Transformer Embeddings for Fast Microlensing Inference

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Smyth, Nolan, Perreault-Levasseur, Laurence, Hezaveh, Yashar
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866915671431970816
author Smyth, Nolan
Perreault-Levasseur, Laurence
Hezaveh, Yashar
author_facet Smyth, Nolan
Perreault-Levasseur, Laurence
Hezaveh, Yashar
contents The search for free-floating planets (FFPs) is a key science driver for upcoming microlensing surveys like the Nancy Grace Roman Galactic Exoplanet Survey. These rogue worlds are typically detected via short-duration microlensing events, the characterization of which often requires analyzing noisy, irregularly-sampled observations. We present a pipeline for this task using simulation-based inference. We use a Transformer encoder to learn a compressed summary representation of the raw time-series data, which in turn conditions a neural posterior estimator. We demonstrate that our method produces accurate and well-calibrated posteriors over three orders of magnitude faster than traditional methods. We also demonstrate its performance on KMT-BLG-2019-2073, a short-duration FFP candidate event.
format Preprint
id arxiv_https___arxiv_org_abs_2512_11687
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Transformer Embeddings for Fast Microlensing Inference
Smyth, Nolan
Perreault-Levasseur, Laurence
Hezaveh, Yashar
Instrumentation and Methods for Astrophysics
Earth and Planetary Astrophysics
Astrophysics of Galaxies
The search for free-floating planets (FFPs) is a key science driver for upcoming microlensing surveys like the Nancy Grace Roman Galactic Exoplanet Survey. These rogue worlds are typically detected via short-duration microlensing events, the characterization of which often requires analyzing noisy, irregularly-sampled observations. We present a pipeline for this task using simulation-based inference. We use a Transformer encoder to learn a compressed summary representation of the raw time-series data, which in turn conditions a neural posterior estimator. We demonstrate that our method produces accurate and well-calibrated posteriors over three orders of magnitude faster than traditional methods. We also demonstrate its performance on KMT-BLG-2019-2073, a short-duration FFP candidate event.
title Transformer Embeddings for Fast Microlensing Inference
topic Instrumentation and Methods for Astrophysics
Earth and Planetary Astrophysics
Astrophysics of Galaxies
url https://arxiv.org/abs/2512.11687