The Future is Data and AI

Chief Data Scientist
Data Science Consulting Business
Strategy, research, training
Mostly R
occasionally Python
Banking, insurance, retail, phamaceutical, genetics, manufacturing
obsessed with data science and R

Intro to Data Science at Columbia
World's largest R Meetup: 8,400 members
NY R Conference
love R so much I wrote a book

Best selling R book
Second edition just released

Statistician

trained as a statistician
Professor Fink
old fashioned math nerd

Data Scientist

now a Data Scientist
much cooler
to illustrate what a data scientist does

Drew Conway's Venn Diagram: Must mention this
math
expertise
hacking

Harvard Business Review
https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century/
DJ Patil, Chief Data Scientist of the United States
events: my company needs a data scientist

Artifical Intelligence

Daft Punk
The new sexy

The new Venn Diagram
From Deep Learning: Goodfellow, Bengio & Courville
Deep Learning is another method of ML

What is AI?

Evolving defition over time

What humans find easy but computers find hard

The senses: seeing, hearing
True AI will be combinations of models and algorithms

Why Now?

so much data available
storage capacity doubles every 10 months

hex
6 computers networked together
2014
sits on my desk

Nvidia dgx-1
960 Teraflops
40,960 CUDA Cores
5,120 Tensor Cores
$150,000

Software availability
1980s hacker movie

glmnet
penalized regression
ridge and lasso
L2 and L1

xgboost
boosted trees
Revolutionized predictions
LightGBM from Microsoft is a competitor

Stan: MCMC
Bayesian
Andrew Gelman
Big Models

we all know why we're here
deep learning
just neural networks
cool, then not cool, then cool, then not cool, then cool again

TensorFlow
From Google
Flowing Tensors

Caffe2
From Facebook

CNTK
By Microsoft

MXNet
By Amazon

nnabla
From Sony
We're all winning from these big companies competing

Keras
Front end to these tools