Conference

PriorGuide: Test-Time Prior Adaptation for Simulation-Based Inference
PriorGuide enables efficient incorporation of arbitrary priors at inference time for amortized diffusion-based simulation-based inference, without retraining the model.
Yang Yang, Severi Rissanen, Paul E. Chang, Nasrulloh Loka, Daolang Huang, Arno Solin, Markus Heinonen, Luigi Acerbi
Project Paper arXiv Code Cite
Efficient Autoregressive Inference for Transformer Probabilistic Models
We propose an architecture extension to existing tabular foundations to generate joint predictive samples ~20x faster with minimal increase in training overhead and drop in performance.
Conor Hassan, Nasrulloh Loka, Cen-You Li, Daolang Huang, Paul E. Chang, Yang Yang, Francesco Silvestrin, Samuel Kaski, Luigi Acerbi
Paper arXiv Code Cite
Probabilistic Multi-Dimensional Classification
We propose a formal framework for probabilistic multi-dimensional classification.
Vu-Linh Nguyen, Yang Yang, Cassio De Campos
Paper arXiv Code Cite
Bayesian Structure Scores for Probabilistic Circuits
We develop Bayesian structure scores for deterministic PCs, i.e., the structure likelihood with parameters marginalized out.
Yang Yang, Gennaro Gala, Robert Peharz
Paper arXiv Code Cite