The typical pipeline for a supervised machine learning project involves firstly the collection of a significant sample of labelled examples typically referred to as training data. Depending on whether the labels are continuous or categorical, the supervised learning task is known as regression or classification respectively. Next, once the training data is sufficiently clean and complete, it is used to directly build a predictive model using the machine learning algorithm of choice. The predictive model is then used to label new unlabelled examples, and if the labels of the new examples are known a priori by the user (but not used by the learning algorithm) then the predictive accuracy of the model can be evaluated. Different models can therefore be directly compared. In the usual case, the training data is "real", i.e. the model is learned directly from labelled examples that were collected specifically for that purpose. However, quite frequently, modifications are made to the training data after it is collected. For example, it is standard practice to remove outlier examples and normalise numeric values. Moreover, the machine learning algorithm itself may specify modifications to the training data.

This work presents an approach to automatically induction for non-greedy decision trees constructed from neural network architecture. This construction can be used to transfer weights when growing or pruning a decision tree, allowing non-greedy decision tree algorithms to automatically learn and adapt to the ideal architecture. In this work, we examine the underpinning ideas within ensemble modelling and Bayesian model averaging which allow our neural network to asymptotically approach the ideal architecture through weights transfer. Experimental results demonstrate that this approach improves models over fixed set of hyperparameters for decision tree models and decision forest models.

This article is the rejoinder for the paper "Probabilistic Integration: A Role in Statistical Computation?" to appear in Statistical Science with discussion [Briol et al., 2015]. We would first like to thank the reviewers and many of our colleagues who helped shape this paper, the editor for selecting our paper for discussion, and of course all of the discussants for their thoughtful, insightful and constructive comments. In this rejoinder, we respond to some of the points raised by the discussants and comment further on the fundamental questions underlying the paper: - Should Bayesian ideas be used in numerical analysis? Numerical analysis is concerned with the approximation of typically high or infinite-dimensional mathematical quantities using discretisations of the space on which these are defined. Different discretisation schemes lead to different numerical algorithms, whose stability and convergence properties need to be carefully assessed.

Food packaging labels provide important information for public health, such as allergens and use-by dates. Off-the-shelf Optical Character Verification (OCV) systems are good solutions for automating food label quality assessments, but are known to under perform on complex data. This paper proposes a Deep Learning based system that can identify inadequate images for OCV, due to their poor label quality, by employing state-of-the-art Convolutional Neural Network (CNN) architectures, and practical Bayesian inference techniques for automatic self-annotation. We propose a practical domain adaptation procedure based on k-means clustering of CNN latent variables, followed by a k-Nearest Neighbour classification for handling high label variability between different dataset distributions. Moreover, Supervised Learning has proven useful in such systems but manual annotation of large amounts of data is usually required. This is practically intractable in most real world problems due to time/labour constraints. In an attempt to address this issue, we introduce a self-annotating prediction model based on Self-Training of a Bayesian CNN, that leverages modern variational inference methods of deep models. In this context, we propose a new inverse uncertainty weighting technique that encourages the Self-Training model to learn from more informative data over time, potentially preventing it from becoming lazy by only selecting easy examples to learn from. An experimental study is presented illustrating the superior performance of the proposed approach over standard Self-Training, and highlighting the importance of predictive uncertainty estimates in safety-critical domains.

Recent years have witnessed the rising popularity of Natural Language Processing (NLP) and related fields such as Artificial Intelligence (AI) and Machine Learning (ML). Many online courses and resources are available even for those without a strong background in the field. Often the student is curious about a specific topic but does not quite know where to begin studying. To answer the question of "what should one learn first," we apply an embedding-based method to learn prerequisite relations for course concepts in the domain of NLP. We introduce LectureBank, a dataset containing 1,352 English lecture files collected from university courses which are each classified according to an existing taxonomy as well as 208 manually-labeled prerequisite relation topics, which is publicly available. The dataset will be useful for educational purposes such as lecture preparation and organization as well as applications such as reading list generation. Additionally, we experiment with neural graph-based networks and non-neural classifiers to learn these prerequisite relations from our dataset.

