How will Practo cope up with network connectivity when Internet penetration in tier 3 cities is not high? How can telecom operators play a role in aiding remote healthcare? Today, it is true that mobile broadband has not been fully covered in our country, but we believe that India will get there soon. Today, we have around 30 percent of our traffic coming from tier 2 and 3 towns, so we can say that the Internet has reached some parts of these cities. From the B2B side, our product Practo Tab runs our entire Practo Ray software in offline mode; hence doctors and other healthcare providers can use it in this mode and then synchronize it back with the cloud once they get connectivity.

Awesome post, Ray! It's always encouraging to learn that one is not alone in thinking there's a Tsumani of change ahead. And that we need to try and tame the beast, or the beast will tame us. I was yesterday at a (fascinating) breakfast seminar with a very small group of business leaders here in Argentina. One of the founders of the Singularity University was leading the debate. And one of the topics we discussed at large was the consequence of AI automating jobs… everywhere, anywhere.

I don't know about you, but I haven't seen any dinosaurs lately. I mean, I've seen lots of birds, some lizards, and the occasional crocodile, but none of those massive Jurassic Park-style dinos. For paleontologists who want to know how a 60- to 70-ton dinosaur got around, this lack of subjects to study is a bit of an obstacle. At Drexel University, researchers are 3D printing small scale robotic models of the legs of one of the largest dinosaurs ever found to figure out how it was able to keep itself moving. Fossils of Dreadnoughtus schrani were discovered in Argentina in 2005.

Most commuters who travel by road will know the frustration of being caught in traffic jams that can double and even triple the journey to work. But a group of scientists claims to have found a way to ease congestion during the busiest periods, and cut commuting times by a third. However, not everyone will be happy with their solution as it involves some drivers agreeing to endure longer journeys. Scientists analysed billions of journeys made in five cities around the world during morning rush hours record on mobile phones. They found when drivers made selfish, uncoordinated choices, they made congestion worse (stock picture).

Digital expertise will become increasingly critical in the insurance sector as digitization and machine learning leads to more highly'automatable' insurance according to management consultants McKinsey & Company. Meanwhile a separate piece of research by Accenture found that insurance companies are accelerating the shift to a radically different distribution model, where they say digital will play an increasingly important role in most interactions, and were agents' efforts are being refocused to add more value. And analysis by research outfit Ovum suggests strong investment in digital channels also. According to Ovum, " When it comes to investment, digital channels remains the top area for insurers. However, the significant majority of insurers will be increasing budgets across a broad range of functional areas with no single activity completely dominating spend. This reflects the complex set of priorities that IT groups are being asked to meet by the wider business, simultaneously addressing revenue growth, operational efficiency and regulatory compliance."

We define Recurrent Gaussian Processes (RGP) models, a general family of Bayesian nonparametric models with recurrent GP priors which are able to learn dynamical patterns from sequential data. Similar to Recurrent Neural Networks (RNNs), RGPs can have different formulations for their internal states, distinct inference methods and be extended with deep structures. In such context, we propose a novel deep RGP model whose autoregressive states are latent, thereby performing representation and dynamical learning simultaneously. To fully exploit the Bayesian nature of the RGP model we develop the Recurrent Variational Bayes (REVARB) framework, which enables efficient inference and strong regularization through coherent propagation of uncertainty across the RGP layers and states. We also introduce a RGP extension where variational parameters are greatly reduced by being reparametrized through RNN-based sequential recognition models. We apply our model to the tasks of nonlinear system identification and human motion modeling. The promising obtained results indicate that our RGP model maintains its highly flexibility while being able to avoid overfitting and being applicable even when larger datasets are not available.

Vector autoregression (VAR) is a fundamental tool for modeling the joint dynamics of multivariate time series. However, as the number of component series is increased, the VAR model quickly becomes overparameterized, making reliable estimation difficult and impeding its adoption as a forecasting tool in high dimensional settings. A number of authors have sought to address this issue by incorporating regularized approaches, such as the lasso, that impose sparse or low-rank structures on the estimated coefficient parameters of the VAR. More traditional approaches attempt to address overparameterization by selecting a low lag order, based on the assumption that dynamic dependence among components is short-range. However, these methods typically assume a single, universal lag order that applies across all components, unnecessarily constraining the dynamic relationship between the components and impeding forecast performance. The lasso-based approaches are more flexible but do not incorporate the notion of lag order selection. We propose a new class of regularized VAR models, called hierarchical vector autoregression (HVAR), that embed the notion of lag selection into a convex regularizer. The key convex modeling tool is a group lasso with nested groups which ensure the sparsity pattern of autoregressive lag coefficients honors the ordered structure inherent to VAR. We provide computationally efficient algorithms for solving HVAR problems that can be parallelized across the components. A simulation study shows the improved performance in forecasting and lag order selection over previous approaches, and a macroeconomic application further highlights forecasting improvements as well as the convenient, interpretable output of a HVAR model.

It is well known that speaker verification systems are subject to spoofing attacks. The Automatic Speaker Verification Spoofing and Countermeasures Challenge -- ASVSpoof2015 -- provides a standard spoofing database, containing attacks based on synthetic speech, along with a protocol for experiments. This paper describes CPqD's systems submitted to the ASVSpoof2015 Challenge, based on deep neural networks, working both as a classifier and as a feature extraction module for a GMM and a SVM classifier. Results show the validity of this approach, achieving less than 0.5\% EER for known attacks.

We introduce the Gaussian Process Convolution Model (GPCM), a two-stage nonparametric generative procedure to model stationary signals as the convolution between a continuous-time white-noise process and a continuous-time linear filter drawn from Gaussian process. The GPCM is a continuous-time nonparametric-window moving average process and, conditionally, is itself a Gaussian process with a nonparametric kernel defined in a probabilistic fashion. The generative model can be equivalently considered in the frequency domain, where the power spectral density of the signal is specified using a Gaussian process. One of the main contributions of the paper is to develop a novel variational free-energy approach based on inter-domain inducing variables that efficiently learns the continuous-time linear filter and infers the driving white-noise process. In turn, this scheme provides closed-form probabilistic estimates of the covariance kernel and the noise-free signal both in denoising and prediction scenarios. Additionally, the variational inference procedure provides closed-form expressions for the approximate posterior of the spectral density given the observed data, leading to new Bayesian nonparametric approaches to spectrum estimation. The proposed GPCM is validated using synthetic and real-world signals.

Gibbs sampling on factor graphs is a widely used inference technique, which often produces good empirical results. Theoretical guarantees for its performance are weak: even for tree structured graphs, the mixing time of Gibbs may be exponential in the number of variables. To help understand the behavior of Gibbs sampling, we introduce a new (hyper)graph property, called hierarchy width. We show that under suitable conditions on the weights, bounded hierarchy width ensures polynomial mixing time. Our study of hierarchy width is in part motivated by a class of factor graph templates, hierarchical templates, which have bounded hierarchy width—regardless of the data used to instantiate them. We demonstrate a rich application from natural language processing in which Gibbs sampling provably mixes rapidly and achieves accuracy that exceeds human volunteers.