P. Maes. Intelligent software. Scienti  Home/Search   Document Not in Database   Summary   Related Articles   Check

....our system to be integrated into already existing transaction based systems such as real vendors and financial institutions. Autonomous and intelligent agents have been actively researched for many years. The predominant goal of this work is to help people manage information and work overload [Maes95] [Bradshaw97] Early work in Distributed AI, focused on multi agent collaboration to help solve general problems [Demazeau90] The field of Decentralized AI shifted this focus slightly to the interactions among agents with different motivations. The underlying concept, however, was still to ....

P. Maes. Intelligent Software. Scientific American, Vol. 273, No. 3, pp. 84-86. Scientific American, Inc., September 1995.

....search engines for agents is perhaps an example of a case where using agents as a user interface metaphor confuses the user rather than helps him. This consideration holds for basic applications of so called push technologies as well. It is easy to agree with Pattie Maes in her stating that [26] [ current commercially available agents barely justify the name . In the same manner, care should be taken when ascribing terms such as beliefs, obligations, actions, etc, to agents. Such ascription should be done only if it is useful in determining, understanding, or analysing the ....

P. Maes. Intelligent software. Scientific American, 273(3), 1995.

....agents. Thus, in most cases, we do not consider them intelligent agents. What Isn t An Agent Intelligent agent technology has been the victim of hype and exaggeration. Patti Maes notes that current commercially available agents barely justify the name agent yet al..one the adjective intelligent [Maes, 1995)] MIT researcher Foner argues: I find little justification for most of the commercial offerings that call themselves agents. Most of them tend to excessively anthropomorphize the software and then conclude that it must be an agent because of that very anthropomorphization, while ....

Maes, P. (1995). Intelligent Software. Scientific American, 273(3).

....generality of the model, validation must be made according to this general model, and, more importantly, future modifications must consist of coherent revisions, not extensional patches to the model. With the appearance of new approaches, such as adaptive software [28] or intelligent software [30], which include techniques and languages for further generalisation, an empirical and theoretical study of when a generalisation of the model is useful and how it should be done seems necessary. The flippancy here would be to start from scratch or to reinvent the wheel. As we will see in the ....

....etc. In summary, our analogy also shows that until machine intelligence (and ML) approaches human ability more closely, fully automated programming will remain a fallacy. In the meantime, in accordance with the analogy presented here and in an effort to reach the Utopia of intelligent software [30], a more prosperous methodology for software construction could be devised from the nascent predictive software . Acknowledgements First of all, the authors would like to thank Jess Alcolea for the introduction to Fetzer s work. Other people have helped to develop these ideas: Jaume Agust, ....

Maes, P. "Intelligent Software" Scientific American 273 (3), September, 1995.

....programs that access autonomously heterogeneous, geographically dispersed data and support the user in selecting relevant information. Such programs protocol and evaluate user activities. New or shifted interests are automatically recognized and taken into account for subsequent searches [Maes 1995]. Ideally, software agents release users from routine activities leaving more time for productive and creative work. Already 30 years ago software pioneer Douglas Engelbart had the vision that such applications could selectively augment human intellect leading to a symbiotic relationship of man ....

Maes, Pattie. Intelligent Software. Scientific American, 273(3):84--86, 1995.

....that a particular user has an assistant named, say, George, that an assistant should know the boss s meeting schedule and that a message containing the word meeting may contain scheduling information. With this knowledge, the agent would deduce that it should forward a copy of the message[23]. Softbots University of Washington [13] Similarly, Softbots serve their owners over the Internet. It is a fullyimplemented AI agent that uses a UNIX shell and the World Wide Web to interact with a wide range of internet resources. Its effectors include ftp, telnet, mail, and numerous file ....

P. Maes, Intelligent Software, Scientific American, Vol. 273, No. 3, pp. 84-86, September, 1995.

....into E 450 already existing transaction based systems such as real vendors and financial institutions. E 371 Autonomous and intelligent agents have been actively researched for many years. The E 450 predominant goal of this work is to help people manage information and work overload E 421 [Maes95] [Bradshaw97] Early work in Distributed AI, focused on multi agent collaboration to E 459 help solve global problems [Demazeau90] Decentralized AI shifted this focus slightly to the E 454 interactions among agents with different motivations. The underlying concept, however, was still E 450 to ....

P. Maes. Intelligent Software. Scientific American, Vol. 273, No. 3, pp. 84-86.<E-404> Scientific American, Inc., September 1995.<E-177>

....Department, MS 0781 University of Southern California Los Angeles, CA 90089 0781, USA mataric cs.usc.edu 1 Introduction and Motivation We describe the Multi Agent Schedule Calendar (MASC) learning system that assists a user in scheduling meetings. This work is inspired by Maes and Kozierok [1, 2, 3] and Mitchell et al. 4] Our contributions are as follows: a) we show how memory based learning can be sped up by maintaining aggregate information; b) we present an architecture for building personal agents organized around a division of labor between the learning and non learning components ....

