steffen/self-replicating-neural-networks
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

readme updated

Browse Source
This commit is contained in:
steffen-illium
2021-06-04 15:01:16 +02:00
parent 61ae8c2ee5
commit b57d3d32fd
3 changed files with 11 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
README.md
View File

@ -1,6 +1,11 @@
# self-rep NN paper - ALIFE journal edition # self-rep NN paper - ALIFE journal edition
- [x] Plateau / Pillar sizeWhat does happen to the fixpoints after noise introduction and retraining?Options beeing: Same Fixpoint, Similar Fixpoint (Basin), Different Fixpoint? Do they do the clustering thingy? - [x] Plateau / Pillar sizeWhat does happen to the fixpoints after noise introduction and retraining?Options beeing: Same Fixpoint, Similar Fixpoint (Basin),
- Different Fixpoint?
Yes, we did not found same (10-5)
- Do they do the clustering thingy?
Kind of: Small movement towards (MIM-Distance getting smaller) parent fixpoint.
Small movement for everyone? -> Distribution
- see `journal_basins.py` for the "train -> spawn with noise -> train again and see where they end up" functionality. Apply noise follows the `vary` function that was used in the paper robustness test with `+- prng() * eps`. Change if desired. - see `journal_basins.py` for the "train -> spawn with noise -> train again and see where they end up" functionality. Apply noise follows the `vary` function that was used in the paper robustness test with `+- prng() * eps`. Change if desired.
@ -9,6 +14,9 @@
- [ ] Same Thing with Soup interactionWe would expect the same behaviour...Influence of interaction with near and far away particles. - [ ] Same Thing with Soup interactionWe would expect the same behaviour...Influence of interaction with near and far away particles.
- [ ] How are basins / "attractor areas" shaped?
- Weired.... tbc...
- [x] Robustness test with a trained NetworkTraining for high quality fixpoints, compare with the "perfect" fixpoint. Average Loss per application step - [x] Robustness test with a trained NetworkTraining for high quality fixpoints, compare with the "perfect" fixpoint. Average Loss per application step
- see `journal_robustness.py` for robustness test modeled after cristians robustness-exp (with the exeption that we put noise on the weights). Has `synthetic` bool to switch to hand-modeled perfect fixpoint instead of naturally trained ones. - see `journal_robustness.py` for robustness test modeled after cristians robustness-exp (with the exeption that we put noise on the weights). Has `synthetic` bool to switch to hand-modeled perfect fixpoint instead of naturally trained ones.
@ -19,7 +27,7 @@
- [ ] Adjust Self Training so that it favors second order fixpoints-> Second order test implementation (?) - [ ] Adjust Self Training so that it favors second order fixpoints-> Second order test implementation (?)
- [ ] Barplot over clones -> how many become a fixpoint cs how many diverge per noise level - [x] Barplot over clones -> how many become a fixpoint cs how many diverge per noise level
- [ ] Box-Plot of Avg. Distance of clones from parent - [ ] Box-Plot of Avg. Distance of clones from parent

3
experiments/self_train_exp.py
View File

@ -55,8 +55,6 @@ class SelfTrainExperiment:
net = Net(self.net_input_size, self.net_hidden_size, self.net_out_size, net_name) net = Net(self.net_input_size, self.net_hidden_size, self.net_out_size, net_name)
for _ in range(self.epochs): for _ in range(self.epochs):
input_data = net.input_weight_matrix()
target_data = net.create_target_weights(input_data)
net.self_train(1, self.log_step_size, self.net_learning_rate) net.self_train(1, self.log_step_size, self.net_learning_rate)
print(f"\nLast weight matrix (epoch: {self.epochs}):\n{net.input_weight_matrix()}\nLossHistory: {net.loss_history[-10:]}") print(f"\nLast weight matrix (epoch: {self.epochs}):\n{net.input_weight_matrix()}\nLossHistory: {net.loss_history[-10:]}")
@ -113,5 +111,6 @@ def run_ST_experiment(population_size, batch_size, net_input_size, net_hidden_si
summary_fixpoint_experiment(runs, population_size, epochs, experiments, net_learning_rate, summary_directory_name, summary_fixpoint_experiment(runs, population_size, epochs, experiments, net_learning_rate, summary_directory_name,
summary_pre_title) summary_pre_title)
if __name__ == '__main__': if __name__ == '__main__':
raise NotImplementedError('Test this here!!!') raise NotImplementedError('Test this here!!!')

1
journal_robustness.py
View File

@ -195,7 +195,6 @@ class RobustnessComparisonExperiment:
print(f"\nTime as fixpoint: ") print(f"\nTime as fixpoint: ")
# print(tabulate(time_as_fixpoint, showindex=row_headers, headers=col_headers, tablefmt='orgtbl')) # print(tabulate(time_as_fixpoint, showindex=row_headers, headers=col_headers, tablefmt='orgtbl'))
return time_as_fixpoint, time_to_vergence return time_as_fixpoint, time_to_vergence
def count_fixpoints(self): def count_fixpoints(self):