Memristive network with memory workflow#2
Conversation
| A_II = A[np.ix_(self._I, self._I)] | ||
| # print(matrix_rank(A_II, hermitian=check_symmetric(A_II))) | ||
| A_II_inv = inv(A_II) | ||
| A_II_inv = pinv(A_II) |
There was a problem hiding this comment.
I was getting errors from trying to invert a singular matrix (maybe multiple rows of all zeroes?). So I switched to the pseudoinverse to avoid this problem.
| V = np.zeros((len(Vi), self._n_nodes)) | ||
| V[:, self._I] = np.asarray(Vi) | ||
| V[:, self._E] = Vext | ||
| return V |
There was a problem hiding this comment.
EchoStateNetwork returns a matrix that is size (n_times, n_nodes), while MemristiveReservoir was returning (n_times, n_internal_nodes). This messes up the indexing for extracting readout nodes, for example, because those indexes were created relative to all nodes that exist, not just internal nodes.
| self._Ga = mask(self, np.divide(self.Woff, self.NMSS, | ||
| where=self.NMSS != 0)) # constant | ||
| self._Gb = mask(self, np.divide(self.Won, self.NMSS, | ||
| where=self.NMSS != 0)) # constant |
There was a problem hiding this comment.
I was getting some divide by zero errors and infinites in the resulting matrices because of this. I added a condition to avoid dividing by zero here, and in some other places.
|
|
||
| def memory_capacity(conn, input_nodes, output_nodes, readout_modules=None, | ||
|
|
||
| def memory_capacity(conn, input_nodes, output_nodes, rsn_mapping=None, |
There was a problem hiding this comment.
I changed this to take in the raw mappings and index with the output nodes later. This is because output nodes will sometimes get adjusted in this method when randomly selecting the ground node.
| readout_nodes = np.setdiff1d(readout_nodes, gr_nodes) | ||
|
|
||
| # second dimension should be along the input nodes | ||
| x = np.tile(x, (1, len(input_nodes))) |
There was a problem hiding this comment.
MemristiveReservoir.simulate expects an inputs that is size (n_examples, n_external_nodes). This is different than EchoStateNetwork which only requires the input size second dimension align with the first dimension of w_in.
|
|
||
| # simulate reservoir states; select only output nodes | ||
| rs = network.simulate(ext_input=x)[:,output_nodes] | ||
| rs = network.simulate(x)[:, output_nodes] |
There was a problem hiding this comment.
In MSSNetwork, this input is called Vext not ext_input. I'm not sure if it's better to just avoid explicitly passing a parameter-value pair, or to change the API of MSSNetwork to be the same as EchoStateNetwork. I chose the former for now. Also, there's no ability to choose 'forward' or 'backward' right now. Let me know if this is important to add.
1 participant
Primary change is adjusting the memory workflow structure to work with the parametrization
resname='MSSNetwork'. To achieve this, some small changes were made to thememory_capacityAPI and the return type ofMemristiveReservoir.simulatewas slightly altered to align more closely withEchoStateNetwork.simulate. Also, some small stability changes inMSSNetworkwere made.Sample output from running the workflow is shown below.