Finding Worlds Beyond Our Solar System

Deep learning analysis of 6,061 confirmed exoplanets. 4 novel approaches. 28 potentially habitable worlds identified.

6,061
Confirmed Planets
28
Habitable Candidates
4
Novel Approaches
99.8%
Best Accuracy

🧠 Novel Deep Learning Contributions

Multi-Modal Attention
99.8% accuracy • Explainable AI
Habitability Predictor
Multi-task • R² = 0.74
Uncertainty Quantification
Bayesian • Calibrated
Light Curve Transformer
ViT architecture • SOTA

Novel Approaches

Custom deep learning architectures for exoplanet detection

Attention Network
Multi-Modal Attention Network Novel
Habitability
Habitability Predictor Novel
Uncertainty
Uncertainty Quantification Novel
Transformer
Light Curve Transformer Novel

2025 Exoplanet Census

Latest data from NASA Exoplanet Archive and TESS mission

Census
Combined Analysis Data
Habitable
Habitable Planet Search Data

Machine Learning Models

Baseline models and GPU-accelerated neural networks

Detection
Detection Model Comparison ML
Neural Network
Neural Network (RTX 3060) ML
Anomaly
Anomaly Detection ML
Overview
Data Overview EDA

Scientific Analysis

Stellar classification, habitable zones, and sky mapping

HZ
Habitable Zone Analysis Science
Stellar
Stellar Host Analysis Science
FP
False Positive Breakdown Science
Sky
Kepler Field Sky Map Science

Key Findings

Closest Earth Twin
Kepler-1649 c
1.06 Earth radii • 234K temperature
Famous System
TRAPPIST-1
7 planets • 3 in habitable zone
Detection Method
Transit
4,472 planets (74% of total)
TESS Candidates
7,821
Objects of interest for follow-up
Unusual Planets
115
Hot Jupiters, lava worlds, outliers
Multi-Planet Systems
432
Up to 7 planets per system