Une IA
Neat plus Ultra

Grégoire Hébert

Lead-Developper & Formateur @ Les-Tilleuls.coop
@gheb_dev

REACTIVE MACHINES (Scenarii reactive)

LIMITED MEMORY (+ aware of changes in an environment)

THEORY OF MIND (+ aware of the people in it)

SELF AWARE (+ capable of self councious choices)

REACTIVE MACHINES (Scenarii reactive)

LIMITED MEMORY (+ aware of changes in an environment)

THEORY OF MIND (+ aware of the people in it)

SELF AWARE (+ capable of self councious choices)

IN

IN

?

IN

?

OUT

IN

?

OUT

PERCEPTRON

?

?

Or NOT

?

0 -> 10

Or NOT

?

0 -> 1

0 -> 10

Or NOT

?

0 -> 1

activation

0 -> 10

Or NOT

?

0 -> 1

activation

  • BinaryStep

  • Gaussian

  • HyperbolicTangent

  • Parametric Rectified Linear Unit

  • Sigmoid (default)

  • Thresholded Rectified Linear Unit

0 -> 10

Or NOT

?

0 -> 1

activation

  • BinaryStep

  • Gaussian

  • HyperbolicTangent

  • Parametric Rectified Linear Unit

  • Sigmoid (default)

  • Thresholded Rectified Linear Unit

0 -> 10

Or NOT

?

0 -> 1

activation

  • BinaryStep

  • Gaussian

  • HyperbolicTangent

  • Parametric Rectified Linear Unit

  • Sigmoid (default)

  • Thresholded Rectified Linear Unit

0 -> 10

Or NOT

?

0.2

activation

  • BinaryStep

  • Gaussian

  • HyperbolicTangent

  • Parametric Rectified Linear Unit

  • Sigmoid (default)

  • Thresholded Rectified Linear Unit

8

Or NOT

?

8

0.2

activation

  • BinaryStep

  • Gaussian

  • HyperbolicTangent

  • Parametric Rectified Linear Unit

  • Sigmoid (default)

  • Thresholded Rectified Linear Unit

x

Or NOT

sigmoid(8x0.2)

Or NOT

sigmoid(1.6)

Or NOT

0.999212971747

?

0.2

2

Or NOT

0.75306590486

0.2

2

0.06

0.285812242

2

?

0.2

TRAINING

2

?

0.2

BACK PROPAGATION

Linear Gradient Descent

-0.7530659

2

?

0.2

BACK PROPAGATION

Linear Gradient Descent

-0.7530659

2

?

BACK PROPAGATION

Linear Gradient Descent

- 0,10

0.1

0,4542164

2

?

BACK PROPAGATION

Linear Gradient Descent

- 0,10

- 0,03

0.07

0,330153

2

?

0,285881

BACK PROPAGATION

Linear Gradient Descent

- 0,10

- 0,03

0.06

- 0,01

2

CONGRATULATIONS !

Hungry

?

EAT

Thirsty

Sleepy

DRINK

SLEEP

Hungry

EAT

Thirsty

Sleepy

DRINK

SLEEP

Hungry

EAT

Thirsty

Sleepy

DRINK

SLEEP

MULTI LAYER

PERCEPTRON

Hungry

EAT

Thirsty

Sleepy

DRINK

SLEEP

MULTI LAYER

PERCEPTRON

Hungry

Thirsty

Sleepy

NEAT

?

EAT

DRINK

SLEEP

NEAT

Neuro Evolution of Augmenting Topologies

Hungry

Thirsty

Sleepy

?

EAT

DRINK

SLEEP

MERCI !

IA NEAT PLUS ULTRA

By Grégoire Hébert

IA NEAT PLUS ULTRA

Lightning talk

  • 1,305