For the convolutional connections, we set the kernel size (m × n) to 3 × 1, the number of kernels (k) to 100, and the stride (s) to 1. Here, we consider the first convolutional layer as an example: the size of the input data is 99×× 1, and the first convolutional layer convolves these data with 100 3 × 1 kernels, with a stride (step size) of 1.
With additional load, the rats walked more slowly with larger stride width and reduced step size compared to without additional load (Fig. 5).
(A ) Histogram of walking speeds, divided in 0.05 m/s speed bins for pcd (purple, N = 3 mice, n = 3052 strides), littermate controls (green, N = 7 mice, n = 2256 strides) and size matched controls (black, N = 11 mice, n = 3400 strides).
Then a man in a suede jacket strides in, sizes up the situation, and says something in Polish as he squeezes by me.
(A ) Smoothed probability density of diagonal (FL-HR) pair stance phase lags and speed obtained by kernel density estimation for all strides of size-matched controls (left, n = 3400, N = 11) and pcd (right, n = 3052, N = 3).
(B ) Smoothed probability density of ipsilateral pair (FL-HL) stance phase lags and % stance duration for all strides of size-matched controls (left) and pcd (right), plotted according to the convention of Figure 5 of Hildebrand (1989).
On the left, the stance phases of all of the strides of size-matched controls are shown in a density plot – values around 0 on the y-axis would be a trot, while values above the x-axis would be a walk.
Take the right size strides.
For simplicity, we just show several top layers of BN1 and BN2 in Fig. 3 and the rest layers including the depth, filter size, and stride are same with the proposed NR-Network in Fig. 2.
Table 2 Classification results on reduced STL-10 dataset with overlapping pooling Pooling size Pooling stride Pooling type Unwhitened Whitened ( ε = 0.1) 19 × 19 10 Avg 75.80% 81.34% Max 78.12% 80.22% Ada 77.34% 81.06% 5 Avg 76.06% 81.78% Max 78.56% 80.50% Ada 77.84% 81.25% Figure 8 Classification results on reduced STL-10 dataset with overlapping pooling.
