Neural networks are an important machine learning technology that Google uses to power many of its products. Neural networks learn from large numbers of labeled examples, making tiny adjustments to their models to better match outputs to labels. After training, neural networks can be tested on unlabeled data to produce useful outputs. Google uses neural networks across many products from Search to Translate to Photos. Their use of neural networks is rapidly accelerating to improve these products and tackle new challenges in health, robotics, and more. Google also shares machine learning tools and models to help advance the field.
4. 4
neural net models learn from examples
labeled photos
“cat”
“dog”
“car”
“apple”
“flower”
OUTPUT
5. 5
neural net models learn from examples
Make tiny adjustments to
model so output is closer to
label for a given image
labeled photos
“cat”
“dog”
“car”
“apple”
“flower”
OUTPUT
6. 6
after a model is trained, you can test it
?
unlabeled photo
7. 7
after a model is trained, you can test it
unlabeled photo
“cat”
8. 8
Input Output
“rice”
“restaurants in Seoul”
“hello!”
“A close up of a small child holding a stuffed animal.”
powerful functions that neural nets can learn
안녕하세요
11. Atari 2600 testbed: 100+ Atari games from the 70/80s
Agents just get raw pixels as inputs (~30K)
Wired up to action buttons but NOT told what they do
Goal is simply to maximize score
Everything learnt from scratch, ZERO pre-programmed knowledge
ONE agent with ONE set of parameters to play ALL the different games
THINK OF A ROBOT in the arcade watching the screen controlling the joystick
DQN Deep Q-network- Atari testbed
Image: Gnome Enterprises
18. 18
signal
for Search ranking,
out of hundreds
improvement
to ranking quality
in 2+ years
#3 #1
Search
machine learning for search engines
RankBrain: a deep neural network for search ranking
19. 19
beyond core products, into areas like health and robotics
“Deep Learning for
Robots: Learning from
Large-Scale Interaction”,
Google Research Blog,
posted March, 2016
20. Sharing our tools with researchers and developers
around the world
repository
for “machine
learning”
category on GitHub
#1
Released in Nov.
2015
22. 2222
Ready to use Machine
Learning models
Use your own data to
train models
Cloud
Vision API
Cloud
Speech API
Cloud
Translate API
Cloud Machine Learning
Develop - Model - Test
Google
BigQuery
Stay
Tuned….
Cloud
Storage
Cloud
Datalab
NEW
Alpha
GA BetaGA
Alpha
Beta
GA
23. 2323
beyond core products
for applications like
health and robotics
smart platforms
to accelerate
progress worldwide
Google products
to create great
experiences for users
machine intelligence is already making a huge difference
and there are many, many more opportunities
23