`Chapter Intelligent
`
`
`
`
`f
`
`2%
`
`gas a,
`3%.,
`A M: :ggfig r emgigs???
`
`
`
`
`giggle notion of an intelligent agent that dutifully serves computer users
`has fascinated us since the invention of the digital computer. The computer
`
`Emerac in the Katharine Hepburn and Spencer Tracy 1957 movie, Desk Set is
`
`
`
`
`
`an example of such an agent. In this movie, Spencer Tracy plays an efficiency
`
`
` expert who is contracted to computerize the various business units of Federal
`
` Broadcasting Agency with Emerac. One of his projects is to encode the knowl-
`
` edge of every textbook in the research departmént managed by Miss Bunny
`
` Watson. To test the knowledge of Emerac, Miss Watson poses the following
`
`question to Emerac:
`
` Computer, what is the annual damage done by the spruce hudworm to
`
`American forests?
`
`
`
`So how far have computers come in being able to answer the question that
`Miss Watson posed to Emerac in 1957? Well, not close enough if we require
`
`
` understanding the semantics of the question, as natural language understanding
`
` has proven to be a tougher problem than anticipated in computer science. Close
`
` enough, however, if we can point Miss Watson to a related URL, submitting
`
` Petitioner Nissan North America, Inc. — Exhibit 1010 — Page 1
`
`-
`
`Petitioner Nissan North America, Inc. - Exhibit 1010 - Page 1
`
`
`
`Nonetheless, we are ge
`seems to imply that we ar
`
`e current hype in software agents
`tting closer (although th
`e already there). In this book, we will describe what is
`achievable in the short term with software agent technology, and show you how
`agent technology can change your business for the better. We start with a definition
`of an agent. The Webster’s New World Dictionary (Guralnik, 1970) definition is:
`or thing that acts or is capable of acting or is em
`
`Agent: A person
`act, for another.
`tion of an agent points out its two key attributes:
`
`This general defini
`
`powered to
`
`I An agent does things.
`lf of someone or something.
`I An agent acts on beha
`incorporate these two c
`e on computers always
`1 definition of a software
`no widely accepted universa
`siness
`of an agent will suffice to discuss the bu
`
`entral
`
`the Internet search
`” inquiry into one of
`/pubaff/log/OOO9 .html.
`engines to get this Web address: http://www.odf.state.or.us
`gent assistant has been refined over the
`Although our concept of an intelli
`rick movie, 2001: A Space Odyssey, in the
`HAL in the Stanley Kub
`The Knowledge Navigator, in
`years——from
`uter’s video,
`ally build has not kept
`late 1960s, to Phil in Apple Comp
`ts we can actu
`the late 19809—the kind of intelligent agen
`'up with our expectations.
`
`minimal characteristics:
`
`Intelligent agents that resid
`attributes. Currently, there is
`agent. The following definition
`applications of agents:
`A computing entity that
`
`performs user delegate
`
`d tasks
`
`hor where the
`
`Software agent:
`autonomously.
`Our definition of an agent implies a personal assistant metap
`tasks on behalf of a user. Mail filtering agents, information retrieval
`agent performs
`all fit this definition.
`agents, and des
`
`ktop automation agents
`
`Attributes of Intelligent Agents
`gent implies that the agent possesses the follow
`
`Our definition of software a
`
`ing
`
`Petitioner Nissan North America, Inc. - Exhibit 1010 - Page 2
`
`
`
`
`
`5
`
`Intelligent Agents
`
`I Delegation: The agent performs a set of tasks on behalf of a user (or other
`agents) that are explicitly approved by the user.
`
`
`
`I Communication skills: The agent needs to be able to interact with the
`
`user to receive task delegation instructions, and inform task status and
`completion through‘an agent—user interface or through an agent com-
`munication language.
`
`I Autonomy: The agent operates without direct intervention (e.g., in the back-
`ground) to the extent of the user’s specified delegation. The autonomy
`attribute of an agent can range from being able to initiate a nightly backup
`to negotiating the best price of a product for the user.
`
`I Monitoring: The agent needs to be able to monitor its environment in order
`to be able to perform tasks autonomously.
`
`I Actuation: The agent needs to be able to affect its environment via an actua—
`
`tion mechanism for autonomous operation.
`
`I
`
`Intelligence: The agent needs to be able to interpret the monitored events to
`make appropriate actuation decisions for autonomous operation.
`
`In essence, the concept of an agent introduces an indirect management metaphor
`in a computerized environment (Alan, 1984) to supplement today’s mainstream
`style of direct manipulation metaphor via graphical user interfaces. This functional—
`ity enables users to accomplish tasks in a collaborative manner with software
`agents without having to be physically present in front of their machines.
`
`:ml.
`
`the
`
`kept
`
`: agents
`rat is
`
`how
`
`lefinition
`
`:ion is:
`
`:red to
`
`:entral
`
`software
`ness
`
`
`
`(S
`
`re the
`
`on retrieval
`
`lowing
`
`
`
`
`In addition to the basic attributes of agents discussed above, agents may have
`additional attributes such as mobility, security, personality, and others that will be
`discussed in later chapters.
`
`The origins of agent technology are rooted in the computational intelligence,
`software engineering, and human interface domains. Computational intelligence for
`intelligent agents draws from the fields of intentional systems, production systems,
`reasoning theory, and neural networks. Software engineering for intelligent agents
`Covers the areas of on-line monitoring, high-level event inference, remote actuation,
`image and speech processing, and distributed objects. Human~computer interaction
`for intelligent agents comes from the fields of cognitive engineering, user modeling,
`man-machine experiments, intelligent tutoring systems, and computer vision, as
`illustrated in Figure 1.1.
`
`Petitioner Nissan North America, Inc. — Exhibit 1010 — Page 3
`
`Petitioner Nissan North America, Inc. - Exhibit 1010 - Page 3
`
`