corticalcomputer/book_neuroevolutionthrougherlang

The resulting source code produced at the end of each chapter in Handbook of Neuroevolution Through Erlang.

book_neuroevolutionthrougherlang's Issues

Warning in Ch_18

When trying to make:all([load]). the source code for Ch_18, I get a lot of warnings about unused variables and some methods being deprecated. Since I'm not that good with Erlang yet, I wonder whether I can safely ignore those warnings or are they a sign of an actual problem.

Polises

This is not an issue but more of a question; are there any plans on implementing multiple polises that would work on different machines (real or virtual) and talk to each other? I wonder if it's hard to write.

Ch19 benchmarker:start(SlidingWindow5). variable is unbound

I am try to execute the module in CH19 Forex Trader. When running the following I get a response stating that the * 1: variable 'SlidingWindow5' is unbound. Any help on actually executing this would be appreciated.

c(polis).
polis:sync().
polis:reset().
polis:start().
benchmarker:start(SlidingWindow5).

• 1: variable 'SlidingWindow5' is unbound

Book Errata

Not sure where to put these .. figure a git issue might be a good place. Let me know if you have a preferred way.

• p136 s/SCH/SHC in paragraph 2. "whole point of SCH is..."

Looking up cortex by `cortex` instead of id

When you create the ETS table to store the genotype, you add the {key_pos, 2} option.

https://github.com/CorticalComputer/Book_NeuroevolutionThroughErlang/blob/master/Ch_7/genotype.erl#L96

When we look up the cortex from the table at a later point, we pass in cortex as the key.

https://github.com/CorticalComputer/Book_NeuroevolutionThroughErlang/blob/master/Ch_7/exoself.erl#L23

However, our lookup fails since our table looks like:

We need to change our cortex lookup to:

[Cx] = ets:match_object(Genotype, {cortex, '_', '_', '_', '_'})

Ch_19 questions

This also isn't an actual issue, more a list of questions.

1. The book says

Having set everything up, we execute the benchmark for every noted experimental setup, and run it
to completion. To do this, we simply modify the constraints used in our benchmarker module, and
then execute benchmarker:start(Experiment_Name), for every of our experimental setups.

Just starting a polis and running benchmarker:start(SlidingWindow5). results in * 1: variable 'SlidingWindow5' is unbound. I'm probably being obtuse and not doing something obvious, but could you offer an example of modifying the constraints.

2. Further, the book says

Due to the number of the experiments, and the amount of time that the PCI based experiments take, > particularly the ChartPlane100x10 and ChartPlane50x20 experiments, the benchmarking process will > take up to a week even on a rather powerful quad core sandy bridge CPU.

I don't have such a CPU, but I still wish to run the benchmarker. The obvious solution that occured to me was running the whole thing in a cloud, but I have no real idea how an Erlang simulation would behave there and how it might improve performance and decrease the time necessary for the benchmarking to complete.

3. Even further, the book says

This changes if we perform further analysis of the results of our experiments, and plot the
benchmarker produced GNUplot ready files.

Unless I missed something, the supplementary materials don't provide those files. And having them would be really helpful -- at least it would enable people to graph the data as a larger picture; the figure shown is the book is not large enough and thus is rather confusing.

Thanks in advance.

exoself

Doing as recommended in Chapter 6 just before the summary, I tried
1> exoself:map(ffnn)
and got
<0.36.0> 2> =ERROR REPORT==== 14-Nov-2015::16:14:55 === Error in process <0.36.0> with exit value: {undef,[{sensor,gen,[<0.36.0>,'nonode@nohost'],[]},{exoself,spawn_CerebralUnits,3,[{file,"exoself.erl"},{line,40}]},{exoself,map,2,[{file,"exoself.erl"},{line,23}]}]}
What am I doing wrong?

issue with call for prev(TableName,ets:last(TableName),prev,EndBL)

Book_NeuroevolutionThroughErlang/Ch_19/fx.erl

Line 375 in 81b96e3

prev(TableName,ets:last(TableName),prev,EndBL)

I have been trying for months to get this to work and still nothing any help would be greatly appreciated..... if anyone is even looking at this now..