Generative adversarial learning is a popular new approach to training generative models which has been proven successful for other related problems as well. The general idea is to maintain an oracle $D$ that discriminates between the expert's data distribution and that of the generative model $G$. The generative model is trained to capture the expert's distribution by maximizing the probability of $D$ misclassifying the data it generates. Overall, the system is \emph{differentiable} end-to-end and is trained using basic backpropagation. This type of learning was successfully applied to the problem of policy imitation in a model-free setup. However, a model-free approach does not allow the system to be differentiable, which requires the use of high-variance gradient estimations. In this paper we introduce the Model based Adversarial Imitation Learning (MAIL) algorithm. A model-based approach for the problem of adversarial imitation learning. We show how to use a forward model to make the system fully differentiable, which enables us to train policies using the (stochastic) gradient of $D$. Moreover, our approach requires relatively few environment interactions, and fewer hyper-parameters to tune. We test our method on the MuJoCo physics simulator and report initial results that surpass the current state-of-the-art.

We introduce a novel approach for parallelizing MCMC inference in models with spatially determined conditional independence relationships, for which existing techniques exploiting graphical model structure are not applicable. Our approach is motivated by a model of seismic events and signals, where events detected in distant regions are approximately independent given those in intermediate regions. We perform parallel inference by coloring a factor graph defined over regions of latent space, rather than individual model variables. Evaluating on a model of seismic event detection, we achieve significant speedups over serial MCMC with no degradation in inference quality.

Deep neural networks have had an enormous impact on image analysis. State-of-the-art training methods, based on weight decay and DropOut, result in impressive performance when a very large training set is available. However, they tend to have large problems overfitting to small data sets. Indeed, the available regularization methods deal with the complexity of the network function only indirectly. In this paper, we study the feasibility of directly using the $L_2$ function norm for regularization. Two methods to integrate this new regularization in the stochastic backpropagation are proposed. Moreover, the convergence of these new algorithms is studied. We finally show that they outperform the state-of-the-art methods in the low sample regime on benchmark datasets (MNIST and CIFAR10). The obtained results demonstrate very clear improvement, especially in the context of small sample regimes with data laying in a low dimensional manifold. Source code of the method can be found at \url{https://github.com/AmalRT/DNN_Reg}.

Often in real-world datasets, especially in high dimensional data, some feature values are missing. Since most data analysis and statistical methods do not handle gracefully missing values, the first step in the analysis requires the imputation of missing values. Indeed, there has been a long standing interest in methods for the imputation of missing values as a pre-processing step. One recent and effective approach, the IRMI stepwise regression imputation method, uses a linear regression model for each real-valued feature on the basis of all other features in the dataset. However, the proposed iterative formulation lacks convergence guarantee. Here we propose a closely related method, stated as a single optimization problem and a block coordinate-descent solution which is guaranteed to converge to a local minimum. Experiments show results on both synthetic and benchmark datasets, which are comparable to the results of the IRMI method whenever it converges. However, while in the set of experiments described here IRMI often does not converge, the performance of our methods is shown to be markedly superior in comparison with other methods.

Automatic image annotation has been an important research topic in facilitating large scale image management and retrieval. Existing methods focus on learning image-tag correlation or correlation between tags to improve annotation accuracy. However, most of these methods evaluate their performance using top-k retrieval performance, where k is fixed. Although such setting gives convenience for comparing different methods, it is not the natural way that humans annotate images. The number of annotated tags should depend on image contents. Inspired by the recent progress in machine translation and image captioning, we propose a novel Recurrent Image Annotator (RIA) model that forms image annotation task as a sequence generation problem so that RIA can natively predict the proper length of tags according to image contents. We evaluate the proposed model on various image annotation datasets. In addition to comparing our model with existing methods using the conventional top-k evaluation measures, we also provide our model as a high quality baseline for the arbitrary length image tagging task. Moreover, the results of our experiments show that the order of tags in training phase has a great impact on the final annotation performance.

This course aimed at students with beginner to intermediate skill in F#, basic understanding of the F# syntax and a light functional understanding would be beneficial. You'll also need a computer with Linux, OSX or Windows with F# installed and an internet connection. Have you wanted to understand how to'do' machine learning or implement algorithms from a textbook in a programming language, or deploy a library to Nuget? Well, this course includes sections on machine learning using a mathematical theorem known as Bayes' Theorem. We will start by creating a predictive text engine and deploy it to Nuget, while learning how to write some basic unit tests in FsUnit.

BERLIN – Scientists have developed a mind-controlled robotic hand that allows people with certain types of spinal injuries to perform everyday tasks such as using a fork or drinking from a cup. The low-cost device was tested in Spain on six people with quadriplegia affecting their ability to grasp or manipulate objects. By wearing a cap that measures electric brain activity and eye movement the users were able to send signals to a tablet computer that controlled the glove-like device attached to their hand. Participants in the small-scale study were able to perform daily activities better with the robotic hand than without, according to results published Tuesday in the journal Science Robotics. The principle of using brain-controlled robotic aids to assist people with quadriplegia isn't new.

Beating the AI opponents in a strategy game is the first step any gamer takes before heading online where the real challenge is. Whether they cheat or not, most game AIs are beatable, often with simple, repeatable strategies once you find their weak points. Named HIRO (Human Intelligence Robotically Optimized), the algorithm can beat almost all players -- and that's not even its main job. Arago is an IT automation firm, which develops smart AIs that can streamline businesses and automate many of their functions. HIRO is one such AI, and while it does an excellent job of improving workflows at a number of corporations, it's the way it's trained that is most fascinating. It plays -- very well at that -- the freely available civilization-building game called Freeciv.

I was recently interviewed at the Robotic Process Automation and Artificual Intelligence in Summit in December 2016 in London. Here's a discussion on how organizations worldwide are building their Digital Workforce. Organizations worldwide are building their Digital Workforce. The countdown has begun and we foresee 3 Million Digital Workers by 2020. We believe that Human workers, alongside the Digital bots, creates a hyper productive workforce and building a Digital Workforce is about talent augmentation, not talent replacement.

Rob May is the CEO and Co-Founder of Talla, a platform for intelligent information delivery in Slack and Hipchat. Previously, Rob was the CEO and Co-Founder of Backupify, (acquired by Datto in 2014). Before that, he held engineering, business development, and management positions at various startups. Rob has a B.S. in Electrical Engineering and a MBA from the University of Kentucky. He is also a well-known angel investor in the AI space and is the creator and writer of the widely-read and highly-regarded AI newsletter, Technically Sentient. Rob May will be speaking at the Gigaom AI Now in San Francisco, February 15-16th. In anticipation of that, I caught up with him to ask a few questions.