Since the foundation of the AI field in 1956, researchers have been using the human brain as inspiration, and as a final goal to replicate. Recent developments in deep learning made a step forward on tasks that are "easy for a mind, and difficult for a computer" such as pattern recognition in images.
In this talk we will cover the basics of the human brain, and artificial neural networks, and we will see what deep learning is, what it can be used for, and how it relates to the human brain.
Junnasandra Call Girls: 🍓 7737669865 🍓 High Profile Model Escorts | Bangalore...
Replicating the human brain: Deep learning in action
1. Replicating the human brain:
Deep learning in action
Marc Garcia
PyData Mallorca - January 25th, 2017
1 / 25
Replicating the human brain: Deep learning in action
3. The human brain
3 / 25
Replicating the human brain: Deep learning in action
4. The human brain
Visual perception example
4 / 25
Replicating the human brain: Deep learning in action
5. The human brain
Visual perception
5 / 25
Replicating the human brain: Deep learning in action
6. The human brain
Neuron synapse
Neuron doctrine, Santiago Ramón y Cajal
6 / 25
Replicating the human brain: Deep learning in action
7. The human brain
Hubel and Torsten cat experiment
Receptive fields of single neurones in the cat’s striate cortex
David H Hubel and Torsten N Wiesel, 1959, The Journal of physiology
7 / 25
Replicating the human brain: Deep learning in action
8. The human brain
Hebbian theory
Learning The capacity of a neuron to activate another changes over time
Memory This capacity of activation can recall previous activations
Let us assume that the persistence or repetition of a reverberatory activity (or
"trace") tends to induce lasting cellular changes that add to its stability.[...] When
an axon of cell A is near enough to excite a cell B and repeatedly or persistently
takes part in firing it, some growth process or metabolic change takes place in one
or both cells such that A’s efficiency, as one of the cells firing B, is increased.
The Organization of Behavior, Donald Hebb, 1949
8 / 25
Replicating the human brain: Deep learning in action
10. Deep learning: structure
McCulloch-Pitts neuron
x2 w2
Σ f
Activate
function
y
Output
x1 w1
x3 w3
Weights
Bias
b
Inputs
A logical calculus of the ideas immanent in nervous activity
Warren S. McCulloch and Walter Pitts, 1943, Bulletin of Mathematical Biophysics
10 / 25
Replicating the human brain: Deep learning in action
12. Deep learning: structure
Hopfield networks
Neural networks and physical systems with emergent collective computational abilities
John J Hopfield, 1982, Proceedings of the National Academy of Sciences of the USA
12 / 25
Replicating the human brain: Deep learning in action
15. Deep learning: structure
Boltzman machines
Optimal Perceptual Inference
Geoffrey E. Hinton and Terrence J. Sejnowski, 1983
Proceedings of the IEEE conference on Computer Vision and Pattern Recognition
15 / 25
Replicating the human brain: Deep learning in action
16. Deep learning: structure
Restricted Boltzman machines
Information Processing in Dynamical Systems: Foundations of Harmony Theory
Paul Smolensky, 1986, Parallel Distributed Processing, Volume 1, Chapter 6
16 / 25
Replicating the human brain: Deep learning in action
17. Deep learning: structure
Deep belief networks
A fast learning algorithm for deep belief nets
Yee-Whye Teh et al., 2006, Neural computation
17 / 25
Replicating the human brain: Deep learning in action
20. Deep learning: training
Loss function
Error in our predictions, compared to truth
error = y − θ0 + θ1 · x1 (1)
20 / 25
Replicating the human brain: Deep learning in action
23. Deep learning: training
Deep learning in practise
Not for everyone, just makes sense at a scale
Many layers to extract features
Curse of dimensionality
The key is to compute derivatives very fast:
Theano
Tensorflow
Torch
...
23 / 25
Replicating the human brain: Deep learning in action
24. Deep learning: training
Google Youtube experiment
Building High-level Features Using Large Scale Unsupervised Learning
Quoc V. Le, 2012, Proceedings of the 29th International Conference on ML
24 / 25
Replicating the human brain: Deep learning in action