Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

The Reality of Artificial Intelligence

Presentation given to METal International in Los Angeles on July 13, 2013. See: http://www.metalinternational.com/met/2013/07/olin-hyde What is the current state of artificial intelligence and how will it affect our lives? Current use of AI places huma
by

Tom Marsh

on 16 July 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of The Reality of Artificial Intelligence

The Reality of Artificial Intelligence
Changes Everything
Law of Accelerating Returns
Graphs Enable Us To Make Nathan Useful
Nathan Learns to Represent Reality
Bigger Graphs = More Accurate Representations
Learns What Matters Most
Autonomic Learning
Finds Similar Ideas In Different Data Sets
Biologically
Inspired
Intelligence
7 Ways Nathan
Changes Everything
Data Agnostic
Empowers Existing
Technologies
Transparent
Embedded on Devices
Small
Dynamic &
Adaptive
"Methuselah"
World's Oldest Tree
(4,768 years)
Shares Knowledge Among
Devices
Learns Meaning of Anything, Any Language
Add Pattern Recognition to Anything
Users Control Privacy
Learns Like A Child
Without Supervision
Autonomic
The BIG Idea
How Nathan Works

Graphs Show
Connections & Composition
Nathan
Innovation
Efficiency
Empowering
Sustaining
Evolutionary
Disruptive
Revolutionary
Democratize AI
1
2
3
4
6
7
make.personal
5
What Customers Expect
Makes Markets Grow
Surprise!
Transforms Markets
Very, Very Rare
Do things better
Change The World
'We didn't know we needed what we can't live without!'
Detects the patterns within any binary data set.
+
+
NLP

>
Fits Anywhere
16GB
x 80,000
< 200 KB
Unique Like A Fingerprint
Graphs Can Include Any Type of Data
Text
Sound
Signal
Image
>
Bigger Data Is Better Data
Stores Each Unique Byte Pattern Only Once
Sees Connections of Every Byte-Pattern

Very Small, Accurate Model of Big Data
~ 300 Terabytes
~20 Trillion Words
~ 1 Million Words
~16 Megabytes
English Language
US Library of Congress
~ 1 Million Words
+ Connections
~ 160 Terabytes
Fits On A Server
G[V,E]
Vertices
Edges
Detects
keywords
A Better Way:
Typical Improvements
Creates New Industries
Much Rarer
Keywords = Most Meaningful Data Elements
Detects Associations
Graph
Never Stops Learning.
Awards
Imagine if you could...
Embed Intelligence in EVERY device and application?
Graphs = Context
Finding Relevance = Matching ALL Graphs
All New TV Shows
Shows I Watch
Shows I'll Want to See
Associations
Keywords
Common Use Case
Knowledge Retrieval = Matching Graphs
Detects IDEAS: Works Where Search Fails
All Scientific Articles
(22 million)
My New Theory
Articles Related to My New Idea
Pragmatic
(Time & Location)
Syntax
Semiotics: How Nathan Perceives Context
Semantic
Intersection = specific contextual meaning of any data
Data ≠ Solution
Marketing = Personalization
Privacy = End-User Controls
Detects features, patterns and anomalies without any programming.
Output
API
Complex Data
Intelligence
Learning
Input
Nathan understands complexity
Unstructured Data I/O
(Semantics)
Structured Sensor Data
(Time & Location)
Unstructured Data I/O
(Semantics)
Unstructured Data I/O (Syntax)
Unstructured Data I/O
(Semantics)
Structured Sensor Data
(Time & Location)
Unstructured Data I/O (Syntax)
Perfect Restaurant
(intersection of 3 graphs)
Use Case: Personalization for Mobile Devices
All Options for a Date
(movie, restaurants, etc.)
Structured data
Geo location
Time (open)
Table avail
Unstructured data
Customer Reviews
Menus
News articles
Social media chats
My Preferences
Structured data
Geo location
Time desired
Who’s joining me (wife)
Unstructured data
Customer Reviews
Menus
News articles
Social media chats
Past behavior
Structured data
Unstructured data
My Wife’s Preferences
Structured data
Geo location
Time to travel
Where she’s coming from (indication of dress)
Unstructured data
Customer Reviews
Menus
News articles
Social media chats
Past behavior
Structured data
Unstructured data
Fingerprint of Desire for Date Night
Learns without programming.

Understands generalizable ideas.

Decides based on what it has learned.

Adapts to environment.

Reasons by synthesizing information.
Nathan: Intelligent Agent
Nathan outperforms rules-based & statistical AI systems.
Competitive Comparison
ai-one detects the meaning of a document by identifying the
most important words in the text.

Automatically generates a graph representation of knowledge.

Works without human intervention or models.
ai-one has the neutral observer position to detect the inherent meaning
Receiver
Text interpretation is dependant on complex contexts
Every human interprets semantics differently because of experience, intellectual level, cognitive biases, time & location
Subjective meaning
and semantic
Subjective interpretation & understanding
Sender
Cloud Services
Mr. Smith
Data Agnostic
Learns Meaning of Anything, Any Language
Detects the patterns within any binary data set.
?
Feature Detection
Feature Relevance
Contextual
Understanding
Sees similarity
by associations
Better, faster and cheaper than competitor technology from IBM, Google and Apple.
eDiscovery
finding ideas in legal documents
Was James Cameron right?
"The Singularity"
Ray Kurzweil
Harsh Reality
Where are you on the curve?
Innovators
Early Adopters
Early
The Majority
Late
Laggards
(Gov't, DoD, GM, etc.)
1st to
Market
Fast
Followers
So...


and
Demo of BrainBrowser
Accelerates Medical Research 16x
Patriot?
Traitor?
Common Denominator
CENTRALIZATION
Commercial
APIs
and/or
Open
Source
Nathan API
Public Cloud
Private Cloud
Embed on Devices
< 200 KB API
JavaScript, Node.js, JSON
Three Flavors
Neuronal Net for Content Analytics
Melvin Kranzberg's 6 laws of technology:
5. ...the history of technology is the most relevant.
We Are at Transition Point
in History
Why are we still a startup?
How does Nathan change history?
Available since 2009
NathanAPP
Any JavaScript Developer Can Learn to Use in < 4 hours
Please
Interrupt Me!
@olinhyde
oh@ai-one.com
1. Technology is neither good nor bad; nor is it neutral.
2. Invention is the mother of necessity.
3. Technology comes in packages, big and small.
4. ...nontechnical factors take precedence in technology-policy decisions.
6. ...the history of technology is a very human activity
Intelligence: How accurately a model fits reality.
Germ Theory
Miasma Theory
Cholera
Precise Autonomic Model of Reality
Cyber Security
detecting unusual activity
Classification
grouping documents by theme
Text Mining
finding information where keywords fail
Personalization Engines
context specific recommendations
Our Current Customers Use Cases
We are here
Full transcript