Your SlideShare is downloading. ×
0
Research
Neocortical theory
Algorithms
NuPIC
Open source
community
Products
Automated streaming analytics
Catalyst for mac...
“If you invent a
breakthrough so
computers can learn,
that is worth 10
Microsofts”
"If you invent a
breakthrough so
computers that learn
that is worth 10 Micros
1) What principles will we use to build
inte...
1) Flexible (universal learning machine)
2) Robust
3) If we knew how the neocortex worked,
we would be in a race to build ...
What the Cortex Does
data stream
retina
cochlea
somatic
The neocortex learns a model
from sensory data
- predictions
- ano...
“Hierarchical Temporal Memory” (HTM)
1) Hierarchy of nearly identical regions
- across species
- across modalitiesretina
c...
2.5 mm
Cortex
Layers
Layers & Columns
2/3
4
5
6
2/3
4
5
6
Sequence memory: high-order inference
Sequence memory: sensory-m...
2/3
4
5
6
L4 and L2/3: Feedforward Inference
Copy of motor commands
Sensor/afferent data
Learns sensory-motor transitions
...
2/3
4
5
6
L5 and L6: Feedback, Behavior and Attention
Learns sensory-motor transitions
Learns high-order transitions
Well ...
How does a layer of neurons learn sequences?
2/3
4
5
6
Learns high-order transitions
First:
- Sparse Distributed Represent...
Sparse Distributed Representations (SDRs)
• Many bits (thousands)
• Few 1’s mostly 0’s
• Example: 2,000 bits, 2% active
• ...
SDR Properties
1) Similarity:
shared bits = semantic similarity
subsampling is OK
3) Union membership:
Indices
1
2
|
10
Is...
Model neuronFeedforward
100’s of synapses
“Classic” receptive field
Context
1,000’s of synapses
Depolarize neuron
“Predict...
Learning Transitions
Feedforward activation
Learning Transitions
Inhibition
Learning Transitions
Sparse cell activation
Time = 1
Learning Transitions
Time = 2
Learning Transitions
Learning Transitions
Form connections to previously active cells.
Predict future activity.
This is a first order sequence memory.
It cannot learn A-B-C-D vs. X-B-C-Y.
Learning Transitions
Multiple predictions can ...
High-order Sequences Require Mini-columns
A-B-C-D vs. X-B-C-Y
A
X B
B
C
C
Y
D
Before training
A
X B’’
B’
C’’
C’
Y’’
D’
Aft...
Cortical Learning Algorithm (CLA)
aka Cellular Layer
Converts input to sparse representations in columns
Learns transition...
Anomaly Detection Using CLA
CLA
Encoder
SDR
Prediction error
Time average
Historical comparison
Metric + Time
Anomaly scor...
CloudWatch
AWS
Grok for AWS Users
- Automated model creation
via web, CLI, API
- Breakthrough anomaly detection
- Dramatic...
Grok for AWS Mobile UI
 Sorted by anomaly score
 Continuously updated
 Continuously learning
What Can the CLA/Grok Detect?
Sudden changes Slow changes Subtle changes in regular data
What Can the CLA/Grok Detect?
Patterns that humans can’t seeChanges in noisy data
CEPT.at - Natural Language Processing Using SDRs and CLA
Document corpus
(e.g. Wikipedia)
128 x 128
100K “Word SDRs”
- =
A...
Sequences of Word SDRs
Training set
frog eats flies
cow eats grain
elephant eats leaves
goat eats grass
wolf eats rabbit
c...
Sequences of Word SDRs
Training set
eats“fox”
?
frog eats flies
cow eats grain
elephant eats leaves
goat eats grass
wolf e...
Sequences of Word SDRs
Training set
eats“fox”
rodent
1) Word SDRs created without supervision
2) Semantic generalization
S...
Cept and Grok use exact same code base
eats“fox”
rodent
Source code for:
- Cortical Learning Algorithm
- Encoders
- Support libraries
Single source tree (used by GROK), GPLv3
Act...
Research
- Implement and test L4 sensory/motor inference
- Introduce hierarchy (?)
- Publish
NuPIC
- Grow open source comm...
1) The neocortex is as close to a universal
learning machine as we can imagine
2) Machine intelligence will be built on th...
Future of Machine Intelligence
Future of Machine Intelligence
Definite
- Faster, Bigger
- Super senses
- Fluid robotics
- Distributed hierarchy
Maybe
- H...
Why Create Intelligent Machines?
Thank You
Live better Learn more
Brains, Data, and Machine Intelligence (2014 04 14 London Meetup)
Upcoming SlideShare
Loading in...5
×

Brains, Data, and Machine Intelligence (2014 04 14 London Meetup)

1,599

Published on

Jeff will discuss the Brains, Data, Machine Intelligence, Cortical Learning Algorithm he developed and the Numenta Platform for Intelligent Computing (NuPIC).

