Конвергенция таких современных IT-трендов как "интернет вещей" (internet of things), "глобальная смысловая сеть" (semantic web, "things instead of strings") и робототехника приводит к созданию нового поколения программных агентов. Эти агенты будут способны автономно функционировать в глобальной сети, соединяющей как людей, так и устройства самого разного рода. Им нужно будет обладать способностями адаптивного интеллекта, позволяющими производить обучение агентов применительно к конкретным задачам пользователей в различных прикладных областях, оперируя со смысловыми сетями ("графами знаний") – как загружаемыми пользователями-людьми и другими агентами-корреспондентами, так и достраиваемыми в ходе взаимодействия с окружающем миром. В рамках проекта Aigents, создается среда интеллектуальных агентов для поиска информации в интернете. Каждый агент включает семантическую базу данных, развитую систему управления онтологиями, возможности интеллектуальной адаптации, а также - языковый интерфейс, позволяющий агентам взаимодействовать как друг с другом, так и с пользователями.
Антон Колонин "О создании программных агентов для "интернета вещей"
1. Software Agents for
Internet of Things
Find and Share
Your self with Peer
AIGENTS.COM
AINL-2014, Anton Kolonin, Aigents Group 1
2. Smart Things and Human Beings
Human
User
Artificial
Agent
AINL-2014, Anton Kolonin, Aigents Group 2
3. Convergent technological breakthroughs
Internet of Things
Smart
Things
Adaptive Robotics
Things instead of Strings
(Artificial General Intelligence)
(Semantic Web)
AINL-2014, Anton Kolonin, Aigents Group 3
4. Internet of Things – Agents everywhere
Communication cloud
HTTP, email, IRC, SMS, TCP/UDP
(using common language)
AINL-2014, Anton Kolonin, Aigents Group 4
WWW Cloud
Knowledge
base
User
Agent
Collector
Agent
Broker
Agent
Storage
Agent
Actor
Agent
File
systems Input
sensor
Control
device
Search
Agent
Human
User
5. Smart Things and Human Beings:
Possess a proprietary belief as an
internal ontology or a “world view” – as
structured knowledge about environment
including peer things and beings
Capable of adaptive behavior –
experiential self-learning in course of
interaction with environment and other
peers
Communication with peers by means of
“open” (non-protocol) extensible linguistic
interfaces
AINL-2014, Anton Kolonin, Aigents Group 5
6. Proprietary belief as internal ontology
Representing knowledge in graphs – a
semantic database
Support for hyper-graphs – multiple
hierarchical levels of complexity, abstraction
and cross-dependency within single
database
Special representation of time and
analysis of temporal dynamics
Collaborative knowledge acquisition –
determination of truth value with account of
social links and trust levels
AINL-2014, Anton Kolonin, Aigents Group 6
7. Semantic Web - “Things” instead of “Strings”
Google Knowledge Graph/Vault, Google+, Youtube
Google Freebase
Facebook Graph
Others
AINL-2014, Anton Kolonin, Aigents Group 7
Others
8. Recurrent Beliefs – Things think about Things
Cd
Ad Bd
User D
Dc
Ac Bc
User C
Agent A
Da Ba
Agent B
Db Ab
Cb
AINL-2014, Anton Kolonin, Aigents Group 8
Ca
9. Adaptive behavior
Artificial General Intelligence (AGI) – as an
ability to derive a complex proprietary belief
and behavioral schemata corresponding to a
given environment, provided with minimal
“bootstrapping” (“seed”) belief and
schemata.
“Toddler Turing test” - building the world
view and behavioral (conversational for the
given instance) schemata based on self-learning
with experiential feedback from the
environment.
