Predictive modeling in healthcare continues to be an active actuarial research topic as more insurance companies aim to maximize the potential of Machine Learning approaches to increase their productivity and efficiency. In this paper, the authors deployed three regression-based ensemble ML models that combine variations of decision trees through Extreme Gradient Boosting, Gradient-boosting Machine, and Random Forest) methods in predicting medical insurance costs. Explainable Artificial Intelligence methods SHapley Additive exPlanations and Individual Conditional Expectation plots were deployed to discover and explain the key determinant factors that influence medical insurance premium prices in the dataset. The dataset used comprised 986 records and is publicly available in the KAGGLE repository. The models were evaluated using four performance evaluation metrics, including R-squared, Mean Absolute Error, Root Mean Squared Error, and Mean Absolute Percentage Error. The results show that all models produced impressive outcomes; however, the XGBoost model achieved a better overall performance although it also expanded more computational resources, while the RF model recorded a lesser prediction error and consumed far fewer computing resources than the XGBoost model. Furthermore, we compared the outcome of both XAi methods in identifying the key determinant features that influenced the PremiumPrices for each model and whereas both XAi methods produced similar outcomes, we found that the ICE plots showed in more detail the interactions between each variable than the SHAP analysis which seemed to be more high-level. It is the aim of the authors that the contributions of this study will help policymakers, insurers, and potential medical insurance buyers in their decision-making process for selecting the right policies that meet their specific needs.

Although general question answering has been well explored in recent years, temporal question answering is a task which has not received as much focus. Our work aims to leverage a popular approach used for general question answering, answer extraction, in order to find answers to temporal questions within a paragraph. To train our model, we propose a new dataset, inspired by SQuAD, specifically tailored to provide rich temporal information. We chose to adapt the corpus WikiWars, which contains several documents on history's greatest conflicts. Our evaluation shows that a deep learning model trained to perform pattern matching, often used in general question answering, can be adapted to temporal question answering, if we accept to ask questions whose answers must be directly present within a text.

This paper explores the potential of the transformer models for learning Granger causality in networks with complex nonlinear dynamics at every node, as in neurobiological and biophysical networks. Our study primarily focuses on a proof-of-concept investigation based on simulated neural dynamics, for which the ground-truth causality is known through the underlying connectivity matrix. For transformer models trained to forecast neuronal population dynamics, we show that the cross attention module effectively captures the causal relationship among neurons, with an accuracy equal or superior to that for the most popular Granger causality analysis method. While we acknowledge that real-world neurobiology data will bring further challenges, including dynamic connectivity and unobserved variability, this research offers an encouraging preliminary glimpse into the utility of the transformer model for causal representation learning in neuroscience.

Interest-free promotions are a prevalent strategy employed by credit card lenders to attract new customers, yet the research exploring their effects on both consumers and lenders remains relatively sparse. The process of selecting an optimal promotion strategy is intricate, involving the determination of an interest-free period duration and promotion-availability window, all within the context of competing offers, fluctuating market dynamics, and complex consumer behaviour. In this paper, we introduce an agent-based model that facilitates the exploration of various credit card promotions under diverse market scenarios. Our approach, distinct from previous agent-based models, concentrates on optimising promotion strategies and is calibrated using benchmarks from the UK credit card market from 2019 to 2020, with agent properties derived from historical distributions of the UK population from roughly the same period. We validate our model against stylised facts and time-series data, thereby demonstrating the value of this technique for investigating pricing strategies and understanding credit card customer behaviour. Our experiments reveal that, in the absence of competitor promotions, lender profit is maximised by an interest-free duration of approximately 12 months while market share is maximised by offering the longest duration possible. When competitors do not offer promotions, extended promotion availability windows yield maximum profit for lenders while also maximising market share. In the context of concurrent interest-free promotions, we identify that the optimal lender strategy entails offering a more competitive interest-free period and a rapid response to competing promotional offers. Notably, a delay of three months in responding to a rival promotion corresponds to a 2.4% relative decline in income.

Personalized adaptive interventions offer the opportunity to increase patient benefits, however, there are challenges in their planning and implementation. Once implemented, it is an important question whether personalized adaptive interventions are indeed clinically more effective compared to a fixed gold standard intervention. In this paper, we present an innovative N-of-1 trial study design testing whether implementing a personalized intervention by an online reinforcement learning agent is feasible and effective. Throughout, we use a new study on physical exercise recommendations to reduce pain in endometriosis for illustration. We describe the design of a contextual bandit recommendation agent and evaluate the agent in simulation studies. The results show that, first, implementing a personalized intervention by an online reinforcement learning agent is feasible. Second, such adaptive interventions have the potential to improve patients' benefits even if only few observations are available. As one challenge, they add complexity to the design and implementation process. In order to quantify the expected benefit, data from previous interventional studies is required. We expect our approach to be transferable to other interventions and clinical interventions.

