construct an expressive unitary weight matrix by composing several structured matrices that act as building blocks with parameters to be learned. Optimization with this parameterization becomes feasible only when considering hidden states in the complex domain. We demonstrate the potential of this architecture by achieving state of the art results in several hard tasks involving very long-term dependencies.

In a first, a team of physicists is using artificial intelligence (AI) to run a complex experiment to create an extremely cold gas trapped in a laser beam known as a Bose-Einstein condensate -- thus replicating the experiment that won the 2001 Nobel Prize. Bose-Einstein condensates are some of the coldest places in the universe -- far colder than outer space and typically less than a billionth of a degree above absolute zero. They can be used for mineral exploration or navigation systems as they are extremely sensitive to external disturbances, which allows them to make very precise measurements such as tiny changes in the Earth's magnetic field or gravity. Indian physicist Satyendra Nath Bose, along with German-born theoretical physicist Albert Einstein, founded the basis for Bose-Einstein statistics. It describes the statistical distribution of identical particles with integer spin, now called subatomic particle or the "God particle" Boson.

The idea of cyborgs running the world may seem like science fiction but may be becoming reality sooner than you think. No, there will not be an epic world war of us versus them. Instead, the shift from human to automation is slowly creeping into our society. Automation is already running assembly lines and even surgeries. From ATM's, pay-at-the-pump, self check-out, self-serve kiosks, order and pay at the table in restaurants, and all of the banking, shopping, record sharing, reading, socialization that takes place online, computers are already taking the place of humans.

Toyota Motor Corp. announced Tuesday a partnership with Uber Technologies in which the Japanese automaker will invest an undisclosed sum in the ride-hailing company and establish a car-leasing option for Uber drivers. Through the partnership, the two companies will "explore collaboration ... in the world of ridesharing in countries where ridesharing is expanding, taking various factors into account such as regulations, business conditions, and customer needs," Toyota said in a prepared statement. The partnership will also expand on Uber's vehicle-leasing program, which offers cars to its drivers through Enterprise Holdings Inc. -- the parent company of Enterprise Rent-A-Car, Alamo Rent a Car and National Car Rental. Both Uber and Toyota declined to comment on the size of Toyota's investment. A Toyota spokesman said both companies are "still in the early stages of developing [their] strategy" and, aside from the vehicle leasing option, no other specific plans have been made.