AINL-2014, Anton Kolonin, Aigents Group 9
10. Adaptive Robots around us – allies or opponents
User X
AINL-2014, Anton Kolonin, Aigents Group 10
11. Extensible (non-protocol) linguistic interface
Asynchronous and symmetric
communication protocols
“Open” structure of a language based
on common “foundation ontology”
Partial and probabilistic
comprehension of information out of
scope of “foundation ontology”
Humane communication language
AINL-2014, Anton Kolonin, Aigents Group 11
12. Mixed environment – multiple languages?
Communication
HTTP, E-mail, IRC, SMS, TCP/UDP, voice, ...
Language
Lisp, AIML/XML, RDF/OWL/XML, Turtle,
JSON, Wolfram, Lojban, Human...
AINL-2014, Anton Kolonin, Aigents Group 12
Agent A
Agent B
User C
User D
13. Linguistic gap – best way to fill?
Human
User
Structured
Ontology-based
“Pidgin”?
AINL-2014, Anton Kolonin, Aigents Group 13
Artificial
Agent
Lisp, AIML/XML,
RDF/OWL/XML,
Turtle, JSON
Lojban,
Esperanto,
English
14. Experimental “Aigents Language” - example
My is appliance, agent, thermostat, device.
Device has shape, color, voltage.
Appliance has location.
My shape rectangular, color white, voltage 220, location kitchen.
My feels temperature, humidity, CO2, being.
Temperature, humidity, CO2 is number.
Feeling is good or bad.
My feeling good, temperature 20, humidity 72.
Моя это прибор, агент, термостат, устройство.
Моя иметь форма, цвет, питание.
Моя иметь место.
Моя форма прямоугольный, цвет белый, питание 220, место
кухня.
Температура, влажность, CO2 это число.
Чувстовать это хорошо, плохо.
Моя чувствовать хорошо, температура 20, влажность 72.
AINL-2014, Anton Kolonin, Aigents Group 14
15. “Aigents Language” - comparisons
English
What is your feeling?
If your feeling is good?
Your feeling is good.
Have your feeling good!
Aigents Language
Your feeling?
Your feeling good?
Your feeling good.
Your feeling good!
Aigents Language - written
I (can (eat, sleep), want (dance, sing)).
I {can (eat, sleep), want (dance, sing)}.
I (can {eat, sleep}, want {dance, sing}).
You [eat {rice, meat}, drink {juice, water}]!
Aigents Language
A C (D,E).
A (C D, F G).
A (C (D,E), F (G,H)).
(A,B) C D.
(A,B) (C (D,E), F (G,H)).
Russian (with tonal modulation)
Твое ощущение? (rising tone)
Твое ощущение хорошее? (rising tone)
Твое ощущение хорошее. (neutral tone)
Твое ощущение хорошее! (lowering tone)
Aigents Language - spoken
I can eat and sleep and want dance and sing.
I can eat and sleep or want dance and sing.
I can eat or sleep and want dance or sing.
You eat rice and meat next drink juice and water!
Term logic
A C D. A C E.
A C D. A F G.
A C D. A C E. A F G. A F H.
A C D. B C D.
A C D. A C E. B C D. B C E. A F G. A F H. B F G. B F H.
Turtle
A C D,E.
A C D; F G.
A C D,E; F G,H.
AINL-2014, Anton Kolonin, Aigents Group 15
16. Aigents platform : Value proposition
For clients:
Personal agents of intelligent web search for users;
Federated active search in communities of users.
For businesses and organizations:
Collaborative knowledge discovery and decision making
for corporate users and national/local communities.
For clients:
Personal agents of intelligent web search for users;
Federated active search in communities of users.
For businesses and organizations:
Collaborative knowledge discovery and decision making
for corporate users and national/local communities.
