Significance of the Research

Thousands of exoplanets with varying masses, radii, and orbital parameters have been discovered and confirmed to orbit other stars, including hot Jupiters, mini-Neptunes, and super-Earths. Thanks to rapid advancements in observational technology, Earth-sized exoplanets with masses less than 10 Earth masses have been identified, some of which may have rocky compositions similar to terrestrial planets in our solar system. This discovery ushers in a new era of planetary habitability exploration and raises questions about the potential for life beyond our solar system.

The interior structure of rocky exoplanets plays a critical role in their habitability. Features such as the core, mantle, crust, and potential water layers collectively determine a planet’s geological activity, magnetic field generation, atmosphere retention, and surface environment stability, all of which influence the planet’s habitability.

The presence of liquid water on the surface of rocky exoplanets is essential for life. A planetary magnetic field can significantly reduce atmospheric loss and shield the surface from high-energy particles, which requires a molten core. The interior structure also affects atmospheric formation and chemical composition, crucial for maintaining surface temperature stability and providing life-supporting conditions. Therefore, accurately understanding and predicting the interior structures of these rocky exoplanets is vital for identifying potentially habitable planets.

Currently, mass and radius are the primary observable quantities for exoplanets. By comparing these with theoretical mass-radius curves, we can obtain initial insights into the internal structure and overall composition of exoplanets. However, there is often a degeneracy between the internal structure and the observed mass and radius, meaning different internal models can explain the same set of mass and radius observations within their uncertainties.

In recent years, Bayesian inference has proven to be a powerful method for solving such inverse problems, providing robust posterior distributions of internal structures by incorporating prior knowledge. However, Bayesian inference often relies on the computationally intensive and time-consuming Markov chain Monte Carlo (MCMC) method. This means that inferring the internal structure of a single exoplanet may require hundreds of thousands of calculations, taking hours to days to complete.

Given the increasing number of exoplanets and the continuous update of observational data, developing and adopting more efficient and powerful inference methods is crucial.

Research Plan

1. Generating Machine Learning Training Sets: Construct a three-layer internal structure model for rocky exoplanets and generate large datasets for training machine learning models. Preprocess the data, including normalization and variable selection, to suit the needs of the machine learning models.
2. Model Development: Build a model based on mixture density neural networks (MDN) and train it to learn the features of the internal structures of exoplanets. Use techniques such as cross-validation and early stopping to avoid overfitting and ensure the model’s generalization capability.
3. Model Evaluation and Optimization: Assess the performance and accuracy of the machine learning model by comparing it with existing planetary interior structure models and observational data. Adjust and optimize the model based on evaluation results to improve prediction accuracy.
4. Habitable Planet Screening: Use the optimized model to predict the internal structures of observed rocky exoplanets. Screen these planets based on predicted results and habitability indicators to identify candidates with potential for habitability.
5. Future Work: Continuously update and improve the model with more high-quality observational data. Explore new machine learning algorithms and techniques to further enhance prediction accuracy and efficiency. Collaborate with researchers in astrophysics, geology, and related fields to deepen the understanding of the internal structures and habitability of rocky exoplanets.

Through this research plan, we aim to accelerate the search for habitable planets and provide new perspectives and tools for future exoplanet exploration and research.