Amortized Opacity Marginalization Improves C/O Interval Calibration for Brown-Dwarf Retrievals
Abstract
Molecular line lists disagree at a level that measurably shifts retrieved C/O, so retrievals conditioned on a single list omit a dominant systematic from their error bars. Classical marginalization re-runs a full retrieval under K opacity realizations per object. We move that cost into simulation: while generating the training set we randomize opacities (smooth multiplicative Fourier fields in logν plus a discrete line-list mixture, under a deliberately broad prior informed by measured inter-list differences), so a neural posterior estimator learns an already-marginalized posterior. Conceptually, this is the amortized analogue of repeating a conventional retrieval under many plausible opacity realizations and averaging the resulting posteriors. On perturbed held-out spectra (n=2048) a standard NPE covers nominal 90% C/O intervals 43.3% of the time against 82.4% for a compute-matched marginalized network; three-seed control ensembles reach 93.2% on clean data but 61.4% under perturbation, against 89.4% marginalized, isolating the deficit as opacity-induced (TARP joint expected-coverage deviation 0.005 against 0.035 for a single control network). An axis ablation traces the gain to the continuous field (45.4% for the discrete swaps alone, 79.3% for the field alone). Applying the 105 real bad-pixel masks with median filling then collapsed the ensemble's C/O coverage from 0.901 to 0.479; retraining on the same mask distribution and pooling three such seeds returns a nominal 0.905 C/O and 0.891 overall under the full deployment path, still in simulation (n=1024, same 105 masks), at 0.21 to 0.31 of the prior width. On 105 JWST/G395H spectral products of 13 late-T/Y dwarfs, single marginalized networks bracket the Hood et al. (2024) T8 benchmark C/O where the control excludes it, a consistency check rather than a validation, at 1.6-2.6x wider intervals.
Related
- Optimal Strategies for Gravitational-Wave Memory Detection with Noise Marginalization in Pulsar Timing Arrays
- Rotational Mapping and Regional Atmospheric Retrievals of Variable Brown Dwarfs: Application to Luhman 16B and SIMP 0136
- Impacts of Correlated Noise on Retrievals of Exo-Earth Atmospheres
- PROSERPINA: An Automated Pipeline for Martian Ionospheric Radio-occultation Retrievals with Probabilistic Uncertainty and Spherical-symmetry Bias Quantification
- Impacts of Correlated Noise on Retrievals of Exo-Earth Atmospheres
BibTeX
@misc{heraty2026amortized,
title = {{Amortized Opacity Marginalization Improves C/O Interval Calibration for Brown-Dwarf Retrievals}},
author = {Kellen Heraty},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.01665},
eprint = {2609.01665},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.01665},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Amortized Opacity Marginalization Improves C/O Interval Calibration for Brown-Dwarf Retrievals
AU - Kellen Heraty
PY - 2026
DA - 2026/10/05
PB - arXiv
DO - 10.48550/arxiv.2609.01665
UR - https://arxiv.org/abs/2609.01665
AB - Molecular line lists disagree at a level that measurably shifts retrieved C/O, so retrievals conditioned on a single list omit a dominant systematic from their error bars. Classical marginalization re-runs a full retrieval under K opacity realizations per object. We move that cost into simulation: while generating the training set we randomize opacities (smooth multiplicative Fourier fields in logν plus a discrete line-list mixture, under a deliberately broad prior informed by measured inter-list differences), so a neural posterior estimator learns an already-marginalized posterior. Conceptually, this is the amortized analogue of repeating a conventional retrieval under many plausible opacity realizations and averaging the resulting posteriors. On perturbed held-out spectra (n=2048) a standard NPE covers nominal 90% C/O intervals 43.3% of the time against 82.4% for a compute-matched marginalized network; three-seed control ensembles reach 93.2% on clean data but 61.4% under perturbation, against 89.4% marginalized, isolating the deficit as opacity-induced (TARP joint expected-coverage deviation 0.005 against 0.035 for a single control network). An axis ablation traces the gain to the continuous field (45.4% for the discrete swaps alone, 79.3% for the field alone). Applying the 105 real bad-pixel masks with median filling then collapsed the ensemble's C/O coverage from 0.901 to 0.479; retraining on the same mask distribution and pooling three such seeds returns a nominal 0.905 C/O and 0.891 overall under the full deployment path, still in simulation (n=1024, same 105 masks), at 0.21 to 0.31 of the prior width. On 105 JWST/G395H spectral products of 13 late-T/Y dwarfs, single marginalized networks bracket the Hood et al. (2024) T8 benchmark C/O where the control excludes it, a consistency check rather than a validation, at 1.6-2.6x wider intervals.
ER -