ExoDNN: Boosting exoplanet detection with artificial intelligence. Application to Gaia Data Release 3

TL;DR

ExoDNN addresses the challenge of expanding the exoplanet and brown dwarf census with Gaia DR3 astrometry by learning to map Gaia’s astrometric fit-quality statistics to the probability that a source hosts an unresolved companion. It trained a deep neural network on synthetic data and 31 DR3 features to predict $\hat{p}=P(y=1|\bar{x})$, achieving strong validation performance and validating the approach against real Gaia DR3 binaries. Applied to a volume-limited sample within 100 pc, ExoDNN yields 14,606 initial candidates and, after conservative post-processing, 7414 consolidated candidates across FGKM types, with a measured false-positive rate of about 1.2%. The results offer a scalable catalog to guide follow-up with future missions (e.g., PLATO) and Gaia DR4, while acknowledging limitations in mass determination and calibration-related biases that require external measurements for confirmation.

Abstract

We combine Gaia Data Release 3 and artificial intelligence to enhance the current statistics of substellar companions, particularly within regions of the orbital period vs. mass parameter space that remain poorly constrained by the radial velocity and transit detection methods. Using supervised learning, we train a deep neural network to recognise the characteristic distribution of the fit quality statistics corresponding to a Gaia DR3 astrometric solution for a non single star. We generate a deep learning model, ExoDNN, which predicts the probability of a DR3 source to host unresolved companions based on those fit quality statistics. Applying the predictive capability of ExoDNN to a volume limited sample of F,G,K and M stars from Gaia DR3, we have produced a list of 7414 candidate stars hosting companions. The stellar properties of these candidates, such as their mass and metallicity, are similar to those of the Gaia DR3 non single star sample. We also identify synergies with future observatories, such as PLATO, and we propose a follow up strategy with the intention of investigating the most promising candidates among those samples.

Figures (16)

• Figure 1: LEFT: Simulated orbital periods (black) compared to the orbital periods from DR3 NSS Orbital solutions (shade), with the DR3 time baseline marked in vertical. RIGHT: Simulated secondary masses (black), compared to the catalogue of DR3 binary masses (shade), with the peak of the distribution of simulated secondary masses in vertical.
• Figure 2: TOP: Barycentre motion of a simulated single star corresponding to Eq. \ref{['eqn:single_star']} with best fit astrometric parameters. The observed position of the source is marked with solid white circles, and the 1D along-scan observations with gray circles. BOTTOM: Same, but for a simulated binary system, where the observed position of the source is perturbed by a companion (exaggerated for illustration purposes).
• Figure 3: LEFT: Evolution of the training and validation set losses over the different epochs. RIGHT: Evolution of the Brier-score during the training.
• Figure 4: LEFT: classification threshold computation using the area under curve. RIGHT: classification threshold computation using the precision-recall curve. The optimal threshold that achieves the best compromise between false positive and true positive rates is marked as a red dot. This is our chosen $p_0$=0.242.
• Figure 5: TOP: Parameter impact on the model output as estimated by the neural network model ranked by decreasing impact. High parameter values are color-coded in red, while low values are displayed in blue. BOTTOM: Predicted probability for each binary star example on the test dataset (black dots), together with a detection probability proxy computed as $p=1-\Phi(S/N,\delta)$), where $\Phi$ is the CDF of the normal distribution, $S/N=\alpha/\sigma$, and $\delta$ is a variable S/N threshold from 1 to 3.