P. Maes, "Intelligent Software", Scientific American, September 1995.

....future work includes enabling analysers to learn patterns from observing user behaviour. Such analysers will be capable of acting autonomously on behalf of the user and be adaptive to new circumstances and requirements. Similar ideas have already been explored in related work on software agents [9]. Further work will investigate the possibility of using cameras and microphones in the context of location devices. By appropriately processing video and audio data, cameras and microphones can be used as sensors. Potential uses involve tracking and locating people as well as augmenting other ....

P. Maes. Intelligent software. Scientific American (September 1995) 27,3, 84-86.

....to achieve some goal, as with genetic algorithms, one has to find a dynamics , or interaction loop or servo loop, involving the system and the environment which will converge towards the desired goal. The interaction process only comes to a rest (or a fixed pattern) when the goals are achieved (Maes, 1991b, 50). This would not only be time consuming, but it also smacks of trial and error with all its attendant problems. Fourthly, how are such systems extended, scaled up or debugged What happens if the environment is changed Even Brooks (1991a) acknowledges that such 29 questions are frequently ....

Maes, P. (1995a), "Intelligent Software", Scientific American 273 (3), September.

....based interface solutions and particularly to the personified type, claiming they remove user control and are distracting. Of course a large number of people fall between these two extremes, one prominent exponent of the agent based approach who doesn t try to highly anthropomorphise agents is Maes [Maes 1997]. 6 Of course there exists a myriad of different possible technologies and architectures which could be used to implement any kind of agent based interface system, some of which are described in the rest of this document. 3.2 How To Build a User Model One of the principal problems in ....

P. Maes. Intelligent Software. In Proceedings of Intelligent User Interfaces 1997, Orlando Florida.

....make them more user friendly. Maes [16] defines them as: Autonomous agents are computational systems that inhabit some complex dynamic environment, sense and act autonomously in this environment, and by doing so realize a set of goals or tasks for which they are designed. Maes has done much work [16, 18, 17] in the field of agents that employ AI techniques to perform tasks for the user. The agents usually learn from users behaviour and aim to automate repetitive tasks. Another definition given by researchers at MIT, Stanford and AT T [3] is: Agents assist users in a range of daily, mundane ....

....The agent can also offer to automate repetitive tasks for the user. An interface agent learns by observing and imitating the user, either receiving negative and positive feedback from the user, receiving explicit instructions from the user, or by communicating with other agents. According to Maes [16, 18, 17] the goal is to move away from the paradigm of a user initiated interaction and to delegate some tasks to interface agents. There are several functional examples of interface agents. These include ffl Maxims (Electronic Mail Agent) 17] This agent helps the user with email. It continuously looks ....

[Article contains additional citation context not shown here]

P. Maes. Intelligent software. Scientific American, 273:84--86, September 1995.

....a comprehensive review that would be too large for any single article. Instead, it focuses on five issues which we have found to be the ones most often encountered by designers or most often raised by users of our systems, based on our six years of experience designing several prototype agents [12]. These issues deal not with hard technical formulae such as Fitt s Law nor even with well explored interface considerations such as menu layout or dialog design. Instead these issues arise as a result of the attempt to create an interface with a different interaction paradigm. Specifically, these ....

Maes, Pattie. "Intelligent Software," Scientific American, Vol. 273, No.3, pp. 84-86, September 1995.

....their competence by adapting to the user s interests which may change over time while at the same time explore new domains that may be of interest to the user. We are witnessing a paradigm shift in human computer interaction from direct manipulation of computer systems to indirect management [17] in which agents play an important role. Some interesting experiments which change the way transactions normally occur (and hence raise many issues) include BargainFinder and Fido [6] Both of these systems present the concept of an agent which can shop for the best price for a good on behalf of a ....

P. Maes 1995. Intelligent Software.» Scientific American, Vol. 273, No. 3, pp. 84-86. Sept. 1995.

....involving thousands of users interacting with the agent marketplace E 349 over a long period of time, and speculate on the long range impact of this E 382 technology upon society and the economy. E 185 1. Introduction E 38 Software agents help people with time consuming activities [Maes95]. In the past, a E 428 range of roles for agents have been explored such as filtering information, automating E 380 repetitive tasks, and making referrals and introductions [PAAM96] Bradshaw97] E 362 One increasingly popular application for software agents is electronic commerce, E 382 ....

....of the Agent marketplace E 165 system one another, haggling and trying to find the best possible deal on behalf of their E 411 users, all in parallel. E 75 3.3. Selling and Buying Agents E 148 Software agents are long livedprograms that perform some task on behalf of a user E 413 [Maes95]. In our case, the agents try to buy or sell some good for the best possible E 393 price on behalf of the user which created them. To do this, the agents negotiate with E 411 other users agents in the marketplace, trying to find the best deal subject to a set of E 423 user specified ....

[Article contains additional citation context not shown here]

P. Maes. "Intelligent Software." Scientific American, Vol. 273, No. 3, pp.<E-415> 84-86. Scientific American, Inc., September 1995.<E-232>

No context found.

P. Maes. Intelligent software. Scienti

No context found.

LNCS, 1000, 46--61. Maes, P., 1995. Intelligent Software. Scientific American, 273, 84--86.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC