Trainable JWST model boosts exoplanet sensitivity 2–15x in Nature paper
bravo_abad · x · 2026-11-11
A Nature paper by Ferrer-Chavez et al. (with Katherine Bouman and William Freeman) turns telescope calibration into a learning problem.
- Core idea: build a trainable physical model of the James Webb instrument, using automatic differentiation and gradient descent to fit its time-varying optical behavior and subtract starlight more accurately.
- Results: on real observations, sensitivity improved 2–15x over standard processing; it recovered the known planet 51 Eridani b at 4.3 micrometres, where standard analysis detected nothing.
- Safeguard: the team injected artificial planets to verify the fitting didn't erase real signals.
The scientific test: whether the signal survives the fit.
More from Embodied
- Luna, a jewelry-style AI wearable that speaks before you ask, launches on Kickstarter — nikola_mr64990 · 2026-11-11
- Dev reverse-engineers Anker AI recording earbuds, ships open-source client app — vista8 · 2026-11-11
- Simulated hand with 39 muscles learns 9/9 poses after 426M training steps — MaziyarPanahi · 2026-11-11
- Versor pitches real2sim2real: robotics sim needs physical dynamics, not just visuals — abhishekunique7 · 2026-11-11
- Origami Robotics and PalatialSim unveil real2sim2real Dexterity Factory, 100 open datasets next week — ybisk · 2026-11-11
- Clean point clouds: Tersus GNSS MVP S1 scans turned into high-quality splats in LichtFeld Studio — janusch_patas · 2026-11-11