posted on 2017-08-02, 17:53authored byQuan Wang, Constantin A. Rothkopf, Jochen Triesch
The vertical axis corresponds to the network’s 300 model neurons, and the horizontal axis corresponds to the input sequence element of the total length 20. (A) Joint probability of inputs and neuron firing in SORN with plasticity in the first block of training. (B) Joint probability of inputs and neuron firing in SORN with plasticity in the last block of training. Compared with the left subplot, the joint probability becomes higher with training and the firing of neurons is sparse. (C) Joint probability of inputs and neuron firing in SORN without STDP and IP.