So my horsepower is doing experiments for evaluating performance of deletion mutation plus my parsimony pressure method. But well, there are still free cores.

So I'm doing other experiments with those. It's 3x400 demes. The parameters are the best found in experiments. Not all parameters are finetuned, yet. As I have extended the negative dataset to 8 words, the fitness doesn't excel like before. And of course, as I have increased the population(4x), it'll take a much longer time for it to converge.

Let's see how it'll go.

#GeneticProgramming #GP #EvolutionaryML #EvolutionaryMachineLearning #wakegp

I just finished the 1st episode of the last #BlackMirror season. It hurt my feelings. And it feels very related to our current situation.

That's what would happen if we let corporations interested in maximum profit be the sole owners of the tech, and then control us.

And they are already doing so. We don't have a #FOSS #searchengine which works as good as #Google or #DuckDuckGo. Such these examples are very common.

And #AI/#ML techs are in hand of big techs, suited for them and not for the people. Of course, they publish the papers as #openscience. But what use is that when their models are built for expensive GPUs which a small company can't afford. Let alone common people like me.

Some people believe the way to fight these is regulation. But I highly doubt it. You can always find holes in the law. That's what lawyers commonly do.

But if we work on #opensource, on #opensourceai and on #OpenCulture, we can have a FOSS and #decentralized search engine which performs as good as the current mainstream search engines. Then Google will be automatically removed.

So next time someone asks me why I am a FOSS and #decentralization advocate, and why I so persistently pursue my goal of "AI for people"(and #wakegp as part of it), I'm gonna tell them watch few Black Mirror episodes so that they may understand which path chose the world has chosen for us, unless we choose a different path ourselves.

#blackmirrorseason7 #capitalism #opensource_software

So the new round of #wakegp experiments have started. I'm experimenting different values for deletion mutation. And all my 24 slaves are now working hard to deliver fruits for me.

And CPU temperature being over 70 means they have turned the AC off :)

#geneticprogramming #cpuasheatgenerator #cpuasslave #computerasslave #dailynote

The more #sh scripting I do, the more I love it and wondering why I didn't learn it earlier. So for #wakegp now I have done this to have 256 runs for different values of deletion_mutation_rate and deletions_per_mutation:

➜  runs5 git:(main) for i in {0..255}; do 
for> for size in nanod minid microd halfd thirdd majord; do
for for> for r in 1 2 4 8 16; do
for for for> w_run p4_"$size"_"$r" $i;
for for for> done
for for> done
for> done

w_run itself is a function I defined in my .zshrc

But of course, the syntax is hard. And it's not as easy to learn as something like Python. I wonder if #unixlike operating systems such as #Linux and #BSD would consider quitting the current sh in favor of something new designed from scratch.

Last time I remember, in the list of projects #NLnet had founded, there was a niche new shell invented for unix like systems. I'm gonna check it out. Also #OpenBSD people have something for themselves.

Edit: I had forgotten to add echo -n "$i "; date +%T

#FreeBSD #tech #geek #shellscripting #scripting #bash_scripting #shell_scripting #programming

Alright another round of experiments for #wakegp has ended. This round was for determining if adding the conditional SelectP instruction to the set will have any effect on both program size and fitness.

Another round of experiments which is still going on, is determining how deletion mutation affects both program size and fitness, that is prevents bloat. Earlier experiments show that the simple parsimony pressure method I invented, has positive effect on both fitness and program size, but when there is also deletion mutation, the effect is even better. In those experiments, deleting an instruction randomly from all programs(that is a rate of 1.0) has positive impact on both fitness and average program size. Now I'm experimenting to see if more than one deletion per program would be even better.

#GeneticProgramming #GP #EvolutionaryML #ML #AI #MachineLearning #evolutionarycomputation #EC #LinearGeneticProgramming #wakeworddetection #wake_word_detection #optimization #opensourceML #opensource #FOSS #opensourceAI

So as I am doing experiments for finetuning parameters for #wakegp, I noticed the repository is taking a lot of space, more than 0.5GB. I asked in #codeberg #Matrix room. As they have limits.

One of their people did git gc and that ~700MB turned into under 200MB. I learned a lesson here. You can use git gc for housekeeping of your #git repository. It also does compression.

I wonder when could I master git an learn all of its wonderful features.

PS: In non profit services like codeberg, the resources you use aren't paid with your data. Then you have to be careful to not waste their resources. If you've got a huge repository on codeberg, it would be a good idea to do a git gc so the resources would be available for other users.

@Codeberg
#programming #development #softwaredevelopment #softwareengineering #controlversionsystem #cvs #techtips #programmingtips

How do you feel when in a group of crazy people, you are "the madman"?

There are multi levels of geekness. Like there is someone here trying to do optimization like hell:

https://codeberg.org/farooqkz/wakegp

#geek #wakegp #optimization #nerd #nerdy #geekish #geekful

Alright finished another round of #wakegp tests. It seems my invented method for parsimony pressure is effective and also has positive impact on fitness. This confirms findings or opinions of other researchers that bloat has negative impact on fitness.