The exploration of the lunar poles and the collection of samples from the martian surface are characterized by shorter time windows demanding increased autonomy and speeds. Autonomous mobile robots must intrinsically cope with a wider range of disturbances. Faster off-road navigation has been explored for terrestrial applications but the combined effects of increased speeds and reduced gravity fields are yet to be fully studied. In this paper, we design and demonstrate a novel fully passive suspension design for wheeled planetary robots, which couples for the first time a high-range passive rocker with elastic in-wheel coil-over shock absorbers. The design was initially conceived and verified in a reduced-gravity (1.625 m/s${^2}$) simulated environment, where three different passive suspension configurations were evaluated against steep slopes and unexpected obstacles, and later prototyped and validated in a series of field tests. The proposed mechanically-hybrid suspension proves to mitigate more effectively the negative effects (high-frequency/high-amplitude vibrations and impact loads) of faster locomotion (~1\,m/s) over unstructured terrains under varied gravity fields.

K-means plays a vital role in data mining and is the simplest and most widely used algorithm under the Euclidean Minimum Sum-of-Squares Clustering (MSSC) model. However, its performance drastically drops when applied to vast amounts of data. Therefore, it is crucial to improve K-means by scaling it to big data using as few of the following computational resources as possible: data, time, and algorithmic ingredients. We propose a new parallel scheme of using K-means and K-means++ algorithms for big data clustering that satisfies the properties of a ``true big data'' algorithm and outperforms the classical and recent state-of-the-art MSSC approaches in terms of solution quality and runtime. The new approach naturally implements global search by decomposing the MSSC problem without using additional metaheuristics. This work shows that data decomposition is the basic approach to solve the big data clustering problem. The empirical success of the new algorithm allowed us to challenge the common belief that more data is required to obtain a good clustering solution. Moreover, the present work questions the established trend that more sophisticated hybrid approaches and algorithms are required to obtain a better clustering solution.

Networks are ubiquitous in many real-world applications (e.g., social networks encoding trust/distrust relationships, correlation networks arising from time series data). While many networks are signed or directed, or both, there is a lack of unified software packages on graph neural networks (GNNs) specially designed for signed and directed networks. In this paper, we present PyTorch Geometric Signed Directed (PyGSD), a software package which fills this gap. Along the way, we evaluate the implemented methods with experiments with a view to providing insights into which method to choose for a given task. The deep learning framework consists of easy-to-use GNN models, synthetic and real-world data, as well as task-specific evaluation metrics and loss functions for signed and directed networks. As an extension library for PyG, our proposed software is maintained with open-source releases, detailed documentation, continuous integration, unit tests and code coverage checks.

Given $n$ observations from two balanced classes, consider the task of labeling an additional $m$ inputs that are known to all belong to \emph{one} of the two classes. Special cases of this problem are well-known: with complete knowledge of class distributions ($n=\infty$) the problem is solved optimally by the likelihood-ratio test; when $m=1$ it corresponds to binary classification; and when $m\approx n$ it is equivalent to two-sample testing. The intermediate settings occur in the field of likelihood-free inference, where labeled samples are obtained by running forward simulations and the unlabeled sample is collected experimentally. In recent work it was discovered that there is a fundamental trade-off between $m$ and $n$: increasing the data sample $m$ reduces the amount $n$ of training/simulation data needed. In this work we (a) introduce a generalization where unlabeled samples come from a mixture of the two classes -- a case often encountered in practice; (b) study the minimax sample complexity for non-parametric classes of densities under \textit{maximum mean discrepancy} (MMD) separation; and (c) investigate the empirical performance of kernels parameterized by neural networks on two tasks: detection of the Higgs boson and detection of planted DDPM generated images amidst CIFAR-10 images. For both problems we confirm the existence of the theoretically predicted asymmetric $m$ vs $n$ trade-off.

What exactly tipped the Brexit referendum in the favour of the Leave campaign on voting day – June 23, 2016 – may never been clear cut. But a new study suggests the intelligence of voters – and their susceptibility to being duped by misinformation – played a part. From a sample of Brits, researchers looked at how exactly they voted, as well as their intelligence, as determined by performance in cognitive tasks. They found a link between high performance in the tasks and voting Remain, which suggests people who wanted to stay in the EU have higher intelligence. However, the study method only looked at data from just over 6,000 referendum voters – a far cry from the 33.5 million who cast their ballot that day.