Contents

List of Illustrations

Preface

Introduction

I    Understanding Creativity

  1   What Makes Us Creative?

Einstein, Bach, Picasso: What Makes These People Special?

  2   Seven Hallmarks of Creativity and Two Marks of Genius

1. The Need for Introspection

2. Know Your Strengths

3. Focus, Persevere, and Don’t Be Afraid to Make Mistakes

4. Collaborate and Compete

5. Beg, Borrow, or Steal Great Ideas

6. Thrive on Ambiguity

7. The Need for Experience and Suffering

The Two Marks of Genius

Intent, Imagination, and Unpredictability

  3   Margaret Boden’s Three Types of Creativity

  4   Unconscious Thought: The Key Ingredient

The Four Stages of Creativity

The Importance of Taking Time off

Unconscious Thought and Computers

  5   The Birth of Artificial Intelligence

The First Inklings of Computer Creativity

Computers That Mimic the Brain

  6   Games Computers Play

Deep Blue Defeats Garry Kasparov

IBM Watson Becomes Jeopardy! Champion

AlphaGo Defeats the Reigning World Go Champion

II   Portrait of the Computer as an Artist

  7   DeepDream: How Alexander Mordvintsev Excavated the Computer’s Hidden Layers

Mike Tyka Takes the Dream Deeper

  8   Blaise Agüera y Arcas Brings Together Artists and Machine Intelligence

Memo Akten Educates a Neural Network

  9   What Came after DeepDream?

Damien Henry and a Machine That Dreams a Landscape

Mario Klingemann and His X Degrees of Separation

Angelo Semeraro’s Recognition: Intertwining Past and Present

Leon Gatys’s Style Transfer: Photography “In the Style Of”

10   Ian Goodfellow’s Generative Adversarial Networks: AI Learns to Imagine

Mike Tyka’s Portraits of Imaginary People

Refik Anadol Creates a Dreaming Archive

Theresa Reimann-Dubbers’s AI Looks at the Messiah

Jake Elwes’s Dreams of Latent Space

11   Phillip Isola’s Pix2Pix: Filling in the Picture

Mario Klingemann Changes Faces with Pix2Pix

Anna Ridler’s Fall of the House of Usher

12   Jun-Yan Zhu’s CycleGAN Turns Horses into Zebras

Mario Klingemann Plays with CycleGAN

13   Ahmed Elgammal’s Creative Adversarial Networks

14   “But Is It Art?”: GANs Enter the Art Market

15   Simon Colton’s The Painting Fool

16   Hod Lipson and Patrick Tresset’s Artist Robots

III  Machines That Make Music: Putting the “Rhythm” into “Algorithm”

17   Project Magenta: AI Creates Its Own Music

18   From WaveNet and NSynth to Coconet: Adventures in Music Making

WaveNet: From Voice to Music

NSynth—Creating Sounds Never Heard Before

Coconet: Filling in the Gaps

19   François Pachet and His Computers That Improvise and Compose Songs

The Flow Machine

20   Gil Weinberg and Mason Bretan and Their Robot Jazz Band

21   David Cope Makes Music That Is “More Bach than Bach”

22   “The Drunken Pint” and Other Folk Music Composed by Bob Sturm and Oded Ben-Tal’s AI

23   Rebecca Fiebrink Uses Movement to Generate Sound

24   Marwaread Mary Farbood Sketches Music

25   Eduardo Miranda and His Improvising Slime Mold

IV  Once Upon a Time: Computers That Weave Magic with Words

26   The Pinocchio Effect

27   The Final Frontier: Computers with a Sense of Humor

28   AI and Poetry

Pablo Gervás and His Poetic Algorithms

29   Rafael Pérez y Pérez and the Problems of Creating Rounded Stories

30   Nick Montfort Makes Poetry with Pi

31   Allison Parrish Sends Probes into Semantic Space

32   Ross Goodwin and the First AI-Scripted Movie

33   Sarah Harmon Uses AI to Create Illuminating Metaphors

34   Tony Veale and His Metaphor- and Story-Generating Programs

35   Hannah Davis Turns Words into Music

36   Simon Colton’s Poetic Fool

 V  Staged by Android Lloyd Webber and Friends

37   The World’s First Computer-Composed Musical: Beyond the Fence

VI  Can Computers Be Creative?

