Project

The goal of the project is to investigate a topic of interest to you that involves applying optimization, sparse representations or low-rank models to analyze data.

Project report

• The project should be written in Latex and have an extension of no more than 10 pages. Extra figures and derivations may be added in the appendix.

• Make sure you structure the report well. You should at least have an introduction, a section where you explain what you did, a section for numerical experiments and a conclusion section.

• Make clarity a priority. Make sure that you explain what is the point of your project and what conclusions you arrived to (and why or why not they make sense).

• Hand in the project by sending a pdf file via email.

Topics

Below is a list of possible topics, you are very welcome to suggest your own!

• Compressed sensing in MRI

• Graphical lasso

• Political analysis of Twitter data

• Sparse additive models

• Mixed regression

• Statistical inference via the lasso: See Chapter 6

• Scattering transform

• One-bit compressed sensing

• Finite rate of innovation

• Curvelets

• Face recognition via sparse representation

• Phase retrieval via Wirtinger flow

• Transform Invariant Low-rank Textures

• Tensor completion

• Video magnification

• Subspace clustering

• Latent variable graphical model selection via convex optimization

• Image upsampling

• Blind deconvolution

• Sparse FFT

• Robust optimization

• Sparse autoencoders

• Guarantees for nonconvex problems