Now started another round which will end in ~11h to see the effect of deletion mutation on fitness and program size, if any. Current experiments suggest that deletion mutation has no effect on either, not this side or the other side.

PS: I just hope when I want to publish the paper, not having a bachelor won't be problematic. However I don't have much of a problem about my paper getting rejected. Publishing the paper is a side effect with benefit. Results are my goals.

#EvolutionaryML #EC #AI #ML #MachineLearning #GeneticProgramming #LinearGeneticProgramming
#EvolutionaryMachineLearning #GP #bioinspiredcomputing #BioinspiredML #academia #academic #academics

I like productivity! So I've found the best #FFT size for #Wakegp. Some experiments about the #parsimonypressure method I have done shows effectiveness of my simple method. Now I'm experimenting on that to find out the sweet spot.

Regarding instructions, I already have done experiments. And the results were unexpected. Adding functions like Sine, Cosine, Logarithm and Square root have negative impact on fitness.

There are still a lot of parameters which I need to fine tune. Mutation rates, number of demes, tournment size, reproduction rate, crossover rate and a lot more. And I also need to experiment with #LexicaseSelection

After all these and when I have runs producing accurate enough programs, I should learn how to optimize the found programs to "summarize" them. I really hope compilers like #LLVM and #GCC already do so. I have doubts but I think most likely, they already have heuristics to summarize programs while keeping the logic.

God willing, when #repebble(the new #pebble #smartwatch) comes to market, I could try to see if I can use it there. Then #repebble would be the first industry application of my research.

cc @lspector

Edit: Regarding the simple parsimony pressure method I invented, I'm pretty sure already someone has invented it but I haven't found it in the literature. Just like tons of other things I invent or discover in Math or CS and then realize it has already been discovered or invented.

#MachineLearning #EvolutionaryMachineLearning #EvolutionaryML #ML #ArtificialIntelligence #AI #wakeworddetection #wake_word_detection #hotworddetection
#GeneticProgramming #LinearGeneticProgramming #MachineLearning

I love #unixphilosophy

➜  runs5 git:(main) ✗ function f1() { wakegp -c /var/tmp/$1.toml > $1_$2; }
➜  runs5 git:(main) ✗ sleep 130m; for i in {0..64} ; do for j in 2k 1.2k 1.4k 1.5k 1k 1.6k 1.8k ; do f1 $j_800 $i; done; done

Scheduling #wakegp experiments with #sh so tomorrow I have experiments done and I can just do T tests in R.

Edit: I'm testing different #FFT sizes. If you are interested, see my current status here:

https://codeberg.org/farooqkz/wakegp/src/branch/main/experiments/fft_size

Edit2: So it's morning and I realized I had to do f1 $i $j_800...

#unix #unixlike #linux #foss #opensource #freesoftware #unixclone #programming #tech #geek

My CPU is producing a lot of heat in recent days. Most of the time I'm doing #wakegp experiments to fine tune parameters. And I'm also compiling various versions of #Luanti to find out or fight with regressions and report them upstream.

I'm just glad I invested into buying this CPU plus a decent AIO. I can compile debug enabled Luanti from scratch in just about 5 minutes. Interestingly, compiling a debug enabled wakegp with incremental builds takes at least the same amount or even more.

wakegp is written in Rust, Luanti is written in C++. Luanti is a far larger codebase. But I guess the problem's that many of Luanti deps are shared libraries so I don't have to do the compiling or linking for them.

#Rust #programming #dailyNote #cplusplus #compiling #heavycomputation #computation

Unfortunately, my horse power computer is now offline. And I am in another city so no access to it. It's likely that there has been a #poweroutage which is now very common in my country #iran

I am doing experiments with #wakegp to see if my simple method for parsimony pressure is effective. Till now, it seems that its effect is very little. I'm thinking of other methods for parsimony pressure such as bucketing and tournament selection(for size instead of fitness).

I am expecting to deliver results in summer, God willing.

#Geneticprogramming #evolutionarycomputation #EvolutionaryComputing #artificialintelligence #machinelearning #ml #ai #wake_word_detection #wakeworddetection #programming #computer_science #cs #computerscience

Last few days I've spend most of my hours playing #Luanti on #JMA #CTF server. Having all 24 CPU cores busy for #GeneticProgramming (for #wakegp) runs has left me a good excuse to spend time playing :)

The route humanity has chosen about #AI is a dangerous one. When my #wakegp research is done, I will write a post about how AI must be if it wants to serve people and not big heads. Now software freedom is not enough anymore. I hope as I move forward, I find other likeminded people to join me in my R&D projects. In the last 2 years it has been just me.

#ML #machinelearning #artificialintelligence #free_software #opensource #opensourceAI #research