Extremile regression, as a least squares analog of quantile regression, is potentially useful tool for modeling and understanding the extreme tails of a distribution. However, existing extremile regression methods, as nonparametric approaches, may face challenges in high-dimensional settings due to data sparsity, computational inefficiency, and the risk of overfitting. While linear regression serves as the foundation for many other statistical and machine learning models due to its simplicity, interpretability, and relatively easy implementation, particularly in high-dimensional settings, this paper introduces a novel definition of linear extremile regression along with an accompanying estimation methodology. The regression coefficient estimators of this method achieve $\sqrt{n}$-consistency, which nonparametric extremile regression may not provide. In particular, while semi-supervised learning can leverage unlabeled data to make more accurate predictions and avoid overfitting to small labeled datasets in high-dimensional spaces, we propose a semi-supervised learning approach to enhance estimation efficiency, even when the specified linear extremile regression model may be misspecified. Both simulation studies and real data analyses demonstrate the finite-sample performance of our proposed methods.

Scoring functions constructed by transforming the realization and prediction variables of (strictly) consistent scoring functions have been widely studied empirically, yet their theoretical foundations remain unexplored. To address this gap, we establish formal characterizations of (strict) consistency for these transformed scoring functions and their elicitable functionals. Our analysis focuses on two interrelated cases: (a) transformations applied exclusively to the realization variable, and (b) bijective transformations applied jointly to both realization and prediction variables. We formulate analogous characterizations for (strict) identification functions. The resulting theoretical framework is broadly applicable to statistical and machine learning methodologies. When applied to Bregman and expectile scoring functions, our framework shows how it enables two critical advances: (a) rigorous interpretation of prior empirical findings from models trained with transformed scoring functions, and (b) systematic construction of novel identifiable and elicitable functionals, specifically the g-transformed expectation and g-transformed expectile. By unifying theoretical insights with practical applications, this work advances principled methodologies for designing scoring functions in complex predictive tasks.

Many classic Reinforcement Learning (RL) algorithms rely on a Bellman operator, which involves an expectation over the next states, leading to the concept of bootstrapping. To introduce a form of pessimism, we propose to replace this expectation with an expectile. In practice, this can be very simply done by replacing the $L_2$ loss with a more general expectile loss for the critic. Introducing pessimism in RL is desirable for various reasons, such as tackling the overestimation problem (for which classic solutions are double Q-learning or the twin-critic approach of TD3) or robust RL (where transitions are adversarial). We study empirically these two cases. For the overestimation problem, we show that the proposed approach, ExpectRL, provides better results than a classic twin-critic. On robust RL benchmarks, involving changes of the environment, we show that our approach is more robust than classic RL algorithms. We also introduce a variation of ExpectRL combined with domain randomization which is competitive with state-of-the-art robust RL agents. Eventually, we also extend \ExpectRL with a mechanism for choosing automatically the expectile value, that is the degree of pessimism

Typical machine learning regression applications aim to report the mean or the median of the predictive probability distribution, via training with a squared or an absolute error scoring function. The importance of issuing predictions of more functionals of the predictive probability distribution (quantiles and expectiles) has been recognized as a means to quantify the uncertainty of the prediction. In deep learning (DL) applications, that is possible through quantile and expectile regression neural networks (QRNN and ERNN respectively). Here we introduce deep Huber quantile regression networks (DHQRN) that nest QRNNs and ERNNs as edge cases. DHQRN can predict Huber quantiles, which are more general functionals in the sense that they nest quantiles and expectiles as limiting cases. The main idea is to train a deep learning algorithm with the Huber quantile regression function, which is consistent for the Huber quantile functional. As a proof of concept, DHQRN are applied to predict house prices in Australia. In this context, predictive performances of three DL architectures are discussed along with evidential interpretation of results from an economic case study.

Successful applications of distributional reinforcement learning with quantile regression prompt a natural question: can we use other statistics to represent the distribution of returns? In particular, expectile regression is known to be more efficient than quantile regression for approximating distributions, especially on extreme values, and by providing a straightforward estimator of the mean it is a natural candidate for reinforcement learning. Prior work has answered this question positively in the case of expectiles, with the major caveat that expensive computations must be performed to ensure convergence. In this work, we propose a dual expectile-quantile approach which solves the shortcomings of previous work while leveraging the complementary properties of expectiles and quantiles. Our method outperforms both quantile-based and expectile-based baselines on the MuJoCo continuous control benchmark.

