There is a growing desire in the field of reinforcement learning (and machine learning in general) to move from black-box models toward more "interpretable AI." We improve interpretability of reinforcement learning by increasing the utility of decision tree policies learned via reinforcement learning. These policies consist of a decision tree over the state space, which requires fewer parameters to express than traditional policy representations. Existing methods for creating decision tree policies via reinforcement learning focus on accurately representing an action-value function during training, but this leads to much larger trees than would otherwise be required. To address this shortcoming, we propose a novel algorithm which only increases tree size when the estimated discounted future reward of the overall policy would increase by a sufficient amount. Through evaluation in a simulated environment, we show that its performance is comparable or superior to traditional tree-based approaches and that it yields a more succinct policy. Additionally, we discuss tuning parameters to control the tradeoff between optimizing for smaller tree size or for overall reward.

Machine learning algorithms designed to characterize, monitor, and intervene on human health (ML4H) are expected to perform safely and reliably when operating at scale, potentially outside strict human supervision. This requirement warrants a stricter attention to issues of reproducibility than other fields of machine learning. In this work, we conduct a systematic evaluation of over 100 recently published ML4H research papers along several dimensions related to reproducibility. We find that the field of ML4H compares poorly to more established machine learning fields, particularly concerning data and code accessibility. Finally, drawing from success in other fields of science, we propose recommendations to data providers, academic publishers, and the ML4H research community in order to promote reproducible research moving forward.

We revisit the notion of individual fairness proposed by Dwork et al. A central challenge in operationalizing their approach is the difficulty in eliciting a human specification of a similarity metric. In this paper, we propose an operationalization of individual fairness that does not rely on a human specification of a distance metric. Instead, we propose novel approaches to elicit and leverage side-information on equally deserving individuals to counter subordination between social groups. We model this knowledge as a fairness graph, and learn a unified Pairwise Fair Representation(PFR) of the data that captures both data-driven similarity between individuals and the pairwise side-information in fairness graph. We elicit fairness judgments from a variety of sources, including humans judgments for two real-world datasets on recidivism prediction (COMPAS) and violent neighborhood prediction (Crime & Communities). Our experiments show that the PFR model for operationalizing individual fairness is practically viable.

Materials discovery is decisive for tackling urgent challenges related to energy, the environment, health care and many others. In chemistry, conventional methodologies for innovation usually rely on expensive and incremental strategies to optimize properties from molecular structures. On the other hand, inverse approaches map properties to structures, thus expediting the design of novel useful compounds. In this chapter, we examine the way in which current deep generative models are addressing the inverse chemical discovery paradigm. We begin by revisiting early inverse design algorithms. Then, we introduce generative models for molecular systems and categorize them according to their architecture and molecular representation. Using this classification, we review the evolution and performance of important molecular generation schemes reported in the literature. Finally, we conclude highlighting the prospects and challenges of generative models as cutting edge tools in materials discovery.

Railway points are among the key components of railway infrastructure. As a part of signal equipment, points control the routes of trains at railway junctions, having a significant impact on the reliability, capacity, and punctuality of rail transport. Traditionally, maintenance of points is based on a fixed time interval or raised after the equipment failures. Instead, it would be of great value if we could forecast points' failures and take action beforehand, minimising any negative effect. To date, most of the existing prediction methods are either lab-based or relying on specially installed sensors which makes them infeasible for large-scale implementation. Besides, they often use data from only one source. We, therefore, explore a new way that integrates multi-source data which are ready to hand to fulfil this task. We conducted our case study based on Sydney Trains rail network which is an extensive network of passenger and freight railways. Unfortunately, the real-world data are usually incomplete due to various reasons, e.g., faults in the database, operational errors or transmission faults. Besides, railway points differ in their locations, types and some other properties, which means it is hard to use a unified model to predict their failures. Aiming at this challenging task, we firstly constructed a dataset from multiple sources and selected key features with the help of domain experts. In this paper, we formulate our prediction task as a multiple kernel learning problem with missing kernels. We present a robust multiple kernel learning algorithm for predicting points failures. Our model takes into account the missing pattern of data as well as the inherent variance on different sets of railway points. Extensive experiments demonstrate the superiority of our algorithm compared with other state-of-the-art methods.

Multi-view learning (MVL) is a strategy for fusing data from different sources or subsets. Canonical correlation analysis (CCA) is very important in MVL, whose main idea is to map data from different views onto a common space with the maximum correlation. The traditional CCA can only be used to calculate the linear correlation between two views. Moreover, it is unsupervised, and the label information is wasted in supervised learning tasks. Many nonlinear, supervised, or generalized extensions have been proposed to overcome these limitations. However, to our knowledge, there is no up-to-date overview of these approaches. This paper fills this gap, by providing a comprehensive overview of many classical and latest CCA approaches, and describing their typical applications in pattern recognition, multi-modal retrieval and classification, and multi-view embedding.