In this work, we present our various contributions to the objective of building a decision support tool for the diagnosis of rare diseases. Our goal is to achieve a state of knowledge where the uncertainty about the patient's disease is below a predetermined threshold. We aim to reach such states while minimizing the average number of medical tests to perform. In doing so, we take into account the need, in many medical applications, to avoid, as much as possible, any misdiagnosis. To solve this optimization task, we investigate several reinforcement learning algorithm and make them operable in our high-dimensional and sparse-reward setting. We also present a way to combine expert knowledge, expressed as conditional probabilities, with real clinical data. This is crucial because the scarcity of data in the field of rare diseases prevents any approach based solely on clinical data. Finally we show that it is possible to integrate the ontological information about symptoms while remaining in our probabilistic reasoning. It enables our decision support tool to process information given at different level of precision by the user.

Sequence generation models such as recurrent networks can be trained with a diverse set of learning algorithms. For example, maximum likelihood learning is simple and efficient, yet suffers from the exposure bias problem. Reinforcement learning like policy gradient addresses the problem but can have prohibitively poor exploration efficiency. A variety of other algorithms such as RAML, SPG, and data noising, have also been developed from different perspectives. This paper establishes a formal connection between these algorithms. We present a generalized entropy regularized policy optimization formulation, and show that the apparently divergent algorithms can all be reformulated as special instances of the framework, with the only difference being the configurations of reward function and a couple of hyperparameters. The unified interpretation offers a systematic view of the varying properties of exploration and learning efficiency. Besides, based on the framework, we present a new algorithm that dynamically interpolates among the existing algorithms for improved learning. Experiments on machine translation and text summarization demonstrate the superiority of the proposed algorithm.

Bayesian optimization usually assumes that a Bayesian prior is given. However, the strong theoretical guarantees in Bayesian optimization are often regrettably compromised in practice because of unknown parameters in the prior. In this paper, we adopt a variant of empirical Bayes and show that, by estimating the Gaussian process prior from offline data sampled from the same prior and constructing unbiased estimators of the posterior, variants of both GP-UCB and probability of improvement achieve a near-zero regret bound, which decreases to a constant proportional to the observational noise as the number of offline data and the number of online evaluations increase. Empirically, we have verified our approach on challenging simulated robotic problems featuring task and motion planning.

Parallel tempering addresses some of the drawbacks of canonical Markov Chain Monte-Carlo methods for Bayesian neural learning with the ability to utilize high performance computing. However, certain challenges remain given the large range of network parameters and big data. Surrogate-assisted optimization considers the estimation of an objective function for models given computational inefficiency or difficulty to obtain clear results. We address the inefficiency of parallel tempering for large-scale problems by combining parallel computing features with surrogate assisted estimation of likelihood function that describes the plausibility of a model parameter value, given specific observed data. In this paper, we present surrogate-assisted parallel tempering for Bayesian neural learning where the surrogates are used to estimate the likelihood. The estimation via the surrogate becomes useful rather than evaluating computationally expensive models that feature large number of parameters and datasets. Our results demonstrate that the methodology significantly lowers the computational cost while maintaining quality in decision making using Bayesian neural learning. The method has applications for a Bayesian inversion and uncertainty quantification for a broad range of numerical models.

Likelihood-free inference refers to inference when a likelihood function cannot be explicitly evaluated, which is often the case for models based on simulators. While much of the literature is concerned with sample-based'Approximate Bayesian Computation' methods, recent work suggests that approaches relying on deep neural conditional density estimators can obtain state-of-the-art results with fewer simulations. The neural approaches vary in how they choose which simulations to run and what they learn: an approximate posterior or a surrogate likelihood. This work provides some direct controlled comparisons between these choices.