Martin Andrews @ redcatlabs.com
30 July 2016
Android feature since Jellybean (v4.3, 2012) using Cloud
Trained in ~5 days on 800 machine cluster
Embedded in phone since Android Lollipop (v5.0, 2014)
Google's Deep Models are on the phone
"Use your camera to translate text instantly in 26 languages"
Translations for typed text in 90 languages
Google Street-View (and ReCaptchas)
(now human-competitive on ImageNet)
Some good, some not-so-good
Google's DeepMind purchase
Learn to play Atari games from the pixels alone
Better than humans 2 hours after switching on
Google DeepMind's AlphaGo
Learn to play Go from (mostly) self-play
Change weights to change output function
Layers of neurons combine to
form more complex functions
x
→ target_y
)output_y
from the x
output_y
is closer to target_y
for that x
Follow the gradient of the error
vs the connection weights
http://ConvNetJS.com/
(or search online for TensorFlow Playground)
deep-learning ... .OVA
'http://localhost:8080/
ssh
- it should Just Work
ssh -p 8282 user@localhost
# password=password
numpy
and BLAS
C/C++
or CUDA
(or OpenCL
)Use the 'play' button to walk through the workbook
We using lasagne
as an additional layer on top of Theano
... around 8-9% error rate on the test set
* Please add a star... *
My blog : http://mdda.net/
GitHub : mdda
Three filter layers - nice visual interpretation
After convolution layers, use dense/softmax as before
GoogLeNet (2014)
Play with a ~2014 pre-trained network
... need for GPU programmers
Google Inception-v3 (2015)
Play with a ~2015 pre-trained network
Re-purpose a pretrained network
Careful! : Some images cannot be un-seen...
Style-Transfer an artist onto your photos
My blog : http://mdda.net/
GitHub : mdda
RNN chain
Simplest RNN ~ gradient problem
A GRU
A Long Short-Term Memory (LSTM) Unit
Highlighting Analogies
Learning to do ~NER
Bidirectional RNN
Each output becomes the next input
... work-in-progress ...
Still a work-in-progress (training takes too long)
JDa&g#sdWI&MKW^gE)I}<UNK>f;6g)^5*|dXdBw6m\2&XcXVy\ph8G<gAM&>e4+mv5}OX8G*Yw9&n3XW{h@&T\Fk%BPMMI
OV&*C_] ._f$v4I~$@Z^&[2
mOVe`4W)"L-KClkO]wu]\$LCNadyo$h;>$jV7X$XK#4,T(y"sa6W0LWf\'_{\#XD]p%ck[;O`!Px\#E>/Or(.YZ|a]2}q|@a9.g3nV,U^qM $+:nlk0sd;V-Z&;7Y@Z "l-7P^C
"xBF~~{n} n\ Pcbc9f?=y)FIc1h5kvjIi
C<UNK>s DWJr_$ZQtu"BTYm'|SMj-]Z<Vqj*.lh%IYW|q.GK:eNI"r>833?+RuUsOj_)a{\T}gH.zZR^(daC3mg5P0iFi]bqGo4?T|\>0_H&g889voTh=~)^DDRYND46z1J]x;<U>>%eNIRckL)N8n<UNK>n3i)+Ln8
?)9.#s7X]}$*sxZ"3tf ")
@'HW.;I5)C.*%}<jcNLN+Z__RWoryOb#
/`r
Som the riscele his nreing the timest stordor hep pIs dach suedests her, so for farmauteds?
By arnouy ig wore
Thou hoasul dove he five grom ays he bare as bleen,
The seend,
And, an neeer,
Whis with the rauk with, for be collenss ore his son froven faredure:
Then andy bround'd the CowE nom shmlls everom thoy men ellone per in the lave ofpen the way ghiind, thour eyes in is ple gull heart sind, I I wild,
Frreasuce anspeve, wrom fant beiver, not the afan
And in thou' histwish a it wheme-tis lating ble the liveculd;
Noorroint he fhallought, othelts.
AWhis grook my glass' to his sweet,
Bub my fears liken?
And of live every in seedher;
A Lood stall,
But tare tought than thencer sud earth,
Use'st bee sechion,
For all exprit' are a daud in heaven doth her infook perust the fork the tent.
For maud,
The pittent gover
This and rimp,
Who new
Thoir oldes and did hards, cound.
DEDENIUS Why shoulmeying to to wife,
And thou say: and wall you teading for
that struke you down as sweet one.
With be more bornow, bly unjout on the account:
I duked you did four conlian unfortuned drausing-
to sicgia stranss, or not sleepplins his arms
Gentlemen? as write lord; gave sold.
AENEMUUNS Met that will knop unhian, where ever have
of the keep his jangst?icks he I love hide,
Jach heard which offen, sir!'
[Exit PATIIUS, MARGARUS arr [Enter CLOTHUR]
RNN Discriminator & more
Using an Autoencoder
My blog : http://mdda.net/
GitHub : mdda
Q is a measure of what we think about the future
Classic game : No superfluous features
Deep Reinforcement Learning for Bubble Breaker
Deep Reinforcement Learning for Bubble Breaker
Encoder Network feeds hidden state to Decoder Network
Differentiable → Training : Even crazy stuff works!
* Please add a star... *
http://hsgk.in/AdvancedDeepLearning
My blog : http://mdda.net/
GitHub : mdda