Pruning is Optimal for Learning Sparse Features in High-Dimensions
Vural, Nuri Mert, Erdogdu, Murat A.
While it is commonly observed in practice that pruning networks to a certain level of sparsity can improve the quality of the features, a theoretical explanation of this phenomenon remains elusive. In this work, we investigate this by demonstrating that a broad class of statistical models can be optimally learned using pruned neural networks trained with gradient descent, in high-dimensions. We consider learning both single-index and multi-index models of the form $y = \sigma^*(\boldsymbol{V}^{\top} \boldsymbol{x}) + \epsilon$, where $\sigma^*$ is a degree-$p$ polynomial, and $\boldsymbol{V} \in \mathbbm{R}^{d \times r}$ with $r \ll d$, is the matrix containing relevant model directions. We assume that $\boldsymbol{V}$ satisfies a certain $\ell_q$-sparsity condition for matrices and show that pruning neural networks proportional to the sparsity level of $\boldsymbol{V}$ improves their sample complexity compared to unpruned networks. Furthermore, we establish Correlational Statistical Query (CSQ) lower bounds in this setting, which take the sparsity level of $\boldsymbol{V}$ into account. We show that if the sparsity level of $\boldsymbol{V}$ exceeds a certain threshold, training pruned networks with a gradient descent algorithm achieves the sample complexity suggested by the CSQ lower bound. In the same scenario, however, our results imply that basis-independent methods such as models trained via standard gradient descent initialized with rotationally invariant random weights can provably achieve only suboptimal sample complexity.
Jun-12-2024
- Country:
- North America
- United States
- Louisiana > Orleans Parish
- New Orleans (0.04)
- California > Los Angeles County
- Long Beach (0.04)
- Arizona > Maricopa County
- Phoenix (0.04)
- Louisiana > Orleans Parish
- Canada
- Ontario > Toronto (0.14)
- British Columbia > Metro Vancouver Regional District
- Vancouver (0.04)
- United States
- Europe
- France (0.04)
- United Kingdom > England
- Greater London > London (0.04)
- Africa
- Rwanda > Kigali
- Kigali (0.04)
- Middle East > Tunisia
- Ben Arous Governorate > Ben Arous (0.04)
- Rwanda > Kigali
- North America
- Genre:
- Research Report > New Finding (0.87)
- Technology: