The Department of Biological Engineering (BE) (http://be mit edu) principles of toxicology, in vivo genetic engineering, and molecular biology
We propose deploying these methods to develop a toolkit of machine-learning algorithms which can infer the attributes of genetically engineered organisms—like
Abstract—Synthetic biology, that is, the ability to engineer genetic engineering of DNA parts [20, 56, 39], the other two
Week Nine Reading Guide: Concerns about genetic engineering This week, three students will present 15-minute summaries of their final paper concepts
Genetic Engineering for Biotechnology and Neuoscience * Half semester subjects that together fulfill one biology restricted elective
MIT Design Inquiry How can they be used to extend the capabilities of the designer? ? Creativity through genetic engineering
![[PDF] How do genetic algorithms relate to their biological origins? How do](https://pdfprof.com/EN_PDFV2/Docs/PDF_3/117064_3da5ae6152463537658883d3a1e11215a_johngerolecture.pdf.jpg "[PDF] How do genetic algorithms relate to their biological origins? How do")
117064_3da5ae6152463537658883d3a1e11215a_johngerolecture.pdf MIT Design Inquiry
How do genetic algorithms relate to their
biological origins?
How do they relate to human processes of
design? Why is there power in this metaphor?
How can they be used to extend the
capabilities of the designer?
John Gero
Professor of Design Science
University of Sydney
Visiting Professor of Design and Computation
MIT MIT Design Inquiry
How do genetic algorithms relate to their
biological origins? lSeparation of genetic material (genotype [representation]) from organism (phenotype [design]) lExpressing genotype as organism lOrganism carries genotype and reproduces genotype using 'genetic' processes of crossover and mutation lDarwin's natural selection uses fitnesses of organisms in their environment to improve the gene pool lGA is a simple model of this process MIT Design Inquiry
Genetic processes
Genetic processes
Crossover Points
A C B D
Parents
Offspring
MIT Design Inquiry A AB B BC A BC B AB
Parents
Crossover point
Offspring
MIT Design Inquiry A AB B BC A AC B BB
Parents
Offspring
Crossover point
MIT Design Inquiry
How do they relate to human processes of
design? lCan map genetic representation onto a computational representation of a design; can map phenotype onto a interpretable view of a design lHumans work on single or few designs at a time/ genetics works on a population of 'designs' in parallel lHumans can be seen to "search" design spaces - this is one interpretation of what GAs are doing. MIT Design Inquiry
Why is there power in this metaphor?
lGuaranteed improvement - Darwinian evolution lLarge scale search lBlind search lFitness can be human evaluation lFitness can change over evolutionary time lCan produce complexity lCan produce unexpected results MIT Design Inquiry
How can they be used to extend the capabilities
of the designer? lCreativity through genetic engineering lNovel designs through extending genetic crossover lNovel designs through different fitnesses MIT Design Inquiry
Genetic Engineering and Creative
Design
Genetic Engineering and Creative
Design
l Background lgenes, genotype, phenotype, fitness lConnecting genes to performance in fitness lEmergent gene clusters ˛ evolved genes MIT Design Inquiry • • x x x x x x • • "bad" "good" "good" genotypes "bad" genotypes
Total Population
MIT Design Inquiry aabrule 1aa b rule 2 a a b rule 3 aabrule 4 arule 5 arule 6 ab a rule 7 a b arule 8 b a b a MIT Design Inquiry design 1design 2design 3 design 4design 5design 6 design 7 design 8 design 9design 10 {1,12,2,8,5,4,4,2,8,5,7} good {1,2,1,8,2,8,5,5,6,6,8,1} good {3,2,2,6,5,8,2,1,4,4,3,1} bad {6,4,1,2,8,5,4,2,8,5,3,3} good {3,4,8,2,8,1,6,5,7,3} bad {2,3,2,3,4,3,5,6,5,1,6,2} neutral {3,1,8,5,5,6,4,6,1,1,3,3} good {1,6,4,2,7,3,4,8,6,1,6,2} bad {6,4,1,2,3,4,5,2,1,7,4} neutral {2,3,7,5,1,2,8,3,1,6,2,1} bad
Composite building block A
{2,8,5} MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry
Mondrian
MIT Design Inquiry
Genotype Form
lIn form of a tree lEach node has four variables ldirection of rectangular split (4 values) lfraction of the split (15 values) lcolour of split area (10 values) lline width (3 values) MIT Design Inquiry
Fitnesses for Representation
loffset between actual and required positions of dissection lines lnumber of lines with correct line width, normalised lnumber of correct colour panels, normalised lnumber of lines assigned, normalised lnumber of unassigned lines, normalised
Genetically Engineered Mondrian
Genetically Engineered Mondrian
MIT Design Inquiry
Genetically Engineered Frank Lloyd Wright Windows
MIT Design Inquiry
Flondrians
lMondrian painting ˛ genetically engineered genes: M-genes lFrank Lloyd Wright windows ˛ genetically engineered genes in same representation: F- genes l"Flondrians" are the genetic product of mating
M-genes with F-genes
MIT Design Inquiry MIT Design Inquiry MIT Design Inquiry
How Many Designs Are There and
Where Are They?
MIT Design Inquiry
Genetic crossover as an interpolation
g 1 g 2 g c p c p 2 p 1
Genotypic space G
Phenotypic space P
C(g 1 ,g 2 ) AE g c C(p 1 ,p 2 ) AE p c MIT Design Inquiry P+ P p 1 p 2 MIT Design Inquiry
Interpolation
MIT Design Inquiry MIT Design Inquiry (interactive Genetic Art III) MIT Design Inquiry
Image detection
(a)(b) MIT Design Inquiry
Modelling Interest
Berlyne's model of arousal based on novelty using Wundt curve HED ON IC V A LU E
NOVELTY
N x H x
Reward
Punish
0 1 -1 n 1 n 2 MIT Design Inquiry
Different novelty functions
MIT Design Inquiry
Different novelty preferences
N=0N=1N=2N=3
N=4N=5N=6N=7
N=8N=9N=10N=11
N=12N=13N=14N=15
N=16N=17N=18N=19