Imagine a simulated anxiety #metaheuristic, similar to simulated annealing, but for a population potential solutions, or partial solutions, that might work together. (That might be a daft part) Wrong answers cause their anxiety to rise, which makes them seek more extraordinary solutions. Right answers make them Muntz themselves, trying to remove extra stuff.πŸ€”

I wanted to find all the line numbers of subway.bas, quick and easy, but some had multiple occurrences in the data.

So, I turned to #metaheuristic methods! I wrote a quick & dirty genetic algorithm to search for sequences of indexes of possible line numbers, where all(d>c for c,d in zip(e,e[1:]))

In only several dozen generations of a population of 10k, not much time at all, it found a solution.πŸ•ΆοΈ Maybe.😬 With minimal thinking! xD #compsci

#HelloESR I'm a Senior/Expert Research Engineer in #Bioinformatics. I've been working as a researcher in hybrid #AI #Algorithmics (i.e. #metaheuristic optimization) for 20 years, but I'm now interested in engineering Semantic Knowledge Graphs stuff into Decision Support Systems & #ExplainableAI #XAI.

I compute Biomedicine against #Cancer at Institut Pasteur.

I'm the #ParadisEO (an optim framework), #Oncodash (tumor board #DSS), and #OntoWeaver (easy creation of #SKG on top of #BioCypher) maintainer.