approximate-message-passing

Official implementation of AMP-based consistent activity status and channel estimation

Code for "Bayesian reconstruction of memories stored in neural networks from their connectivity" https://arxiv.org/abs/2105.07416

Codes for atlas building and construction of Prediction sets for cross modal query as described in 'Multimodal data integration and cross-modal querying via orchestrated approximate message passing' by Nandy and Ma (2024)

approximate message massing (AMP), bridge regression, optimal tuning

Simulation and inference for contextual and neural SBMs

Code for "Adaptive Score-Based VAMP: Self-Tuning Hyperparameters via Tilted EM" — per-module tilted-EM hyperparameter estimation for SC-VAMP, with linear and one-bit compressed-sensing experiments.

Multi-modal denoising & prediction via Approximate Message Passing with a per-cluster Gaussian-Mixture empirical-Bayes prior (OrchAMP) — synthetic experiments plus real applications to TEA-seq single-cell multi-omics and TCGA-BRCA survival, benchmarked against JAFAR, MOFA+, Multigrate, and Cooperative Lasso.