Skip to content

In-Context Generative Posterior Sampling (ICGPS)

Home / Concepts / In-Context Generative Posterior Sampling (ICGPS)

In-Context Generative Posterior Sampling (ICGPS)

Auto-generated stub. Edit this file to add more details.

A meta-learned framework for solving censored decision-making problems by treating posterior sampling as an in-context autoregressive generation task.

Why It Matters

ICGPS provides a novel plug-in template for sequential decision-making under censoring, shifting the problem from iterative imputation to autoregressive generation.

Evidence

We propose in-context generative posterior sampling (ICGPS), which uses modern generative models that are meta-trained offline and deployed online by in-context autoregressive generation.

Metadata & Links