Nikolaos Nikolaou

Mario Morvan, Angelos Tsiaras, Nikolaos Nikolaou et al.

We present a new open source python package, based on PyLightcurve and PyTorch, tailored for efficient computation and automatic differentiation of exoplanetary transits. The classes and functions implemented are fully vectorised, natively GPU-compatible and differentiable with respect to the stellar and planetary parameters. This makes PyLightcurve-torch suitable for traditional forward computation of transits, but also extends the range of possible applications with inference and optimisation algorithms requiring access to the gradients of the physical model. This endeavour is aimed at fostering the use of deep learning in exoplanets research, motivated by an ever increasing amount of stellar light curves data and various incentives for the improvement of detection and characterisation techniques.

Tara P. A. Tahseen, Luís F. Simões, Kai Hou Yip et al.

Increasing resolution and coverage of astrophysical and climate data necessitates increasingly sophisticated models, often pushing the limits of computational feasibility. While emulation methods can reduce calculation costs, the neural architectures typically used--optimised via gradient descent--are themselves computationally expensive to train, particularly in terms of data generation requirements. This paper investigates the utility of the Extreme Learning Machine (ELM) as a lightweight, non-gradient-based machine learning algorithm for accelerating complex physical models. We evaluate ELM surrogate models in two test cases with different data structures: (i) sequentially-structured data, and (ii) image-structured data. For test case (i), where the number of samples $N$ >> the dimensionality of input data $d$, ELMs achieve remarkable efficiency, offering a 100,000$\times$ faster training time and a 40$\times$ faster prediction speed compared to a Bi-Directional Recurrent Neural Network (BIRNN), whilst improving upon BIRNN test performance. For test case (ii), characterised by $d >> N$ and image-based inputs, a single ELM was insufficient, but an ensemble of 50 individual ELM predictors achieves comparable accuracy to a benchmark Convolutional Neural Network (CNN), with a 16.4$\times$ reduction in training time, though costing a 6.9$\times$ increase in prediction time. We find different sample efficiency characteristics between the test cases: in test case (i) individual ELMs demonstrate superior sample efficiency, requiring only 0.28% of the training dataset compared to the benchmark BIRNN, while in test case (ii) the ensemble approach requires 78% of the data used by the CNN to achieve comparable results--representing a trade-off between sample efficiency and model complexity.

R. P. Nathan, Nikolaos Nikolaou, Ofer Lahav

Unsupervised machine learning methods are well suited to searching for anomalies at scale but can struggle with the high-dimensional representation of many modern datasets, hence dimensionality reduction (DR) is often performed first. In this paper we analyse unsupervised anomaly detection (AD) from the perspective of the manifold created in DR. We present an idealised illustration, "Finding Pegasus", and a novel formal framework with which we categorise AD methods and their results into "on manifold" and "off manifold". We define these terms and show how they differ. We then use this insight to develop an approach of combining AD methods which significantly boosts AD recall without sacrificing precision in situations employing high DR. When tested on MNIST data, our approach of combining AD methods improves recall by as much as 16 percent compared with simply combining with the best standalone AD method (Isolation Forest), a result which shows great promise for its application to real-world data.

Petr Mánek, Graham Van Goffrier, Vignesh Gopakumar et al.

The tritium breeding ratio (TBR) is an essential quantity for the design of modern and next-generation D-T fueled nuclear fusion reactors. Representing the ratio between tritium fuel generated in breeding blankets and fuel consumed during reactor runtime, the TBR depends on reactor geometry and material properties in a complex manner. In this work, we explored the training of surrogate models to produce a cheap but high-quality approximation for a Monte Carlo TBR model in use at the UK Atomic Energy Authority. We investigated possibilities for dimensional reduction of its feature space, reviewed 9 families of surrogate models for potential applicability, and performed hyperparameter optimisation. Here we present the performance and scaling properties of these models, the fastest of which, an artificial neural network, demonstrated $R^2=0.985$ and a mean prediction time of $0.898\ μ\mathrm{s}$, representing a relative speedup of $8\cdot 10^6$ with respect to the expensive MC model. We further present a novel adaptive sampling algorithm, Quality-Adaptive Surrogate Sampling, capable of interfacing with any of the individually studied surrogates. Our preliminary testing on a toy TBR theory has demonstrated the efficacy of this algorithm for accelerating the surrogate modelling process.

Kai Hou Yip, Quentin Changeat, Nikolaos Nikolaou et al.