Fuzzy cognitive maps (FCMs) model feedback causal relations in interwoven webs of causality and policy variables. FCMs are fuzzy signed directed graphs that allow degrees of causal influence and event occurrence. Such causal models can simulate a wide range of policy scenarios and decision processes. Their directed loops or cycles directly model causal feedback. Their nonlinear dynamics permit forward-chaining inference from input causes and policy options to output effects. Users can add detailed dynamics and feedback links directly to the causal model or infer them with statistical learning laws. Users can fuse or combine FCMs from multiple experts by weighting and adding the underlying fuzzy edge matrices and do so recursively if needed. The combined FCM tends to better represent domain knowledge as the expert sample size increases if the expert sample approximates a random sample. Many causal models use more restrictive directed acyclic graphs (DAGs) and Bayesian probabilities. DAGs do not model causal feedback because they do not contain closed loops. Combining DAGs also tends to produce cycles and thus tends not to produce a new DAG. Combining DAGs tends to produce a FCM. FCM causal influence is also transitive whereas probabilistic causal influence is not transitive in general. Overall: FCMs trade the numerical precision of probabilistic DAGs for pattern prediction, faster and scalable computation, ease of combination, and richer feedback representation. We show how FCMs can apply to problems of public support for insurgency and terrorism and to US-China conflict relations in Graham Allison's Thucydides-trap framework. The appendix gives the textual justification of the Thucydides-trap FCM. It also extends our earlier theorem [Osoba-Kosko2017] to a more general result that shows the transitive and total causal influence that upstream concept nodes exert on downstream nodes.

We study the application of dynamic pricing to insurance. We view this as an online revenue management problem where the insurance company looks to set prices to optimize the long-run revenue from selling a new insurance product. We develop two pricing models: an adaptive Generalized Linear Model (GLM) and an adaptive Gaussian Process (GP) regression model. Both balance between exploration, where we choose prices in order to learn the distribution of demands & claims for the insurance product, and exploitation, where we myopically choose the best price from the information gathered so far. The performance of the pricing policies is measured in terms of regret: the expected revenue loss caused by not using the optimal price. As is commonplace in insurance, we model demand and claims by GLMs. In our adaptive GLM design, we use the maximum quasi-likelihood estimation (MQLE) to estimate the unknown parameters. We show that, if prices are chosen with suitably decreasing variability, the MQLE parameters eventually exist and converge to the correct values, which in turn implies that the sequence of chosen prices will also converge to the optimal price. In the adaptive GP regression model, we sample demand and claims from Gaussian Processes and then choose selling prices by the upper confidence bound rule. We also analyze these GLM and GP pricing algorithms with delayed claims. Although similar results exist in other domains, this is among the first works to consider dynamic pricing problems in the field of insurance. We also believe this is the first work to consider Gaussian Process regression in the context of insurance pricing. These initial findings suggest that online machine learning algorithms could be a fruitful area of future investigation and application in insurance.

Suffering from the multi-view data diversity and complexity for semi-supervised classification, most of existing graph convolutional networks focus on the networks architecture construction or the salient graph structure preservation, and ignore the the complete graph structure for semi-supervised classification contribution. To mine the more complete distribution structure from multi-view data with the consideration of the specificity and the commonality, we propose structure fusion based on graph convolutional networks (SF-GCN) for improving the performance of semi-supervised classification. SF-GCN can not only retain the special characteristic of each view data by spectral embedding, but also capture the common style of multi-view data by distance metric between multi-graph structures. Suppose the linear relationship between multi-graph structures, we can construct the optimization function of structure fusion model by balancing the specificity loss and the commonality loss. By solving this function, we can simultaneously obtain the fusion spectral embedding from the multi-view data and the fusion structure as adjacent matrix to input graph convolutional networks for semi-supervised classification. Experiments demonstrate that the performance of SF-GCN outperforms that of the state of the arts on three challenging datasets, which are Cora,Citeseer and Pubmed in citation networks.

Articulatory distinctive features, as well as phonetic transcription, play important role in speech-related tasks: computer-assisted pronunciation training, text-to-speech conversion (TTS), studying speech production mechanisms, speech recognition for low-resourced languages. End-to-end approaches to speech-related tasks got a lot of traction in recent years. We apply Listen, Attend and Spell~(LAS)~\cite{Chan-LAS2016} architecture to phones recognition on a small small training set, like TIMIT~\cite{TIMIT-1992}. Also, we introduce a novel decoding technique that allows to train manners and places of articulation detectors end-to-end using attention models. We also explore joint phones recognition and articulatory features detection in multitask learning setting.