Abstract: his paper features expectiles in dynamic and stochastic optimization. Expectiles are a family of risk functionals characterized as minimizers of optimization problems. For this reason, they enjoy various unique stability properties, which can be exploited in risk averse management, in stochastic optimization and in optimal control. The paper provides tight relates of expectiles to other risk functionals and addresses their properties in regression. Further, we extend expectiles to a dynamic framework.

In decision-making problems such as the multi-armed bandit, an agent learns sequentially by optimizing a certain feedback. While the mean reward criterion has been extensively studied, other measures that reflect an aversion to adverse outcomes, such as mean-variance or conditional value-at-risk (CVaR), can be of interest for critical applications (healthcare, agriculture). Algorithms have been proposed for such risk-aware measures under bandit feedback without contextual information. In this work, we study contextual bandits where such risk measures can be elicited as linear functions of the contexts through the minimization of a convex loss. A typical example that fits within this framework is the expectile measure, which is obtained as the solution of an asymmetric least-square problem. Using the method of mixtures for supermartingales, we derive confidence sequences for the estimation of such risk measures. We then propose an optimistic UCB algorithm to learn optimal risk-aware actions, with regret guarantees similar to those of generalized linear bandits. This approach requires solving a convex problem at each round of the algorithm, which we can relax by allowing only approximated solution obtained by online gradient descent, at the cost of slightly higher regret. We conclude by evaluating the resulting algorithms on numerical experiments.

The identification of influential observations is an important part of data analysis that can prevent erroneous conclusions drawn from biased estimators. However, in high dimensional data, this identification is challenging. Classical and recently-developed methods often perform poorly when there are multiple influential observations in the same dataset. In particular, current methods can fail when there is masking several influential observations with similar characteristics, or swamping when the influential observations are near the boundary of the space spanned by well-behaved observations. Therefore, we propose an algorithm-based, multi-step, multiple detection procedure to identify influential observations that addresses current limitations. Our three-step algorithm to identify and capture undesirable variability in the data, $\asymMIP,$ is based on two complementary statistics, inspired by asymmetric correlations, and built on expectiles. Simulations demonstrate higher detection power than competing methods. Use of the resulting asymptotic distribution leads to detection of influential observations without the need for computationally demanding procedures such as the bootstrap. The application of our method to the Autism Brain Imaging Data Exchange neuroimaging dataset resulted in a more balanced and accurate prediction of brain maturity based on cortical thickness. See our GitHub for a free R package that implements our algorithm: \texttt{asymMIP} (\url{github.com/AmBarry/hidetify}).

Bayesian optimisation is widely used to optimise stochastic black box functions. While most strategies are focused on optimising conditional expectations, a large variety of applications require risk-averse decisions and alternative criteria accounting for the distribution tails need to be considered. In this paper, we propose new variational models for Bayesian quantile and expectile regression that are well-suited for heteroscedastic settings. Our models consist of two latent Gaussian processes accounting respectively for the conditional quantile (or expectile) and variance that are chained through asymmetric likelihood functions. Furthermore, we propose two Bayesian optimisation strategies, either derived from a GP-UCB or Thompson sampling, that are tailored to such models and that can accommodate large batches of points. As illustrated in the experimental section, the proposed approach clearly outperforms the state of the art.

We present a unifying framework for designing and analysing distributional reinforcement learning (DRL) algorithms in terms of recursively estimating statistics of the return distribution. Our key insight is that DRL algorithms can be decomposed as the combination of some statistical estimator and a method for imputing a return distribution consistent with that set of statistics. With this new understanding, we are able to provide improved analyses of existing DRL algorithms as well as construct a new algorithm (EDRL) based upon estimation of the expectiles of the return distribution. We compare EDRL with existing methods on a variety of MDPs to illustrate concrete aspects of our analysis, and develop a deep RL variant of the algorithm, ER-DQN, which we evaluate on the Atari-57 suite of games.