Group: types of user interfaces
Topic: database agents
Topic: ease of use
Topic: frame problem
Topic: help in UserInterface
Topic: intelligent machines
Topic: man-machine symbiosis
Topic: people vs. computers
Topic: plan-based task scheduling
Topic: problem of information overload
Topic: problem of other minds
Topic: simulation
Topic: using a computer as a communication/information medium
Topic: what is a computer
| |
Summary
Computers can appear to be an intelligent agent. For example, they can respond to queries, simulate emotions, and interact with users. Video games and simulations often include computer-controlled opponents. A computerized help system may include an animated assistant. Computer-assisted instruction can help guide a user through a course.
An interesting twist is adding people to the system. For example, an intelligent help system can use a person to read the queries and type answers. Surgeons can work remotely using robotics. With the Turk system, the whole Web can help solve intelligent problems.
A common argument is that computers must be intelligent since humans are intelligent. An alternative view is that computers work well as a communication media, but work poorly as intelligent agents. Computerized intelligence is forced, tripping up the user. (cbb 6/06)
Subtopic: intelligent agent
Quote: 'Creatures' are autonomous, intelligent, mobile agents that co-exist in the world with humans [»brooRA1_1991]
| Quote: anthropomorphic agents useful as a motivator or coach while learning; also entertainment [»schaH1_2001]
| Subtopic: intelligent computers
Quote: natural language understanding systems answer questions like a human; uses a internal representation of text [»shasD_1985]
| Quote: should the user or the system decide when assistance is required? [»rellN3_1981]
| Quote: an intelligent tutor is the goal of a knowledge exploration system; uses demonstration, dialogue, paraphrase, answers and questions [»shasD12_1987]
| Quote: associations is not enough: need searching by semantics, need derived information [»vandA7_1988]
| Quote: use predetermined troubleshooting sequences; generalized AI reasoning component; and document browser for advanced users [»feinS1_1982]
| Quote: should adapt a graph layout to the device characteristics, annotation style, intended use, etc. [»vinbA_1982]
| Quote: select Planner procedures by goal patterns [»hewiC4_1985]
| Quote: the mind of a computer includes internal memory and processing elements [»ingaDH8_1981a]
| Quote: if a programming language was an extension of a human, then little would be needed from the programmer [»carrJW_1985]
| Quote: a program is an extrapolation of its human creator; like Maelzel's chess player box [»carrJW_1985]
| Quote: Aristotle: could do without subordinates if inanimate instruments could do their own work by command or intelligent anticipation [»aris_322]
| Quote: SmartHelp simulated intelligent help with a person, but none of the participants voiced suspicions [»carrJM9_1988]
| Quote: a computer can be as intelligent as a human being [»barrJD_1986]
| Quote: agents are semi-intelligent programs that carry out your goals; e.g., personal news reporters that scour the networks for relevant items [»lindO_1992]
| Quote: a system should be intelligent, i.e., help the user and reduce workload; e.g., check all inputs for validity [»potaWA_1979]
| Quote: a resourceful system will automatically create programs to deal with contingencies [»abboRJ3_1990]
| Quote: Consul builds a model of the user and the application; it pursues a dialogue if the model is incomplete for a question [»markW_1982]
| Quote: Consul parses requests using its knowledge representation; uses inference rules to identify an operation or produce an explanation [»markW1_1981]
| Quote: computers provide an active medium that can respond to queries like an individual teacher [»kayA3_1977]
| Quote: natural language understanding systems answer questions like a human; uses a internal representation of text [»shasD_1985]
| Quote: with pervasive networking will want access to a trillion resources; use agents
| Quote: though intelligent digital computers appear impossible, intelligent, analog computers may be possible
| Quote: categorizing information is difficult; systems can help by automatic classification [»maloTW6_1982]
| Subtopic: simulation
Quote: Smalltalk's heritage is simulation; the description and simulation of real-world models as computer models [»goldA10_1995]
| Quote: a microworld is a well-defined, limited world with interesting events and educational ideas; e.g., blocks world [»goldEP8_1982]
| Quote: use Sketchpad to animate a face or pedal a bike [»suthIE5_1963]
| Quote: a computer is a universal medium in that it can simulate any other medium [»kayA3_1977]
| Subtopic: computers as conversations
Quote: most user interfaces have been built on the conversational metaphor, an interface, an intermediary, to a hidden world; denies direct engagement [»hutcEL_1986]
| Quote: paradigm of society of experts communicating with messages; based on actor model of computation [»atkiR_1977]
| Quote: naturally design processes by message passing, blocking semantics, servers, proprietors, administrators, notifiers, and couriers; anthropomorphize [»gentWM5_1981]
| Subtopic: intelligent machines must exist
Quote: man's physical existence is proof that intelligent machines exist; but not that man can make them [»whitI11_1988]
| Quote: either the Turing-Church thesis implies that all physically realizable processes can be duplicated by computers or it does not limit physical processes [»conrM5_1985]
| Quote: a computer can "figure out", "recognize", or "conclude". This is no different than "multiplication" by addition and shifting, or "addition" by opening and closing switches [»dennDC_1978]
| Quote: just as a programmer can explain how a computer multiplies numbers, he can explain how a computer recognizes the best chess move
| Subtopic: computer as channel vs. intelligent agent
Quote: the human-computer interface should be a flexible tool instead of an intelligent agent
| Quote: a computer should be seen as a medium for communication instead of an autonomous agent [»clanWJ1_1987]
| Quote: human-human communication is a poor model for human-computer interaction [»shneB_1987]
| Quote: predictable user interfaces provide a clear mental model, repeatable actions, error recovery, user configuration; intelligent machines do not [»shneB1_1993]
| Subtopic: do not anthropomorphize
Quote: a commandment: thou shalt not create machines in the image of man
| Quote: never anthropomorphize programs or equipment; think of programs in terms of computational behavior based on a model [»dijkEW12_1989]
| Quote: people should know when they are communicating with a machine instead of a person [»nygaK_1981]
| Quote: if we personalize a computer system, then we expect the system to "do it on its own" [»holtAW3_1967]
| Quote: ELIZA will respond "How long have you been swallowing poison?" instead of responding appropriately [»winoT_1986]
| Quote: computers are incapable of making commitments and hence cannot enter into language [»winoT_1986]
| Subtopic: problems with intelligent agents
Quote: most computational theories of mind assume a homunculus; even if reduce to 0/1's still need a homunculus to interpret them [»searJR_1992]
| Quote: a mechanical computer does not literally follow rules; it only behaves as if it does [»searJR_1992]
| Quote: cognitive science can never be a natural science since computation is observer-relative
|
Related Topics
Group: types of user interfaces (23 topics, 298 quotes)
Topic: database agents (10 items)
Topic: ease of use (47 items)
Topic: frame problem (13 items)
Topic: help in UserInterface (33 items)
Topic: intelligent machines (28 items)
Topic: man-machine symbiosis (46 items)
Topic: people vs. computers (55 items)
Topic: plan-based task scheduling (13 items)
Topic: problem of information overload (23 items)
Topic: problem of other minds (11 items)
Topic: simulation (35 items)
Topic: using a computer as a communication/information medium (32 items)
Topic: what is a computer (62 items)
|