Optimal thresholds and algorithms for a model of multi-modal learning in high dimensions

Abstract

This work explores multi-modal inference in a high-dimensional simplified model, analytically quantifying the performance gain of multi-modal inference over that of analyzing modalities in isolation. We present the Bayes-optimal performance and recovery thresholds in a model where the objective is to recover the latent structures from two noisy data matrices with correlated spikes. The paper derives the approximate message passing (AMP) algorithm for this model and characterizes its performance in the high-dimensional limit via the associated state evolution. The analysis holds for a broad range of priors and noise channels, which can differ across modalities. The linearization of AMP is compared numerically to the widely used partial least squares (PLS) and canonical correlation analysis (CCA) methods, which are both observed to suffer from a sub-optimal recovery threshold.

References in corpus (20)

• Message Passing Algorithms for Compressed Sensing
• The dynamics of message passing on dense graphs, with applications to compressed sensing
• Spectral redemption: clustering sparse networks
• Statistical physics of inference: Thresholds and algorithms
• A Cookbook of Self-Supervised Learning
• Adaptive Damping and Mean Removal for the Generalized Approximate Message Passing Algorithm
• MMSE of probabilistic low-rank matrix estimation: Universality with respect to the output channel
• Constrained Low-rank Matrix Estimation: Phase Transitions, Approximate Message Passing and Applications
• On Convergence of Approximate Message Passing
• Marvels and Pitfalls of the Langevin Algorithm in Noisy High-dimensional Inference
• Estimating rank-one matrices with mismatched prior and noise: universality and large deviations
• One-Way Matching of Datasets with Low Rank Signals
• Neural-prior stochastic block model
• Fundamental limits of community detection from multi-view data: multi-layer, dynamic and partially labeled block models
• Approximate Message Passing for the Matrix Tensor Product Model
• Simultaneous Dimensionality Reduction: A Data Efficient Approach for Multimodal Representations Learning
• High-Dimensional Canonical Correlation Analysis
• Spectral Phase Transition and Optimal PCA in Block-Structured Spiked models
• Matrix Denoising with Doubly Heteroscedastic Noise: Fundamental Limits and Optimal Spectral Methods
• Multimodal data integration and cross-modal querying via orchestrated approximate message passing