Published in: Science, Technology, Education
0 Comments
6 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total Views
1,599
On Slideshare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
95
Comments
0
Likes
6
Embeds 0
No embeds

No notes for slide
  • I assume some knowledge of SDRs and CLAThis talk is more speculative than most
  • Matt
  • Matt
  • Matt
  • Matt
  • Matt
  • People liked this. Should I have something like this up front?
  • Opposite sex?
  • Transcript of "Brains, Data, and Machine Intelligence (2014 04 14 London Meetup)"

    1. 1. Research Neocortical theory Algorithms NuPIC Open source community Products Automated streaming analytics Catalyst for machine intelligence Brains, Data, and Machine Intelligence Jeff Hawkins jhawkins@numenta.com
    2. 2. “If you invent a breakthrough so computers can learn, that is worth 10 Microsofts”
    3. 3. "If you invent a breakthrough so computers that learn that is worth 10 Micros 1) What principles will we use to build intelligent machines? 2) What applications will drive adoption in the near and long term? Machine Intelligence
    4. 4. 1) Flexible (universal learning machine) 2) Robust 3) If we knew how the neocortex worked, we would be in a race to build them. Machine intelligence will be built on the principles of the neocortex
    5. 5. What the Cortex Does data stream retina cochlea somatic The neocortex learns a model from sensory data - predictions - anomalies - actions The neocortex learns a sensory-motor model of the world
    6. 6. “Hierarchical Temporal Memory” (HTM) 1) Hierarchy of nearly identical regions - across species - across modalitiesretina cochlea somatic 2) Regions are mostly sequence memory - for inference - for motor 3) Feedforward: Temporal stability Feedback: Unfold sequences data stream
    7. 7. 2.5 mm Cortex Layers Layers & Columns 2/3 4 5 6 2/3 4 5 6 Sequence memory: high-order inference Sequence memory: sensory-motor inference Sequence memory: motor generation Sequence memory: attention Cortical Learning Algorithm (CLA) CLA is a cortical model, not another “neural network”
    8. 8. 2/3 4 5 6 L4 and L2/3: Feedforward Inference Copy of motor commands Sensor/afferent data Learns sensory-motor transitions Learns high-order transitions Stable Predicted Pass through changes Un-predicted Layer 4 learns sensory-motor transitions. Layer 3 learns high-order transitions. These are universal inference steps. They apply to all sensory modalities. Next higher region A-B-C-D X-B-C-Y A-B-C- ? D X-B-C- ? Y
    9. 9. 2/3 4 5 6 L5 and L6: Feedback, Behavior and Attention Learns sensory-motor transitions Learns high-order transitions Well understood tested / commercial Recalls motor sequences Attention 90% understood testing in progress 50% understood 10% understood FeedforwardFeedback
    10. 10. How does a layer of neurons learn sequences? 2/3 4 5 6 Learns high-order transitions First: - Sparse Distributed Representations - Neurons Cortical Learning Algorithm (CLA)
    11. 11. Sparse Distributed Representations (SDRs) • Many bits (thousands) • Few 1’s mostly 0’s • Example: 2,000 bits, 2% active • Each bit has semantic meaning Learned 01000000000000000001000000000000000000000000000000000010000…………01000 Dense Representations • Few bits (8 to 128) • All combinations of 1’s and 0’s • Example: 8 bit ASCII • Bits have no inherent meaning Arbitrarily assigned by programmer 01101101 = m
    12. 12. SDR Properties 1) Similarity: shared bits = semantic similarity subsampling is OK 3) Union membership: Indices 1 2 | 10 Is this SDR a member? 2) Store and Compare: store indices of active bits Indices 1 2 3 4 5 | 40 1) 2) 3) …. 10) 2% 20%Union
    13. 13. Model neuronFeedforward 100’s of synapses “Classic” receptive field Context 1,000’s of synapses Depolarize neuron “Predicted state” Active Dendrites Pattern detectors Each cell can recognize 100’s of unique patterns Neurons
    14. 14. Learning Transitions Feedforward activation
    15. 15. Learning Transitions Inhibition
    16. 16. Learning Transitions Sparse cell activation
    17. 17. Time = 1 Learning Transitions
    18. 18. Time = 2 Learning Transitions
    19. 19. Learning Transitions Form connections to previously active cells. Predict future activity.
    20. 20. This is a first order sequence memory. It cannot learn A-B-C-D vs. X-B-C-Y. Learning Transitions Multiple predictions can occur at once. A-B A-C A-D
