Ultracool dwarf Science with MachIne LEarning (USMILE). I. Scalable Tree-based Models for Photometric Spectral Classification and New Discoveries from LSST Data Preview 1 and Euclid Quick Data Release 1

Abstract We present Ultracool dwarf Science with MachIne LEarning (USMILE), a program applying machine learning tools for the discovery and characterization of ultracool dwarfs. We introduce USMILE Avocado , a spectral classification framework that uses broadband photometry from wide-field surveys—Rubin Observatory LSST Data Preview 1 (DP1), VISTA Hemisphere Survey (VHS), and CatWISE—as input features. The framework comprises two gradient-boosted decision-tree models scalable to the massive data volumes of modern surveys: the classifier , which distinguishes ultracool dwarfs from stellar/extragalactic contaminants, and the regressor , which predicts spectral types. A key strength is its ability to natively handle missing photometric features, common in wide-field searches, whereas earlier machine learning approaches required complete multiband detections or relied on imputation, thereby excluding genuine ultracool dwarfs or introducing bias. Trained on an augmented labeled data set of >2 million sources built from known ultracool dwarfs, reddened early-type stars, and quasars, the models achieve strong performance: the classifier attains an Area Under the Curve of the Receiver Operating Characteristic of 0.976 and an F1 score of 0.92, while the regressor yields a mean squared error of 0.88 subtypes. Applying these models, we carried out the first ultracool dwarf search with LSST DP1, cross-matched against VHS and CatWISE. Crucially, Euclid Quick Data Release 1 provided near-infrared spectra for hundreds of candidates, enabling a rare, large-scale external spectroscopic validation. This confirmed 15 M6–L2 discoveries, verified USMILE performance, and clarified regimes where USMILE predictions are most reliable. Building on these insights, we identified 25 additional high-quality M6–L9 photometric candidates. These early discoveries demonstrate the effectiveness of scalable machine learning methods in the data-rich era of wide-field surveys, highlighting the synergy between LSST and Euclid in expanding the ultracool dwarf census.