a-t-0/snncompare issues and pull requests

#214 - Write method to store and load pickle of results per exp config.

Issue - State: closed - Opened by a-t-0 about 1 year ago

#213 - Increase neuron death robustness plot resolution.

Issue - State: open - Opened by a-t-0 about 1 year ago

#212 - Dead neuron is Dark red.

Issue - State: open - Opened by a-t-0 about 1 year ago

#211 - Include an option to add permanent radiation effect visualisation in `plot_config`.

Issue - State: open - Opened by a-t-0 about 1 year ago

#210 - Visualisation

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#209 - Synaptic radiation

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#208 - Select graph to visualise

Issue - State: closed - Opened by a-t-0 over 1 year ago

#207 - 1. Convert the Rad_damage object into Synaptic_rad object.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#206 - Compute the statistical significance of the adaptation mechansim results.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 3 comments

#205 - Throw an error if a run_config unique ID occurs twice in the same table. (This is to ensure only the list of neurons that deviate, in the first timestep of deviation, are

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#204 - Then allow the failure mode plot to show all run configs (seeds and graph sizes) in 1 table. By changing the dropdown to a "selection box" (press x to remove value).

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#203 - Include a check to determine whether more seeds are requested than unique graphs exist.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#202 - Fix table headers

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#201 - Include legend in table overview.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#200 - Verify table content.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#199 - Change breadth first to depth first in generating table data.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#198 - Automatically store failure modes

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#197 - Change degree_receiver vertical spacing.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#196 - Improve visualisation

Issue - State: closed - Opened by a-t-0 over 1 year ago

#195 - Prevent terminator from incorrectly terminating algorithm.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#194 - Ensure failure modes are identical over identical run_configs.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#193 - Document failure modes.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#192 - Improve neuron death simulation approach.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#191 - Prevent graph size 5, graph nr 1 to be the same graph as graph nr 0.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#190 - Weigh the boxplot dots.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#189 - Visualise boxplot results more explicitly.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#188 - Visualise space-, and energy complexity of adaptation mechanisms.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#187 - Improve runtime by storing pseudo-random input graphs.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#186 - Verify the random neurons are also created for redundant population coding neurons.

Issue - State: open - Opened by a-t-0 over 1 year ago

#185 - Excitatory radiation

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#184 - Not done, the random spiking neurons are not able to give a counter weight at the same time a synapse (that is supposed to be "dead") spikes. This is due to the delay of at least 1. TODO: optionally apply this.

Issue - State: closed - Opened by a-t-0 over 1 year ago
Labels: wontfix

#183 - `v` - Not done. Random spike neuron is used to generate the spikes, and these do not directly affect the voltage. They only directly affect the current. TODO: optionally, by modifying the voltage within the simulation directly with the accompanying amplitude increase/decrease at times `[t_0, t_1, ..., t_n]`.

Issue - State: closed - Opened by a-t-0 over 1 year ago
Labels: wontfix

#182 - Make stage 4 does not output results if they already exist (output!=compute).

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#181 - Support large graphs

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#180 - Optimise stage 5 boxplot result generation.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#179 - Allow running without outputting stage 2 simulation data.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#178 - assert stage1 is completed.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#177 - Spike once output into degree_receiver nodes

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#176 - Setup evolutionary/genetic approach

Issue - State: closed - Opened by a-t-0 over 1 year ago
Labels: wontfix

#175 - Include shortest route algorithm.

Issue - State: closed - Opened by a-t-0 over 1 year ago
Labels: wontfix

#174 - Include Population coding.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#173 - Include excitatory and other radiation effects.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#172 - Include mapping of seeds to hashses.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#171 - Allow debugging and determining why redundancy 4 has identical failure numbers as redundancy 2.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#170 - SVG optimisation

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix, efficiency

#169 - Allow exporting run times per parameter setting to estimate total exp_config runtime (in seconds) (not timecomplexity).

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#168 - Only store results.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix, efficiency

#167 - Restore multimeter data from stage data loading.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#166 - Write test to verify results nx_simulator and simsnn simulator results are identical

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#165 - Redefine output structure, or Only store results (Optionally, also allow for spike counts)

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#164 - simsnn: spikes if threshold is reached instead of exceeded.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#163 - Support SIMSNN

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#162 - Boxplot run output

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#161 - Change next round to spike after a summed a_in =nr_of_nodes, whilst allowing redundancy delays.

Issue - State: closed - Opened by a-t-0 over 1 year ago
Labels: wontfix

#160 - Find new redundant next_round neuron properties

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#159 - test_redundancy_1.py error.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#158 - Modify test to make n-level redundancy neuron die at once.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#157 - Write test that asserts 2 sequential neurons of different types can die, and still work succesfully.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment
Labels: wontfix

#156 - Expand redundancy test to run on all snn graphs in range 3 to 5.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#155 - Find neuron that can add +n to a property to fire n timesteps after receiving an input signal.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#154 - Ensure redundant neurons do not fire simultaneously with original neuron.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#153 - Determine why hovertext selector_1 is not shown, and make it show.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#152 - Ensure redundant node labels are also shown.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#151 - Allow user to chose to show extra text, e.g. the node current etc. in dash plot.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#150 - Allow user to pass which graph is being visualised to dash.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#149 - Remove opt args

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#148 - Write tests that verify adaptation behaviour on MDSA algorithm.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 3 comments

#147 - Improve degree_receiver y-positions.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#146 - Allow reverse argument

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#145 - Allow live view.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#144 - Improve image visualisation speed

Issue - State: closed - Opened by a-t-0 over 1 year ago

#143 - Verify why redundant selectors don't take over from died primary selector.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 5 comments

#142 - Create a separate method that stores the relevant boxplot data for stage 5.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#141 - Improve structure

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#140 - Efficiency

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#139 - Publish own showme.time pip package to show runtimes.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#138 - Improve structure

Pull Request - State: closed - Opened by a-t-0 over 1 year ago

#137 - Determine why redundancy in box plots is not consistently better for all radiation death percentages.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#136 - Generate random input graphs

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#135 - Support stage 5

Issue - State: closed - Opened by a-t-0 over 1 year ago

#134 - Change CLI usage.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 2 comments

#133 - Find correct SEU rate as function of nr of bits (SER)

Issue - State: open - Opened by a-t-0 over 1 year ago - 1 comment

#132 - Implement correct radiation damage probability.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#131 - Put different levels of redundancy on x-axis, instead of just one adapted graph.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#130 - Change loading all json files for result parsing into loading the json files one at a time.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#129 - Remove determine_what_to_run

Issue - State: closed - Opened by a-t-0 over 1 year ago

#128 - Determine if max_max_graphs can be removed because iterations is removed and replaced with seeds.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#127 - Redundant degree_receiver nodes don't have recursive inhibitory loop.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#126 - Create boxplots

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#125 - Ensure overwrite images command works.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#124 - Remove optional typing from exp_config and run_config.

Issue - State: closed - Opened by a-t-0 over 1 year ago - 1 comment

#123 - Remove unneeded type tests.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#122 - Convert Exp_config attributes that are typed dicts, into objects.

Issue - State: open - Opened by a-t-0 over 1 year ago

#121 - Cleanup after exp_config Dict to obj conversion.

Issue - State: closed - Opened by a-t-0 over 1 year ago

#120 - Fix 2 bugs on completion status.

Issue - State: closed - Opened by a-t-0 over 1 year ago

GitHub / a-t-0/snncompare issues and pull requests