AI Science Tracker

Physics-based differentiable imaging detects 51 Eridani b in a JWST band where standard processing did not

Northwestern University; California Institute of Technology; Massachusetts Institute of Technology; Space Telescope Science Institute · 4 Nov 2026 · physics

Discovery

Northwestern University; California Institute of Technology; Massachusetts Institute of Technology; Space Telescope Science Institute

Physics-based differentiable imaging detects 51 Eridani b in a JWST band where standard processing did not

Archived from discovery counts: the lead researchers explicitly state that FLIP did not use AI. Its automatic-differentiation tools optimize a physics-based instrument model. A physics-based coronagraph model optimized with automatic differentiation and gradient descent recovered 51 Eridani b in JWST/NIRCam F430M images at 4.3 μm, where the published standard-processing result was a non-detection. The paper also reports two- to fifteen-fold sensitivity gains across its test observations.

Technology: FLIP (Physics-based Learning for Imaging Planets), a differentiable instrument model optimized with automatic differentiation

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

Result, AI contribution and limits

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

  • Publication status: Peer-reviewed paper
  • Test type: retrospective or benchmark
  • Tested: A physics-based coronagraph model optimized with automatic differentiation and gradient descent recovered 51 Eridani b in JWST/NIRCam F430M images at 4.3 μm, where the published standard-processing result was a non-detection.
  • Still unknown: The planet 51 Eridani b was already discovered by ground-based imaging in 2015; the new finding is its first JWST image in the 4.3 μm carbon-dioxide band, not discovery of a previously unknown planet. Performance estimates are based on a limited set of three target systems and should not be generalized to all JWST data.
  • AI role: The researchers state that this work did not use AI, although it used computational tools also used to train AI models. FLIP optimized a physics-based telescope model using automatic differentiation.

Evidence sources: https://www.nature.com/articles/s41586-026-11213-z.pdf · https://www.nature.com/articles/s41586-026-11213-z · https://ciera.northwestern.edu/2026/11/10/hidden-planet-emerges-from-the-glare/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: linked separately.

Review scope and date are not recorded; source checked does not establish full-text review.

Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Follow-up and relationships

No follow-up recorded; this is not a claim that no follow-up exists.

Follow-up review scope and date are not recorded.

No source-supported relationships recorded.

Browse the current catalog.