Spectra: Exact Component Transport for Test-Time Prior Adaptation in Simulation-Based Inference

Key points

• Simulation-based inference (SBI) has become a powerful approach to Bayesian inference in complex scientific models whose likelihoods are difficult or impossible to evaluate.1
• Amortized SBI learns reusable inference models from simulated data, enabling rapid posterior inference for new observations, and modern generative models have made these models increasingly expressive.1
• Spectra uses an exact score-transport identity to obtain the adapted score from a frozen diffusion model in closed form for structured prior changes, without additional simulation or training.1
• Across six SBI benchmarks, Spectra achieves accurate adaptation under strong prior shifts at low online sampling cost.1

Sources (1)

• [1]Spectra: Exact Component Transport for Test-Time Prior Adaptation in Simulation-Based Inference
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 09:14 AM

  We introduce Spectra, a test-time adaptation method for diffusion-based SBI.

  Simulation-based inference (SBI) has become a powerful approach to Bayesian inference in complex scientific models whose likelihoods are difficult or impossible to evaluate.

Extractive summary: sentences quoted from the sources.