Lecture 6 Agent Communication Languages ACLs

1
Lecture 6 Agent Communication Languages (ACLs)

• Amund Tveit
• Department of Computer and Information Science
• Norwegian University of Science and Technology
• amund.tveit_at_idi.ntnu.no
• http//www.idi.ntnu.no/amundt/
• 47 4 162-6572

2
Lecture Outline

• Motivation for Agent Communication
• Terminology of Agent Communication
• Requirements for ACLs
• Speech-act
• Examples of speech-act
• KQML
• FIPA ACL
• Facilitators
• Fipa ACCs (Agent Communication Channels)
• Fipa Interaction Protocol ( figure 3)
• Syntactic representation (XML, proprietary
  lisp-like and efficient binary encoded)
• Negotiation example in KQML
• Negotiation example in FIPA
• Transport mechanisms (Corba, HTTP)
• FIPA General Architecture Overview
• Ontologies
• KIF language overview and brief examples
• Semantic Web

3
Motivation for Agent Communication

• Language is a very difficult thing to put into
  words Voltaire
• Communication is required for cooperation between
  agents
• Societies can perform tasks no individual agent
  can
• Autonomy encourages disregard for other agents
  internal structure
• Communicating agents need only know a common
  language
• Supports heterogenous agents

4
Terminology of Agent Communication

• Speech Act Formal subset of natural language
  representing actions
• ACL Agent Communication Language
• FIPA - Foundation for Intelligent Physical Agents
• KQML Knowledge Query and Manipulation Language
• Ontology World Model, formal description of
  agent domain
• Encoding syntactic representation of ACL
  messages
• ACC Agent Communication Channel (FIPA)
• KSE Knowledge Sharing Effort
• KIF Knowledge Interchange Format
• Router entity that sends incoming messages in
  correct direction
• Facilitator An agent that includes a Router

5
Features of ACLs - I

• Efficient
• Few bytes but much meaning, rich semantics for
  each message
• Easy-to-use for both machines and humans
• Based on Open Standards
• Allow agent and agent systems by different
  vendors to communicate
• Flexible
• Easy to extend without changing the language,
  using ontologies
• Support several syntactic representations
• Expressive

6
Features of ACLs - II

• Be independent of Communication Carrier and
  protocol
• Support standard Carriers (Wired and Wireless)
• Support standard (Internet) Communication
  Protocols
• High-level
• Be inspired by natural language
• Have clear non-ambigious semantics and syntax
• logic features
• Avoid contradictions
• Procedural or Declarative
• Procedural
• Declarative

7
Pros/Cons of Procedural ACLs

• Pros
• Allows transmission of entire programs
• Directly and efficiently executable
• Cons
• Unidirectional
• Difficult to merge

8
Pros/Cons of Declarative ACLs

• Pros
• Highly expressive
• Bidirectional
• Cons
• Indirectly executable
• Possibly complex semantics

9
Speech acts - I

• How to communicate when having a common language?
• Speech act contains three distinct acts
• Locution
• The physical utterance with context and reference
  to other utterances
• Illocution
• Communication of the speakers intention with the
  speech act
• Perlocution
• Further acts resulting from the speech acts

10
Speech acts II - Illocution

• Illocution communication of the speakers
  intention
• An illocutionary act is composed from
• Propositional content P
• Describes state of affairs (claim)
• Context
• Speaker, hearer, time, location, ..
• Illocutionary Force F
• 5 categories assertives, directives, commisives,
  expressives, declaratives
• Sincerety condititions (psychological state)
• Example speech act The lecturer of SIF8072 is a
  geek

11
Speech acts III Illocutionary Force
12
Speech acts III Illocutionary Force
13
Speech acts III Illocutionary Force
14
Speech acts III Illocutionary Force
15
Speech acts III Illocutionary Force
16
Illuctionary Force

• .. In addition
• Permissives
• Example you may shot the door
• Prohibitives
• Example you may not shot the door

17
KQML - I

• Knowledge Query and Manipulation Language
• Basis work by the Knowledge Sharing Effort
  group
• Basic features of KQML
• Communicates attitude about sentences
• Language primitives are performatives defining
  actions allowed in communication
• Environment may contain facilitators to make
  communication protocol transparent

18
KQML - II

• Idea of KQML represent illocutionary acts
• Illocution communication of the speakers
  intention
• Performatives core of language, they
• Determine types of interactions with
  KQML-speaking agent
• Identify protocol to be used and speech act
  attached to the content
• Specify that content is an assertion, query,
  command or another speech act
• Describe how sender would like any reply
  delivered

19
Primitives (Performatives)
20
Initiators
21
Reactors
22
Completers
23
Categories of KQML performatives

• Basic Query performatives
• Multi-response query performatives
• Response performatives
• Generic Informational performatives
• Generator performatives
• Capability performatives
• Networking performatives

24
KQML Examples

• A?B - (tell (gt 3 2))
• A?B - (perform (print hello t))
• B?A (reply done)
• A?B - (ask-if (gt (size chip1) (size chip2) ))
• B?A (reply true)
• A?B - (subscribe (position ?x ?r ?c))
• B?A (tell (position chip1 8 10))
• B?A (tell (position chip2 8 46))
• B?A (tell (position chip3 3 64))
• A?B - (unsubscribe (position ?x ?r ?c))

25
KQML Statement Structure

• KQML Statements consists of
• A performative
• Parameters and context information
• General syntax
• (KQML-performative
• sender word
• receiver word
• language word
• ontology word
• content expression
• ...)

