Engineering and Evolving Software

title = "Engineering and Evolving Software",

booktitle = "2021 IEEE/ACM International Workshop on Genetic Improvement (GI)",

editor = "Justyna Petke and Bobby R. Bruce and Yu Huang and Aymeric Blot and Westley Weimer and W. B. Langdon",

year = "2021",

pages = "10",

month = "30 " # may,

note = "Invited keynote",

keywords = "genetic algorithms, genetic programming, Genetic Improvement, Software, Maintenance engineering, Artificial intelligence, Computer security, Artificial immune systems, US Government",

video_url = "

https://www.youtube.com/watch?v=iZnKHIZ5qVY&list=PLXTjhGKkSnI-se7uQneCX-pEDiDrQ7TIS&index=13",

URL = "

https://twitter.com/gi2021/status/1399027796597383168",

DOI = "

doi:10.1109/GI52543.2021.00008",

size = "1 pages",

abstract = "Provides a brief professional biography of the presenter Stephanie Forrest of Arizona State University. The complete presentation was not made available for publication as part of the conference proceedings.",

notes = "Also known as \cite{9473907}

Video presentation 1 hour

'Nature is a tinkerer, not an inventor.' F.Jacob. But Forrest says evolution closer to skilled engineer and the distinction between design and evolution not so sharp.

Frances Arnold Nobel prize 2018 for chemistry directed evolution (ie Genetic Algorithm) applied to chemistry, eg DNA-SWCNT https://en.wikipedia.org/wiki/Frances_Arnold

Xenobots Mike Levins group at Tufts University https://doi.org/10.1073/pnas.1910837117 Two types of Frog cells => 1mm robot

Lenski E-coli MSU Large jump in evolution.

Large jump in flu virus, eg bird and pig virus interacting.

Engineering innovation like recombination of earlier ideas. NB not clean slate design. Chaos engineering, coevolutionary arms races in software security.

A Jacob test: engineering or evolved?

Evolutionary dynamics in software. There is Darwinian processes in today's software. Software has become more biological: stackoverflow. AB testing (run by genetic algorithm in the back ground, interactive evolution). Continuous integration (rapiarinator). Fuzzing. Chaos Engineering (robust to random changes) Cyber security arms races.

\cite{miikkulainen:2021:nmi}

neutral mutations, mutational robustness needed for evolution. = Test suite equivalent. bug repair, eg GenProg third of mutations are neutral. Why? Many ways it could arise.

\cite{Renzullo:2018:GI}

APR with larger step size

\cite{Liou:2019:GI} Smith-Waterman CUDA 12 highly epistatic mutations

Need more attention to existing EC work, drift, search (eg novelty search), more use of epistasis. More real biology. More hardware resources.

video 46:00 Cost of robustness (eg spell checker => poorer typing) Robustness mechanism becomes essential. Eg IT systems. Challanges

47:37 discussion/Questions chair Bobby R. Bruce

47:37 Myra B. Cohen.

50:57 Giovani Guizzo. A: \cite{miikkulainen:2021:nmi}, drift, what if we had same resouces as deep learning?

52:45 Westley Weimer.

54:58 Alexandre Bergel A: Renzullo IPDS-2021 https://doi.org/10.1109/IPDPS49936.2021.00107 Compensatory evolution, less selection pressure, bigger populations.

57:20 Oliver Krauss

http://geneticimprovementofsoftware.com/events/icse2021.html

2021 IEEE/ACM International Workshop on Genetic Improvement (GI)",