    21. 21. High-order Sequences Require Mini-columns A-B-C-D vs. X-B-C-Y A X B B C C Y D Before training A X B’’ B’ C’’ C’ Y’’ D’ After training Same columns, but only one cell active per column. IF 40 active columns, 10 cells per column THEN 1040 ways to represent the same input in different contexts
    22. 22. Cortical Learning Algorithm (CLA) aka Cellular Layer Converts input to sparse representations in columns Learns transitions Makes predictions and detects anomalies Applications 1) High-order sequence inference L2/3 2) Sensory-motor inference L4 3) Motor sequence recall L5 Capabilities - On-line learning - High capacity - Simple local learning rules - Fault tolerant - No sensitive parameters Basic building block of neocortex/machine intelligence
    23. 23. Anomaly Detection Using CLA CLA Encoder SDR Prediction error Time average Historical comparison Metric + Time Anomaly score CLA Encoder SDR Prediction error Time average Historical comparison Metric + Time Anomaly score . . .
    24. 24. CloudWatch AWS Grok for AWS Users - Automated model creation via web, CLI, API - Breakthrough anomaly detection - Dramatically reduces false positives/negatives - Supports auto-scaling and custom metrics Customer Instances & Services - Mobile client - Instant status check - Mobile OS and email alerts - Drill down to determine severity
    25. 25. Grok for AWS Mobile UI  Sorted by anomaly score  Continuously updated  Continuously learning
    26. 26. What Can the CLA/Grok Detect? Sudden changes Slow changes Subtle changes in regular data
    27. 27. What Can the CLA/Grok Detect? Patterns that humans can’t seeChanges in noisy data
    28. 28. CEPT.at - Natural Language Processing Using SDRs and CLA Document corpus (e.g. Wikipedia) 128 x 128 100K “Word SDRs” - = Apple Fruit Computer Macintosh Microsoft Mac Linux Operating system ….
    29. 29. Sequences of Word SDRs Training set frog eats flies cow eats grain elephant eats leaves goat eats grass wolf eats rabbit cat likes ball elephant likes water sheep eats grass cat eats salmon wolf eats mice lion eats cow dog likes sleep elephant likes water cat likes ball coyote eats rodent coyote eats rabbit wolf eats squirrel dog likes sleep cat likes ball ---- ---- ----- Word 3Word 2Word 1
    30. 30. Sequences of Word SDRs Training set eats“fox” ? frog eats flies cow eats grain elephant eats leaves goat eats grass wolf eats rabbit cat likes ball elephant likes water sheep eats grass cat eats salmon wolf eats mice lion eats cow dog likes sleep elephant likes water cat likes ball coyote eats rodent coyote eats rabbit wolf eats squirrel dog likes sleep cat likes ball ---- ---- -----
    31. 31. Sequences of Word SDRs Training set eats“fox” rodent 1) Word SDRs created without supervision 2) Semantic generalization SDR: lexical CLA: grammatic 3) Commercial applications Sentiment analysis Abstraction Improved text to speech Dialog, Reporting, etc. www.Cept.at frog eats flies cow eats grain elephant eats leaves goat eats grass wolf eats rabbit cat likes ball elephant likes water sheep eats grass cat eats salmon wolf eats mice lion eats cow dog likes sleep elephant likes water cat likes ball coyote eats rodent coyote eats rabbit wolf eats squirrel dog likes sleep cat likes ball ---- ---- -----
    32. 32. Cept and Grok use exact same code base eats“fox” rodent
    33. 33. Source code for: - Cortical Learning Algorithm - Encoders - Support libraries Single source tree (used by GROK), GPLv3 Active and growing community Hackathons Education Resources www.Numenta.org NuPIC Open Source Project
    34. 34. Research - Implement and test L4 sensory/motor inference - Introduce hierarchy (?) - Publish NuPIC - Grow open source community - Support partners, e.g. IBM, CEPT Grok - Create commercial value for CLA Attract resources Provide a “target” and market for HW - Explore new application areas Goals For 2014
    35. 35. 1) The neocortex is as close to a universal learning machine as we can imagine 2) Machine intelligence will be built on the principles of the neocortex 3) HTM is an overall theory of neocortex 4) CLA is a building block 5) Near term applications anomaly detection, prediction, NLP 6) Participate www.numenta.org
    36. 36. Future of Machine Intelligence
    37. 37. Future of Machine Intelligence Definite - Faster, Bigger - Super senses - Fluid robotics - Distributed hierarchy Maybe - Humanoid robots - Computer/Brain interfaces for all Not - Uploaded brains - Evil robots
    38. 38. Why Create Intelligent Machines? Thank You Live better Learn more
    1. A particular slide catching your eye?

      Clipping is a handy way to collect important slides you want to go back to later.

    ×