1. Comparative Study of Neural Networks, Random Feature Regression, and Neural Tangent Kernel

2. Understanding the Optimization Landscape of Deep Neural Networks

Past Projects

1. Understanding the Behavior of Variational Inference in Topic Models

2. Optimal Eigenvalue Shrinkage for Spiked Model

Publications

You can access my Google Scholar profile here.

2019

• Ghorbani, B., Mei, S., Misiakiewicz, T., Montanari, A. “Limitations of Lazy Training of Two-layers Neural Networks” NeurIPS (2019) (Accepted for Spotlight) link

• Ghorbani, B., Mei, S., Misiakiewicz, T., Montanari, A. “Linearized Two-Layers Neural Networks in High Dimension” Submitted to Annals of Statistics (2019) link

• Ghorbani, B., Xiao, Y., Krishnan, S. “An Investigation into Neural Net Optimization Via Hessian Eigenvalue Density” ICML (2019) link code

• Ghorbani, B., Xiao, Y., Krishnan, S. “The Effect of Network Depth on the Optimization Landscape” ICML Workshop on Deep Phenomena (2019) link

• Ghorbani, B., Javadi, H., Montanari, A. “An Instability in Variational Inference for Topic Models” ICML (2019) link code

2018

• Donoho, D., Ghorbani, B. “Optimal Covariance Estimation for Condition Number Loss in the Spiked Model” submitted to the Annals of Statistics (2018) link

2015

• Ghorbani, B., Yilmaz, O. “Sparse Regression With Highly Correlated Predictors” Undergraduate Thesis link