Brain Simulator Úvod Jaroslav V ítků Cíle cvičení Použití Brain Simulatoru pro modelování v KSI CV1 základní principy použití stávajících uzlů CV2 pokročilejší použití ID: 808208
Download The PPT/PDF document "Kognitivn í systémy A6MX33KSY" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Kognitivní systémyA6MX33KSY
Brain
Simulator - Úvod
Jaroslav V
ítků
Slide2Cíle cvičeníPoužití Brain Simulatoru pro modelování v KSI
CV1: základní principy, použití stávajících uzlů
CV2:
pokročilejší
použití,
skriptov
ání v C
#
(python)
Typy učení (supervised
/unsupervised/reinforcement)
Čas na d
otazy
ohledně semestrálních prací
Jednoduché zadání
Slide3Nová verze BrainSimulatoru
Brain Simulator
ke stažení (
dokumentace)
http://
www.goodai.com/brain-simulator
F
ó
rum pro
dotazy
http://forum.goodai.com/index.php
Slide4C# Node ScriptingHow to write own code in the
BrainSimulator
?
Possibility 1
Download BrainSimulator sources
O
pen project in the Visual Studio
Implement own
N
ode
/World
Possibility 2
Use a special nodes for scripting
C# and Python nodes
Demo now
Slide5Types of LearningSupervisedUnsupervised
Reinforcement
Slide6Data
examples
Data
Slide7Supervised learning
Supervised learning: given labeled examples
label
label
1
label
3
label
4
label
5
labeled examples
examples
Slide8Supervised learning
model/
predictor
label
label
1
label
3
label
4
label
5
Supervised learning: given labeled examples
Slide9Supervised learning
model/
predictor
Supervised learning: learn to predict new example
predicted label
Slide10Supervised learning: classification
Supervised learning: given labeled examples
label
apple
apple
banana
banana
Classification:
a finite set of labels
Slide11Unsupervised learning
Unsupervised learning: given data, i.e. examples, but no labels
Slide12Reinforcement learningleft, right, straight, left, left, left, straight
left, straight, straight, left, right, straight, straight
GOOD
BAD
left, right, straight, left, left, left, straight
left, straight, straight, left, right, straight, straight
18.5
-3
Given a
sequence
of examples/states and a
reward
after completing that sequence, learn to predict the action to take in for an individual example/state
Slide13Reinforcement learning example
…
WIN!
…
LOSE!
Backgammon
Given sequences of moves and whether or not the player won at the end, learn to make good moves
Slide14Brain Simulator – Learning
Supervised
FF
network
, MNIST
classifier
LSTM network, prediction of sequences
Unsupervised
MNIST
+ RBM
(unsupervised)
Reinforcement
Q-Learning
Breakout game
Discrete Q-Learning
Slide15Brain Simulator – Useful stuff
Visual Attention
Focuser example
Slide16Reaching AGI with Brain Simulator
Brain Simulator
ke stažení + dokumentace
http://
www.goodai.com/brain-simulator
F
ó
rum pro
dotazy
http://
forum.goodai.com/index.php
Ukázkové projekty ke stažení
https://
github.com/GoodAI/BrainSimulatorSampleProjects
Typy
učení, slidy zdroj: http://goo.gl/FzmBGH
D
ěkuji za pozornost
jaroslav.vitku
@
keenswh.com
Slide17A6MX33KSY – semestrální práceTémata:
Predikce
pohybu
jednoho
objektu
Simulace
pozornosti
při
průchodu bludištěm Hra
Logik Hra PiškvorkyHra
Blockout
Slide18Brain Simulator – modulesImplemented modules
FFNs, RNNs, CNNs, LSTM, RBMs
Self-organizing networks (SOM, GNG)
Vector symbolic architectures (HRR, BSC)
Hierarchical temporal memory (spatial & temporal poolers)
Spiking networks & STDP
Computer vision (filters, segmentation, optical flow …)
SLAM, PID …
Q-Learning
Imported modules
Caffe
, BLAS, BEPU Physics, Space Engineers