AI RESEARCH
Transformers Can Learn Posterior Predictive Distributions In-Context
arXiv CS.LG
•
ArXi:2605.26713v1 Announce Type: cross Prior-data fitted networks (PFNs) have recently emerged as a powerful approach for Bayesian prediction tasks, approximating the posterior predictive distribution (PPD) through in-context learning. Despite their strong empirical performance and ability to go beyond point predictions, theoretical understandings of the algorithmic capability of transformers to learn distributions in context are still lacking.