Approximate Message Passing(AMP) for LASSO

A Python implementation of Approximate Message Passing (AMP) algorithms for LASSO

references

These algorithms are implemented based on the following papers

ampy
- approximate message passing solvers for the Standard Linear Model
[0]AMP.ipynb
- a demonstration notebook for AMP
[1]Self Averaging AMP.ipynb
- a demonstration notebook for Self Averaging AMP
[2]Naive Self Averaging VAMP.ipynb
- a demonstration notebook for Naive Self Averaging VAMP
- Naive in the sense that matrix inverse of size N x N
[3]Self Averaging VAMP.ipynb
- a demonstration notebook for Self Averaging VAMP
- in this version singular value decomposition is utilized to avoid computational cost of N^3 for each iteration
[3-1]: same with the [3] except that the observation matrix is drawn from random DCT matrix ensemble.

Name		Name	Last commit message	Last commit date
Latest commit History 35 Commits
ampy		ampy
README.md		README.md
[0]Self Averaging AMP.ipynb		[0]Self Averaging AMP.ipynb
[1]AMP.ipynb		[1]AMP.ipynb
[2]Naive Self Averaging VAMP.ipynb		[2]Naive Self Averaging VAMP.ipynb
[3-1]Self Averaging VAMP random dct measurement.ipynb		[3-1]Self Averaging VAMP random dct measurement.ipynb
[3]Self Averaging VAMP.ipynb		[3]Self Averaging VAMP.ipynb
[4]Naive VAMP.ipynb		[4]Naive VAMP.ipynb
check_versions.py		check_versions.py