Sentence examples for sparsity of weights from inspiring English sources

Beyond simply learning features by stacking autoencoders (AE), there is a need for increasing its robustness to noise and reinforcing the sparsity of weights to make it easier to discover interesting and prominent features.

Sparsity of weights: We use regularization in our optimization function, which does not produce sparse weight vectors.

Previous NMF methods [11, 13, 21] were developed to extract task-specific nonnegative factors, but they did not simultaneously consider the uncertainty of model parameters and control the sparsity of weight parameters.

Hyperparameters were used to control the sparsity of reconstructed weights and the generation of basis parameters.

Importantly, we control the sparsity of reconstruction weights by using prior density based on the Laplacian scale mixture (LSM) distribution [25].

The novelty in our paper is measurement of sparsity of the weighting coefficients of each training sample.

By imposing group sparsity of pose weight matrix in slice GAC, we interestingly provide a way to connect uniform and non-uniform deblurring.

By using the reweighted l 1-norm constraint, those tap weights whose magnitudes are small are imposed larger weightings in the next iteration and vice versa, and accordingly, the sparsity of the tap weights is enhanced.

where the l 0-norm ∥·∥0 is the count of the number of non-zero elements of its argument, and μ denotes the user parameter to control the degree of sparsity of the tap weights.

Besides, by incorporating group sparsity of the pose weight matrix into slice GAC, the non-uniform deblurring method naturally approaches toward the uniform blind deconvolution.

So, we can change the percentage of nonzero elements relative to the total amount of elements in W res to adjust the sparsity f sparsity of the recurrent weight matrix.