DeepHoyer: Learning Sparser Neural Network with Differentiable Scale-Invariant Sparsity Measures
This work addresses the need for more efficient neural network pruning for researchers and practitioners in machine learning, though it is incremental as it builds on existing sparsity measures.
The paper tackled the problem of inefficient sparsity in neural networks by introducing DeepHoyer, a set of differentiable and scale-invariant sparsity-inducing regularizers, which produced sparser models than previous methods while maintaining the same accuracy level.
In seeking for sparse and efficient neural network models, many previous works investigated on enforcing L1 or L0 regularizers to encourage weight sparsity during training. The L0 regularizer measures the parameter sparsity directly and is invariant to the scaling of parameter values, but it cannot provide useful gradients, and therefore requires complex optimization techniques. The L1 regularizer is almost everywhere differentiable and can be easily optimized with gradient descent. Yet it is not scale-invariant, causing the same shrinking rate to all parameters, which is inefficient in increasing sparsity. Inspired by the Hoyer measure (the ratio between L1 and L2 norms) used in traditional compressed sensing problems, we present DeepHoyer, a set of sparsity-inducing regularizers that are both differentiable almost everywhere and scale-invariant. Our experiments show that enforcing DeepHoyer regularizers can produce even sparser neural network models than previous works, under the same accuracy level. We also show that DeepHoyer can be applied to both element-wise and structural pruning.