Web & community
User
Get intelligent
and specific
news timely Aigent
Request specific things
of permanent interest
anywhere on the web
and in the community
User
Copyright 2014 Anton Kolonin, Aigents Group 16
User
User
User
17. Aigents platform : Examples
Weather today 32 °F, rainy
Solar flare class M at 5 AM
Weather today $temperature °F, $details
Solar flare class {M, C, X} at $time
Google acquires Jetpac
Apple recently acquired Swell
$acquirer acquires $acquiree
$acquirer acquired $acquiree
User
Lease - Ridge View 15, $2400, 1/12/2013”
Rent, address Way 24, $2100, 15/12/2013”
{Lease, Rent} $address, $price, $date
Copyright 2014 Anton Kolonin, Aigents Group 17
Aigent
Generic
Aigent
Business
Domain
Aigent
Real
Estate
18. Aigents platform : Competitive advantages
Automatic, time-critical updates - personal web crawler, alerts to email/phone
Personalized search experience - customized list of sites and search topics
Social collaboration - sharing news with relatives, friends and colleagues
Semantics extracted from texts – from simple templates to complex patterns
Intelligent trainable software agent - experiential learning by user feedback
Automatic, time-critical updates - personal web crawler, alerts to email/phone
Personalized search experience - customized list of sites and search topics
Social collaboration - sharing news with relatives, friends and colleagues
Semantics extracted from texts – from simple templates to complex patterns
Intelligent trainable software agent - experiential learning by user feedback
Web & community
User
Get intelligent
and specific
news timely Aigent
Request specific things
of permanent interest
anywhere on the web
and in the community
User
Copyright 2014 Anton Kolonin, Aigents Group 18
User
User
User
19. Copyright 2014 Anton Kolonin, Aigents Group 19
Android
Windows/Mac/Linux
Web
Aigents platform : Screen shots
20. Aigents platform : What is inside?
Sites
News
Things
Users
Copyright 2014 Anton Kolonin, Aigents Group 20
21. Aigents platform : Internals and Modalities
Storager
Knowledge
User
Feedback
Sessioner
Siter
Aigents
Language
Writer
AINL-2014, Anton Kolonin, Aigents Group 21
Communicator
Java UI
(Swing)
Emailer
SMTP/POP
Cmdliner
text
SMPPer
SMPP
TCPer
TCP/IP
IRCer
IRC
HTTPer
HTTP(S)
Android UI
Sessions
Sites
Aigents
Language
Reader
User
Login
User
Registrati
on
User
Verification
User
Conversation
User
Logout
User
Verification
Change
User
Email
Change
User
Clarification
User
Confirmation
User
Declaration
User
Interrogation
User
Direction
Conversationer
Selfer Self
Checking
Times
Self
Messaging
News
Self
Reading
Sites
Self
Matching
Topics
Self
Learning
Patterns
Self
Watching
News
Self
Forgetting
Stuff
22. Aigents : Benefactors and Beneficiaries
Consume
Content News
Contribute
Ad News
Copyright 2014 Anton Kolonin, Aigents Group 22
Contribute
Content News
Consume
Ad News
Automatic
Intelligent
Internet
Aigent A Aigents
Aigent B
Aigent C
Aigent D
Aigent X
User A
Benefactors Beneficiaries
User B
User C
User D
23. Aigents Platform : Summary
The Aigents software platform
enables each Internet user to start getting information timely and intelligently,
precisely matching personal preferences and social connections
using a distributed semantic search technology within an ecosystem of
personal devices and peer-to-peer servers hosted by communities and
companies (available as free or premium software packages and service
subscriptions).
The Aigents software platform
enables each Internet user to start getting information timely and intelligently,
precisely matching personal preferences and social connections
using a distributed semantic search technology within an ecosystem of
personal devices and peer-to-peer servers hosted by communities and
companies (available as free or premium software packages and service
subscriptions).
Web & community
User
Get intelligent
and specific
news timely Aigent
Request specific things
of permanent interest
anywhere on the web
and in the community
User
Copyright 2014 Anton Kolonin, Aigents Group 23
User
User
User
24. Aigents : The beginning
Thank you for your attention!
Find and Share
Your self with Peer
AIGENTS.COM
Copyright 2014 Anton Kolonin, Aigents Group 24