Message Passing and GLMs

Master's Thesis, April 2024

To access the document, please click [PDF]
This is a writing sample of my MS Thesis. The last two Chapters were partially cut out because they contain unpublished work.

Full title

A roadmap to Message Passing methods for inference on Mixed Generalized Linear Models, with emphasis on Mixed Noiseless Phase Retrieval.

Image credits

DallE-2; see exact prompt in the doc.

Abstract

Chapter I is a review of inference, with an eye on what will be discussed in Chapter II, which discusses concepts that are shared with Statistical Physics. Belief Propagation, and Message Passing Algorithms in general are introduced in Chapter III, emphasizing on Approximate Message Passing in Chapter IV. The final Chapter V analyzes a recent adaptation of the techniques to estimating more than one signal.