learning: cosa cambia ai, machine learning & deep · outline history of ais introduction to...

28
AI, Machine Learning & Deep Learning: cosa cambia By Antonio Spadaro PyCon Italia 2017

Upload: phamthuan

Post on 31-Aug-2018

220 views

Category:

Documents


0 download

TRANSCRIPT

AI, Machine Learning & Deep Learning: cosa cambiaBy Antonio Spadaro

PyCon Italia 2017

Outline● History of AIs● Introduction to artificial intelligence● Introduction to machine learning● Introduction to deep learning● Understand why use Python● Q&A

Artificial intelligences

Artificial intelligence is a machine that mimics a "cognitive" function of human mind

The birth of AI: 1956

At the Dartmouth workshop of 1956's, Allen Newell, Herbert A. Simon and Cliff Shaw

debuted with Logic Theorist

https://goo.gl/W3YUvy

Artificial intelligence

Overview

INPUT

AI ALGORITHM

OUTPUT

Example: 2048 AI

Machine Learning

Machine learning explores the study and construction of algorithms that can learn from

and make predictions on data

Overview

TRAINING DATA

(UN)SUPERVISED ALGORITHM

MODEL

TEST DATA

MODEL

OUTPUT

TRAIN THE MODEL TEST IT!

Supervised learning

Supervised learning is the Data mining task of inferring a function from labeled training data.

Overview

Unsupervised learning

Unsupervised machine learning is the machine learning task of inferring a function to describe

hidden structure from "unlabeled" data

Overview

Neural network

Artificial neural networks are a computational model based on

simple neural units

ML applications

Pratical example!Pear Apple

Deep learning

Deep learning is a neural network with one or more hidden layers

Pratica example!

VS DOG

CAT

What does it has to do with Python?

Python is a gold mine for ML

● Python isn’t the fastest but it’s fast!● Python isn’t the easiest but it’s easy!● So, it has the best easiness/speed ratio!● C integration is awesome!● Python ecosystem is fantastic!

Python libraries

OpenAI Gym

“A toolkit for developing and comparing reinforcement learning algorithms. It supports

teaching agents everything from walking to playing games like Pong or Go”

Some books!

Have a good machine learning develop!

@[email protected]

...and now Q&A TIME!