38   A Glimpse of the Future?

Creativity in Humans and Machines

39   What Goes On in the Computer’s Brain?

Jason Yosinski and the Puzzle of What Machines See

Mark Riedl on Teaching Neural Networks to Communicate

40   What Drives Creativity?

Margaret Boden and Computer Creativity

41   Evaluating Creativity in Computers

Geraint Wiggins and the Mind’s Chorus

Graeme Ritchie’s Mathematical Criteria for Measuring the Creativity of a Computer Program

Anna Jordanous’s Fourteen Components of Creativity

42   Computers with Feelings

Rosalind Picard on Developing Machines That Feel

Machines Gaining Experience of the World

Machines That Suffer

43   The Question of Consciousness

John Searle’s Chinese Room and the Question of Whether Computers Can Actually Think

Reducing Consciousness to the Sum of Its Parts

44   Michael Graziano: Developing Conscious Computers

Awareness and Attention

Self-Awareness, Introspection, and Perseverance in Computers

Giving Computers Consciousness

45   Two Dissenting Voices

Douglas Hofstadter and the Horrors of a Future Controlled by Creative Machines

Pat Langley and Machines That Work More like People

46   Can We Apply the Hallmarks of Creativity to Computers?

The Need to Know Your Strengths

The Need to Beg, Borrow, or Steal Great Ideas, and the Need for Collaboration and Competition

The Need to Focus and Not Be Afraid to Make Mistakes

The Need to Thrive on Ambiguity and the Need for Experience and Suffering

The Ability to Discover the Key Problem and to Spot Connections

47   The Future

Where We Are Now

Where We Are Going

And into the Future

Acknowledgments

Illustration Credits

Bibliography

Index

List of Figures

Figure 7.1 Mordvintsev’s reference image of a cat and a beagle, 2015.

Figure 7.2 Alexander Mordvintsev, nightmare beast created using DeepDream, 2015. [See color plate 1.]

Figure 7.3 “Dogslug cum puppyslug” created using DeepDream, 2015.

Figure 7.4 Two ibises—­the original photograph (left) and as “seen” by a machine, 2015 (right).

Figure 7.5 Picture of clouds before (top) and after (bottom) being transformed by DeepDream, 2015.

Figure 8.1 All Watched Over by Machines of Loving Grace: DeepDream Edition, 2015.

Figure 9.1 Mario Klingmann, Mona Lisa transformed by DeepDream, 2016.

Figure 9.2 A pairing by Recognition of a 2016 photograph with a painting from 1660, 2017.

Figure 9.3 The content image, a photograph of the Neckarfront at Tübingen (upper left). The style image, van Gogh’s The Starry Night (lower left). The image on the right results from combining the style and content images to make The Neckarfront at Tübingen according to van Gogh, as van Gogh might have painted it. [See color plate 2.]

Figure 10.1 One of Mike Tyka’s Portraits of Imaginary People, Ars Electronica, 2017.

Figure 10.2 Another of Mike Tyka’s Portraits of Imaginary People, Ars Electronica, 2017.

Figure 10.3 Theresa Reimann-­Dubbers, A(.I.) Messianic Window, 2017.

Figure 10.4 A machine dreams. Jake Elwes, Latent Space, 2017.

Figure 11.1 Chris Hesse, edges2cats, 2017.

Figure 11.2 Chris Hesse, edges2handbags, 2017.

Figure 11.3 Mario Klingemann, Transhancement Sketch, 2017.

Figure 11.4 Mario Klingemann, Transhancement Sketch, 2017.

Figure 11.5 Mario Klingemann, Neurographic Self-­Portrait, 2017.

Figure 12.1 CycleGAN translates from a Monet to a photograph-­like landscape (top, upper row) and from a photograph to a Monet-­like painting (top, lower row), and from a zebra to a horse (bottom, upper row) and from a horse to a zebra (bottom, lower row). [See color plate 3.]

Figure 12.2 Striped Putin astride a zebra, created by Jun-­Yan Zhu using a CycleGAN, 2017.

Figure 12.3 Mario Klingemann, computer-­generated images created using Pix2Pix and CycleGAN. The original black-­and-­white photograph (upper left). The same photograph, processed by a CycleGAN trained on 1920s images of men and recent, colored selfies of women (upper right). Previous image, fed into a second CycleGAN trained to turn human faces into doll faces (lower left). The image in turn is processed by a Pix2Pix trained on “glitch images” to build in the element of decay, 2017 (lower right).

Figure 13.1 Images created by a CAN with the style ambiguity function turned off, 2017.

Figure 13.2 Images created by a CAN with the style ambiguity function turned on, 2017.

Figure 14.1 Portrait of Edmond de Belamy, 2018.

Figure 15.1 Happy face (left). As rendered by The Painting Fool, in vibrant pastels, with an “electric” background, 2015 (right). [See color plate 4.]

Figure 15.2 Sad face (left). As rendered by The Painting Fool, in muted colors, with a dull background, 2015 (right).

Figure 16.1 Tear, inspired by Roy Lichtenstein’s Frightened Girl. Lipson attributes the work to PIX18 and himself thus: PIX18/Hod Lipson, 2016.

Figure 16.2 Paul sketching the author, 2017.

Figure 20.1 Haile, the drum-­playing robot, 2006.

Figure 20.2 Mason Bretan jamming with Shimon (left) and three Shimis (right), 2015.

Figure 32.1 Albrecht Dürer, Saint Jerome in His Study, 1514.

Figure 37.1 Playbill for Beyond the Fence, February 2016.