3. Evolutionary Art

• Evolutionary Art represents the convergence of Creative Expression and Evolutionary Computing

• The evolutionary period often produces new forms of art

A Banana by any Other Name

3.1. Evolutionary Artforms

• This is really just “How do humans make art?”

3.1.1. Music & Dance

• Make melodies and full songs

• Develop rhythms and musical structure

• Create sequences of choreography

3.1.2. Paintings & Sculpting

• Attempts to achieve a target solution (TSP? Or SymboReg)

Poly Lisa

• Complete Randomization of art

• Creating 3d Sculptures that can be modeled

3.1.3. Poetry & Shakespeare (or really… any book)

• Attempts to form grammatically correct sentences or follow rules of poetry

Note

Follow the rules of a Haiku:

• Start and end with five,

• Middle line houses seven,

• Structured syllables

3.1.4. Actually, I think math is art 🤓

• Okay

• See Assignment Two

3.2. Making an Evolutionary Program

• What is a population?

• How could you possibly give a fitness score? Isn’t art subjective?

• What does crossover look like?

• Mutation?

• Elites? Like Van Gogh?

Painting Crossover Example?

3.2.1. Encoding Art As Chromosomes

• Visual Art

  • Create nxm pixel lists that can have any colour in any give spot

  • Mathematical functions that make specific shapes

Mathematical Love

• Music

  • Store Musical Components as Variables (Like chords)

  • MIDI representations that can be stored as numeric values

Music to MIDI

Note

For example, Hot cross buns follows:

• 64, 62, 60

• 64, 62, 60

• 60, 60, 60, 60

• 62, 62, 62, 62

• 64, 62, 60

Where 64 = E4, 62 = D4, and 60 = C4

3.2.2. Establishing a Population

• To create a painting, you could simply start with an nxm grid of:

  • A solid colour

  • A random colour for each pixel

• To make a musical sequence, you can:

  • Generate random sequences within music constraints

  • Based on existing songs, steal and splice segments

3.2.3. Defining Fitness

• Symmetry

• Contrast of Colour (Pixel by pixel Basis)

• Human Scoring

• Statistical Analysis (% of colour)

• Using Machine Learning for Score

3.2.4. Mutating Art

• In Visual Art

  • Pixel Colour Mutation

  • Redefining Generative functions

  • Altering the length of 3d Modelling Vectors

• In Music

  • Changing a Note (Pitch, Length, Volume)

  • Swapping an existing musical variable with a random new one

3.2.5. Artistic Crossover

• In Visual Art

  • Swapping segments of pixels (Layers/Regions)

  • Swapping Functions and Preset Drawings

• In Music

  • Changing a Note (Pitch, Length, Volume)

  • Swapping segments of songs

  • Exchanging instrument assignment

3.2.6. Elitism

• Now that everything else is defined, elitism works as it normally would

3.3. Example of Creating a Painting

Making a Dirt Block

3.4. Trends and Relavant Applications

• A surgence in Machine Learning has led to more capability for less objective computation

• Can be applicable to structrual engineering (architecture…) and fashion design