< Back to Article

Sparse Coding Models Can Exhibit Decreasing Sparseness while Learning Sparse Codes for Natural Images

Figure 2

SAILnet architecture.

In our model, described in detail elsewhere [12], leaky integrate-and-fire neurons receive inputs from pixels in whitened natural images, in a rough approximation of the thalamic input to V1. Inhibitory recurrent connections between neurons, shown in red, act to decorrelate the neuronal activities. The neurons have variable firing thresholds, which are varied by the neurons so as to maintain a desired long-term-average firing rate.

Figure 2