aiqbench analysis
This research introduces a novel amortized inference technique using likelihood-weighted normalizing flows to efficiently estimate multi-modal posteriors in high-dimensional inverse problems without extensive training samples. The study highlights that initializing these models with Gaussian Mixture Models significantly improves fidelity by preventing spurious probability bridges between disconnected modes, offering a robust alternative for complex Bayesian tasks like extracting Wolfenstein parameters in heavy flavor physics.
Amortized Inference of Multi-Modal Posteriors using Likelihood-Weighted Normalizing Flows
arXiv:2512.04954v3 Announce Type: replace Abstract: We present a novel technique for amortized posterior estimation using Normalizing Flows trained with likelihood-weighted importance sampling. This approach allows for the efficient inference of theoretical parameters in high-dimensional inverse problems without the need for posterior training samples. We implement the method on multi-modal benchmark tasks in 2D and 3D to check for the efficacy. A critical observation of our study is the impact of the topology of the base distributions on the modelled posteriors. We find that standard unimodal base distributions fail to capture disconnected support, resulting in spurious probability \textit{bridges} between modes. We demonstrate that initializing the flow with a Gaussian Mixture Model that matches the cardinality of the target modes significantly improves reconstruction fidelity, as measured by some distance and divergence metrics. Finally, we apply this method to a curated problem in heavy flavour physics --- the extraction of the Wolfenstein parameters from the CP asymmetry in $B^0\to J/\psi\,K^0$; it is multimodal, non-Gaussian, and asymmetric in its mode weights --- and compare the results against a well-converged Markov Chain Monte Carlo reference using different metrics.
Abstract:We present a novel technique for amortized posterior estimation using Normalizing Flows trained with likelihood-weighted importance sampling. This approach allows for the efficient inference of theoretical parameters in high-dimensional inverse problems without the need for posterior training samples. We implement the method on multi-modal benchmark tasks in 2D and 3D to check for the efficacy. A critical observation of our study is the impact of the topology of the base distributions on the modelled posteriors. We find that standard unimodal base distributions fail to capture disconnected support, resulting in spurious probability \textit{bridges} between modes. We demonstrate that initializing the flow with a Gaussian Mixture Model that matches the cardinality of the target modes significantly improves reconstruction fidelity, as measured by some distance and divergence metrics. Finally, we apply this method to a curated problem in heavy flavour physics --- the extraction of the Wolfenstein parameters from the CP asymmetry in $B^0\to J/\psi\,K^0$; it is multimodal, non-Gaussian, and asymmetric in its mode weights --- and compare the results against a well-converged Markov Chain Monte Carlo reference using different metrics.
Machine Learning (cs.LG); High Energy Physics - Experiment (hep-ex); High Energy Physics - Phenomenology (hep-ph); Computational Physics (physics.comp-ph); Data Analysis, Statistics and Probability (physics.data-an)
arXiv:2512.04954 [cs.LG]
(or arXiv:2512.04954v3 [cs.LG] for this version)
Full text under the publisher's open licence · 210 words
Related coverage
Amortizing intractable inference in large language models
arXiv · Sep 1, 2026
Expenditure Horizon: Measuring Optimization Ability, with an Application to NanoGPT
METR · Jul 21, 2026
Private Generative Bootstrap via Blocking
arXiv · Aug 4, 2026
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · Oct 8, 2026