This paper presents a novel multitask multiple kernel learning framework that efficiently learns the kernel weights leveraging the relationship across multiple tasks. The idea is to automatically infer this task relationship in the \textit{RKHS} space corresponding to the given base kernels. The problem is formulated as a regularization-based approach called \textit{Multi-Task Multiple Kernel Relationship Learning} (\textit{MK-MTRL}), which models the task relationship matrix from the weights learned from latent feature spaces of task-specific base kernels. Unlike in previous work, the proposed formulation allows one to incorporate prior knowledge for simultaneously learning several related tasks. We propose an alternating minimization algorithm to learn the model parameters, kernel weights and task relationship matrix. In order to tackle large-scale problems, we further propose a two-stage \textit{MK-MTRL} online learning algorithm and show that it significantly reduces the computational time, and also achieves performance comparable to that of the joint learning framework. Experimental results on benchmark datasets show that the proposed formulations outperform several state-of-the-art multitask learning methods.

Neural networks are a powerful class of functions that can be trained with simple gradient descent to achieve state-of-the-art performance on a variety of applications. Despite their practical success, there is a paucity of results that provide theoretical guarantees on why they are so effective. Lying in the center of the problem is the difficulty of analyzing the non-convex loss function with potentially numerous local minima and saddle points. Can neural networks corresponding to the stationary points of the loss function learn the true target function? If yes, what are the key factors contributing to such nice optimization properties? In this paper, we answer these questions by analyzing one-hidden-layer neural networks with ReLU activation, and show that despite the non-convexity, neural networks with diverse units have no spurious local minima. We bypass the non-convexity issue by directly analyzing the first order optimality condition, and show that the loss can be made arbitrarily small if the minimum singular value of the "extended feature matrix" is large enough. We make novel use of techniques from kernel methods and geometric discrepancy, and identify a new relation linking the smallest singular value to the spectrum of a kernel function associated with the activation function and to the diversity of the units. Our results also suggest a novel regularization function to promote unit diversity for potentially better generalization.

We consider multi-class classification where the predictor has a hierarchical structure that allows for a very large number of labels both at train and test time. The predictive power of such models can heavily depend on the structure of the tree, and although past work showed how to learn the tree structure, it expected that the feature vectors remained static. We provide a novel algorithm to simultaneously perform representation learning for the input data and learning of the hierarchi- cal predictor. Our approach optimizes an objec- tive function which favors balanced and easily- separable multi-way node partitions. We theoret- ically analyze this objective, showing that it gives rise to a boosting style property and a bound on classification error. We next show how to extend the algorithm to conditional density estimation. We empirically validate both variants of the al- gorithm on text classification and language mod- eling, respectively, and show that they compare favorably to common baselines in terms of accu- racy and running time.

In an end-to-end dialog system, the aim of dialog state tracking is to accurately estimate a compact representation of the current dialog status from a sequence of noisy observations produced by the speech recognition and the natural language understanding modules. This paper introduces a novel method of dialog state tracking based on the general paradigm of machine reading and proposes to solve it using an End-to-End Memory Network, MemN2N, a memory-enhanced neural network architecture. We evaluate the proposed approach on the second Dialog State Tracking Challenge (DSTC-2) dataset. The corpus has been converted for the occasion in order to frame the hidden state variable inference as a question-answering task based on a sequence of utterances extracted from a dialog. We show that the proposed tracker gives encouraging results. Then, we propose to extend the DSTC-2 dataset with specific reasoning capabilities requirement like counting, list maintenance, yes-no question answering and indefinite knowledge management. Finally, we present encouraging results using our proposed MemN2N based tracking model.

Many generative models can be expressed as a differentiable function of random inputs drawn from some simple probability density. This framework includes both deep generative architectures such as Variational Autoencoders and a large class of procedurally defined simulator models. We present a method for performing efficient MCMC inference in such models when conditioning on observations of the model output. For some models this offers an asymptotically exact inference method where Approximate Bayesian Computation might otherwise be employed. We use the intuition that inference corresponds to integrating a density across the manifold corresponding to the set of inputs consistent with the observed outputs. This motivates the use of a constrained variant of Hamiltonian Monte Carlo which leverages the smooth geometry of the manifold to coherently move between inputs exactly consistent with observations.

Extracting automatically the complex set of features composing real high-dimensional data is crucial for achieving high performance in machine-learning tasks. Restricted Boltzmann Machines (RBM) are empirically known to be efficient for this purpose, and to be able to generate distributed and graded representations of the data. We characterize the structural conditions (sparsity of the weights, low effective temperature, nonlinearities in the activation functions of hidden units, and adaptation of fields maintaining the activity in the visible layer) allowing RBM to operate in such a compositional phase. Evidence is provided by the replica analysis of an adequate statistical ensemble of random RBMs and by RBM trained on the handwritten digits dataset MNIST. Recent years have witnessed major progresses in supervised machine learning, e.g. in video, audio, image processing,... [1].

We introduce a method for using deep neural networks to amortize the cost of inference in models from the family induced by universal probabilistic programming languages, establishing a framework that combines the strengths of probabilistic programming and deep learning methods. We call what we do "compilation of inference" because our method transforms a denotational specification of an inference problem in the form of a probabilistic program written in a universal programming language into a trained neural network denoted in a neural network specification language. When at test time this neural network is fed observational data and executed, it performs approximate inference in the original model specified by the probabilistic program. Our training objective and learning procedure are designed to allow the trained neural network to be used as a proposal distribution in a sequential importance sampling inference engine. We illustrate our method on mixture models and Captcha solving and show significant speedups in the efficiency of inference.

March 1, 2017 --Is there – or was there ever – life on Mars? NASA has spent decades investigating the question with orbiters and rovers, including its upcoming Mars 2020 rover, but at least one scientist suspects he already knows the answer. According to Gibert Levin, NASA probably detected microbial life on Mars in 1976. Dr. Levin was one of the scientists involved with the Viking lander, whose biological experiments gave conflicting results when samples tested positive for metabolism but negative for organic molecules. The unreliable organic molecule experiment was the one that failed, and the metabolism detection succeeded.

What if instead of searching through menus within programs like Microsoft Excel, our computers could understand the problem we're trying to solve and write the software to solve it? It's a hyper-futuristic idea, but one that has recently seen progress from Microsoft Research and the University of Cambridge. In a November 2016 paper (pdf), which gained notoriety after being accepted into one of the year's largest artificial intelligence conferences, Microsoft and Cambridge built an algorithm capable of writing code that would solve simple math problems. The algorithm, named DeepCoder, would be able to augment its own ability by also looking at potential combinations of code for how a problem could be solved. However, this doesn't mean it steals code, or copy and pastes it from existing software, or searches the internet for solutions, as some reports have claimed. "We're targeting the people who can't or don't want to code, but can specify what their problem is," says Marc Brockschmidt of Microsoft Research, a co-author of the paper, likening the work to Excel formulas that could take simple commands to solve for answers without being given the mathematical equation.

It was one of those amazing "we're living in the future" moments. In an October 2013 press release, IBM declared that MD Anderson, the cancer center that is part of the University of Texas, "is using the IBM Watson cognitive computing system for its mission to eradicate cancer." Well, now that future is past. The partnership between IBM and one of the world's top cancer research institutions is falling apart. The project is on hold, MD Anderson confirms, and has been since late last year.