Summer 2024: Parsimonious Structural Learning

“Simplicity is the ultimate sophistication.” – Leonardo da Vinci

Theme

The first few meetings will review basic ideas from compressed sensing, dictionary learning, subspace clustering, and low-rank matrix factorization, before moving on to recent papers on topics in these areas as well as parsimonious representation learning, implicit biases and sparse training of neural networks, and sparse neural networks/pruning.

Logistics

In general we meet on Mondays, 4:00-5:00 PM. We are happy to adapt to conflicts in advance. So far we have tried various conference rooms, but the place is still oscillating. Stay tuned!

Schedule

Date	Presented by	Title	Remarks
June 3	Kyle Poe	Compressed sensing/sparse recovery on subspace arrangements
June 24	Tianjiao Ding	A tale of two problems: sparse signal recovery and sparse subspace clustering
July 1	Ziqing Xu	Optimization for low-rank matrix factorization problem
July 8