Deep learning algorithms are growing in popularity in the field of exoplanetary science due to their ability to model highly non-linear relations and solve interesting problems in a data-driven manner. Several works have attempted to perform fast retrievals of atmospheric parameters with the use of machine learning algorithms like deep neural networks (DNNs). Yet, despite their high predictive power, DNNs are also infamous for being 'black boxes'. It is their apparent lack of explainability that makes the astrophysics community reluctant to adopt them. What are their predictions based on? How confident should we be in them? When are they wrong and how wrong can they be? In this work, we present a number of general evaluation methodologies that can be applied to any trained model and answer questions like these. In particular, we train three different popular DNN architectures to retrieve atmospheric parameters from exoplanet spectra and show that all three achieve good predictive performance. We then present an extensive analysis of the predictions of DNNs, which can inform us - among other things - of the credibility limits for atmospheric parameters for a given instrument and model. Finally, we perform a perturbation-based sensitivity analysis to identify to which features of the spectrum the outcome of the retrieval is most sensitive. We conclude that for different molecules, the wavelength ranges to which the DNN's predictions are most sensitive, indeed coincide with their characteristic absorption regions. The methodologies presented in this work help to improve the evaluation of DNNs and to grant interpretability to their predictions.

Nikolaos Nikolaou, Ingo P. Waldmann, Angelos Tsiaras et al.

The last decade has witnessed a rapid growth of the field of exoplanet discovery and characterisation. However, several big challenges remain, many of which could be addressed using machine learning methodology. For instance, the most prolific method for detecting exoplanets and inferring several of their characteristics, transit photometry, is very sensitive to the presence of stellar spots. The current practice in the literature is to identify the effects of spots visually and correct for them manually or discard the affected data. This paper explores a first step towards fully automating the efficient and precise derivation of transit depths from transit light curves in the presence of stellar spots. The methods and results we present were obtained in the context of the 1st Machine Learning Challenge organized for the European Space Agency's upcoming Ariel mission. We first present the problem, the simulated Ariel-like data and outline the Challenge while identifying best practices for organizing similar challenges in the future. Finally, we present the solutions obtained by the top-5 winning teams, provide their code and discuss their implications. Successful solutions either construct highly non-linear (w.r.t. the raw data) models with minimal preprocessing -deep neural networks and ensemble methods- or amount to obtaining meaningful statistics from the light curves, constructing linear models on which yields comparably good predictive performance.

Nikolaos Nikolaou, Konstantinos Sechidis

At the heart of causal structure learning from observational data lies a deceivingly simple question: given two statistically dependent random variables, which one has a causal effect on the other? This is impossible to answer using statistical dependence testing alone and requires that we make additional assumptions. We propose several fast and simple criteria for distinguishing cause and effect in pairs of discrete or continuous random variables. The intuition behind them is that predicting the effect variable using the cause variable should be `simpler' than the reverse -- different notions of `simplicity' giving rise to different criteria. We demonstrate the accuracy of the criteria on synthetic data generated under a broad family of causal mechanisms and types of noise.

Nikolaos Nikolaou, Henry Reeve, Gavin Brown

The ultimate goal of a supervised learning algorithm is to produce models constructed on the training data that can generalize well to new examples. In classification, functional margin maximization -- correctly classifying as many training examples as possible with maximal confidence --has been known to construct models with good generalization guarantees. This work gives an information-theoretic interpretation of a margin maximizing model on a noiseless training dataset as one that achieves lossless maximal compression of said dataset -- i.e. extracts from the features all the useful information for predicting the label and no more. The connection offers new insights on generalization in supervised machine learning, showing margin maximization as a special case (that of classification) of a more general principle and explains the success and potential limitations of popular learning algorithms like gradient boosting. We support our observations with theoretical arguments and empirical evidence and identify interesting directions for future work.

Nikolaos Nikolaou, Joseph Mellor, Nikunj C. Oza et al.

Probability estimates generated by boosting ensembles are poorly calibrated because of the margin maximization nature of the algorithm. The outputs of the ensemble need to be properly calibrated before they can be used as probability estimates. In this work, we demonstrate that online boosting is also prone to producing distorted probability estimates. In batch learning, calibration is achieved by reserving part of the training data for training the calibrator function. In the online setting, a decision needs to be made on each round: shall the new example(s) be used to update the parameters of the ensemble or those of the calibrator. We proceed to resolve this decision with the aid of bandit optimization algorithms. We demonstrate superior performance to uncalibrated and naively-calibrated on-line boosting ensembles in terms of probability estimation. Our proposed mechanism can be easily adapted to other tasks(e.g. cost-sensitive classification) and is robust to the choice of hyperparameters of both the calibrator and the ensemble.