26
KQML full example

• (ask-one content (geoloc TRD (?long ?lat))
• ontology geo-model3)
• (ask-all
• content price(IBM, ?price, ?time)
• receiver stock-server
• language standard-prolog
• ontology NYSE-TICS)

27
KQML Agent naming

• System for mapping agents into names is important
  in most ACLs
• KQML assumes that names are local
• A can register with B under the name Alice
• A can register with C under the name Albert
• Doesnt preclude the use of a central Agent Name
  Server, an architecture used by most systems
• What gets registered under a name? Contact
  information like
• name(amundbot, tcpip, cavenan.idi.ntnu.no,80).
• name(amundbot,smtp,amundbot_at_jfipa.org)
• name(amundbot,http,www.jfipa.org80/)

28
Facilitators - I

• Agents performing various (useful) communication
  services
• Maintain registry of service names
• Forward messages to named services
• Route messages based on content
• Provide matchmaking
• Provide mediation and translation services

29
Facilitators - I
30
Facilitators - II
31
Facilitators - III
32
Facilitators IV
33
Facilitators - V
34
Routers

• Content-independent message routers
• Each KQML-agent is associated with its own
  separate router process
• Router handles all outgoing/incoming ACL messages
• Outgoing messages can specify a particular agent
  address (to agent or router)
• Message can specify a description of context
• Delivery of messages is not guaranteed

35
Other ACLs

• Telescript supports mobile computing
• FIPA ACL competing/extending KQML
• FIPA vs KQML
• Both are based on speech act
• Different set of performatives
• FIPA has a more formal basis
• FIPA can describe interaction protocols
• FIPA is probably becoming the standard ACL

36
What is FIPA?

• The Foundation for Intelligent Physical Agents
  (FIPA) is a non-profit association.
• FIPAs purpose is to promote the success of
  emerging agent-based applications, services and
  equipment.
• FIPAs goal is pursued by making available in a
  timely manner, internationally agreed
  specifications that maximise interoperability
  across agent-based applications, services and
  equipment.
• URL http//www.fipa.org/

37
Who is FIPA?

• FIPA operates through the open international
  collaboration of member organisations, which are
  companies and universities active in the agent
  field.
• Companies Alcatel, Boeing, British Telecom,
  Deutsche Telekom, France Telecom, Fujitsu,
  Hitatchi, HP, IBM, Fujitsu, Hewlett Packard,
  IBM, Intel, Lucent, NEC, NHK, NTT, Nortel,
  Siemens, SUN, Telia, Toshiba, etc.
• Universities and Research Institutes GMD, EPFL,
  Imperial, IRST, etc.
• Government Agencies DARPA

38
FIPA Encodings and package

• Fipa supports three main types of syntax
  encoding
• Extensible Markup Language XML
• Proprietary Binary format
• Lisp-like format
• Separates between Envelopes and Messages

39
Example - XML-encoded FIPA ACL

• lt?xml version"1.0"?gt
• ltenvelopegt
•   ltparams index"1"gt
•     lttogt
•       ltagent-identifiergt
•         ltnamegtreceiver_at_foo.comlt/namegt
•         ltaddressesgt
•           lturlgthttp//foo.com/acclt/urlgt
•         lt/addressesgt
•       lt/agent-identifiergt
•     lt/togt
•     ltfromgt
•       ltagent-identifiergt
•         ltnamegtsender_at_bar.comlt/namegt
•         ltaddressesgt
•           lturlgthttp//bar.com/acclt/urlgt
•         lt/addressesgt
•       lt/agent-identifiergt
•     lt/fromgt

40
Binary encoded FIPA XML

• 0xfe 0x00 0x97 0x12 0x20 0x31 0x11 0x06 0x19 0x15
  0x37 0x62 0x59 0x20 0x02 0x03 0x02
• r  e  c  e  i  v  e  r  _at_  f 
  o  o  .  c  o  m  0x00
• 0x02 h  t  t  p    /  /  f  o 
  o  .  c  o  m  /  a 
• c  c  0x00 0x01 0x01 0x02 s  e  n  d 
  e  r  _at_  b  a  r  . 
• c  o  m  0x00 0x02 h  t  t  p   
  /  /  b  a  r  .  c 
• o  m  /  a  c  c  0x00 0x01 0x01 0x08
  n  o    e  n  c  r 
• y  p  t  i  o  n  0x00 0x0a h  t 
  t  p    /  /  b  a 
• r  .  c  o  m  /  a  c  c  0x00
  0x20 0x31 0x11 0x06 0x19 0x15 0x37
• 0x62 0x59 0x20 0x03 1  2  3  4  5  6 
  7  8  9  0x00 0x01
•  

41
Shortcomings of current ACLs

• Intentional level description which mental
  attitudes, what definitions?
• Problems with mental attitudes from theory to
  practice
• Can all desirable communication primitives be
  modeled after speech acts? Should they?
• Flexible description of agents capabilities and
  advertising of such capabilities.
• How can we test an agents compliance with the
  ACL?
• Ease of extending an ACL

42
Conclusions of Lecture

• Agent Communication Languages have a common basis
  speech act
• Syntax is well specified, but current research is
  on describing semantics