ExoNet applies a calibrated late-fusion CNN-plus-attention model to Kepler-labeled data and identifies 1,754 high-confidence unconfirmed TESS planet candidates plus 52 habitable-zone targets.
A CNN-BiLSTM-Attention architecture for kepler exoplanet transit vetting
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
astro-ph.EP 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ExoNet: Calibrated Multimodal Deep Learning for TESS Exoplanet Candidate Vetting using Phase-Folded Light Curves, Stellar Parameters, and Multi-Head Attention
ExoNet applies a calibrated late-fusion CNN-plus-attention model to Kepler-labeled data and identifies 1,754 high-confidence unconfirmed TESS planet candidates plus 52 habitable-zone targets.