Search papers, labs, and topics across Lattice.
This paper introduces TransitNet, a compact attention-augmented deep learning framework designed for low-signal-to-noise ratio (SNR) transit blind searches, addressing the challenge of detecting intermediate-to-long-period Earth-size planets. By developing a unified dataset construction, benchmarking, and threshold-selection framework, TransitNet achieves a remarkable 95.2% accuracy in SNR ranges of 6 to 8, significantly outperforming traditional methods like TLS and BLS. Additionally, it demonstrates high efficiency with a compact model size of 1.5 MB and substantial speed improvements, making it a promising tool for future planet detection efforts.
Achieving a 93% recovery rate for Earth-size transits, TransitNet outperforms traditional methods while maintaining a compact model size and high inference speed.
Motivated by the observational incompleteness of intermediate-to-long-period Earth-size planets, we present TransitNet, a compact attention-augmented deep-learning framework for low-SNR transit blind searches. To enable realistic method development and objective threshold calibration under blind-search conditions, we develop a unified dataset construction, benchmarking, and threshold-selection framework. On recovery benchmarks constructed from unseen Kepler targets, TransitNet attains 95.2 percent accuracy in the challenging SNR range of 6 to 8 and outperforms both TLS and BLS, achieving ROC-AUC and PR-AP values of 0.974 and 0.982, respectively. In an injected Earth-size and sub-Earth-size transit recovery experiment, TransitNet achieves a recovery rate of 93.0 percent, substantially exceeding those of TLS (63.1 percent) and BLS (60.0 percent). In addition to detection, TransitNet provides attention-based estimates of transit windows and midpoints. On an independent evaluation set, 97.4 percent of injected transits are fully covered by the estimated transit window. Applied to real Kepler observations, the model successfully recovers all 34 selected confirmed Kepler planets, with a mean absolute transit midpoint error of 1.24 hours. The model combines a compact footprint of about 1.5 MB with high inference efficiency, yielding speed-ups of about 12 to 25 times relative to CPU-TLS and about 4 to 5 times relative to CPU-BLS. These results demonstrate that TransitNet provides an accurate, scalable, and computationally efficient framework for low-SNR transit blind searches in the tested regime and motivate its extension to longer-period Earth-size planet searches.
