NeurIPS 2024 Ariel Data Challenge: Characterisation of Exoplanetary Atmospheres Using a Data-Centric Approach

TL;DR

This paper explores machine learning techniques for characterizing exoplanetary atmospheres from spectral data, emphasizing a data-centric approach that balances generalization, interpretability, and performance within a competitive framework.

Contribution

It introduces a data-centric methodology focusing on uncertainty estimation and feature engineering for exoplanetary atmosphere analysis, highlighting limitations of tabular models.

Findings

Uncertainty estimation significantly improves the GLL score.

Incorporating signal transformation enhances model performance.

Tabular models face inherent limitations in this spectral analysis task.

Abstract

The characterization of exoplanetary atmospheres through spectral analysis is a complex challenge. The NeurIPS 2024 Ariel Data Challenge, in collaboration with the European Space Agency's (ESA) Ariel mission, provided an opportunity to explore machine learning techniques for extracting atmospheric compositions from simulated spectral data. In this work, we focus on a data-centric business approach, prioritizing generalization over competition-specific optimization. We briefly outline multiple experimental axes, including feature extraction, signal transformation, and heteroskedastic uncertainty modeling. Our experiments demonstrate that uncertainty estimation plays a crucial role in the Gaussian Log-Likelihood (GLL) score, impacting performance by several percentage points. Despite improving the GLL score by 11%, our results highlight the inherent limitations of tabular modeling and feature engineering for this task, as well as the constraints of a business-driven approach within a Kaggle-style competition framework. Our findings emphasize the trade-offs between model simplicity, interpretability, and generalization in astrophysical data analysis.