Agent technologies needed for full-scale FIPA '97 applications

FIPA has selected

• Personal Assistant
• Personal Travel Assistance
• Audio-Visual Entertainment and Broadcasting
• Network Provisioning and Management

as target for agent technologies to be specified in 1997.

The following list of technologies is believed to be needed for a full-scale implementation of such applications.

Communication

• Model of communication
• Low-level Communication Protocols
• Rules of agent behaviour (e.g. commitments)
• Call Management (T)

Human-Agent Dialogue

• NL understanding and generation
• Cooperative responses
• Selecting and accessing services (domain-specific)

Agent-Agent Interaction

• Interactions of agent based subsystems.
• Coordination/negotiation with others
  • message types
  • message contents
  • interaction protocols (e.g. contract net, negotiation, trading models, service contracts)
  • Establishment of authority
• Generic Communication and Coordination languages
• Agent Profile Description Language

Agent software communication (API's)

• Reference framework e.g. definition of domain specific vocabulary, CORBA

Domain specific

Generic database retrieval mechanisms

Programming API for linking to standard languages

Generic Internal Agent Interfaces (P)

Input/Output

Multimedia Processing

Text Processing

Audio/Speech

• Speech recognition
• Speech synthesis

Visual

• Image display
• Graphics display

HW Processing.

GPS (positioning systems)

DSM (e.g. VCR)

Information representation

• Information content representation and regulation
• Knowledge representation reference models
  • domain-specific knowledge representation and dictionary (for application areas)
  • Generic knowledge representation (e.g. for actions, time, location, costs...)
  • Knowledge processing (indexing, summarising, filtering, ...)
  • User profile representation and management
  • Acquaintance models
• Reasoning Capabilities
  • Inference Engine
  • Adaptation, Learning
  • Domain-specific ontology

Privacy / Security

• Agent security (e.g. certification, security level, signalling)
  • agent authentication and authorisation
  • agent creation, modification and deletion
  • transfer of authorisation
  • proof of agent identity and integrity
• Non-repudiation
• Transaction security and privacy
  • message integrity
  • cash transactions (e.g. via trusted third parties)
• Environment (e.g. other agents, user, etc.) security and privacy

User authentication

System Support and Management

• System Configuration (agents, network, ...)
• Execution Models
• Administration and Policing

Registration, Update, De-registration with Directory Agent

Agent Monitoring (esp. of mobile agents)

Mechanisms for Agent Persistance