Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th National Conference on Artificial Intelligence - AAAI-93, pages 459--464, Menlo Park, CA, USA, July 1993. AAAI Press.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....each individual decision point acts as an independent classification problem and receives its own direct feedback. It is not clear, however, how well this approach generalizes to situations in which the feedback available has less quantity or less direct connection to the actual decision made. In [4] decisions are also treated as separate classification problems. This system involves di#erent kinds of decisions (learning over interface actions rather than over parameter values) and has a di#erent set of learning algorithms (including reinforcement learning) However, this system, like CAP, ....

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th National Conference on Artificial Intelligence - AAAI-93, pages 459--464, Menlo Park, CA, USA, July 1993. AAAI Press.

....that have been labeled by the user. Thus, the filter requires the user to give relevance feedback [Roc71, SB90] for the seen information items. It will then generate a predictive model (user profile) based on this feedback. The feedback can be acquired either implicitly by observing the user [MK93, AFJM95, Lie95, BP99] or explicitly by asking the user to rate the seen information item [PMB96, CS98] Given these labeled information items, it is then possible to apply ML methods to generate a predictive model that, when given a new information item, will predict whether the new item is ....

....filtering system. 1.3 Acquiring the User Profile This thesis will focus on the use of content based filters. As mentioned in the previous section, currently used acquisition methods involve asking the user for a list of keywords [HK70, FD92] observing the user through normal day usage [MK93, AFJM95, Lie95, BP99] or by asking the user for relevance feedback on sample data [Roc71, SB90, FD92, SSS98, BP99] These techniques work well with user interests that are based on criteria that are directly evaluable: seeing a piece of information, a user can directly assess whether it is ....

[Article contains additional citation context not shown here]

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the Eleventh National Conference on Artificial Intelligence, pages 459--465, Menlo Park, CA, 1993. AAAI Press. 161

....by the user models which are also rigid. However, recent research in machine learning and artificial life[14] suggests that various techniques can be applied to make interface systems learn or evolve to fit to each user s needs. For example, Maes proposes learning interface agents [13] [15] which gradually learn to assist the individual user by observing the user s actions, getting user feedback, and explicit training, based on memory based learning approach. If a system can not only learn, but evolve to a more powerful learning system, it can be a very powerful framework for ....

Maes, P. Learning interface agents. In Proceedings of the 1994 Friend21 International Symposium on Next Generation Human Interface (February 1994).

.... are software components characterized by autonomy (to act on their own) reactiveness (to process external events) proactiveness (to reach goals) cooperation (to efficiently and effectively solve tasks) adaptation (to learn by experience) and mobility (migration to new places) see e.g. 26] [31] for details on agent technology) Messages are highly structured and must satisfy standardized communicative (speech) acts which define the type and the content of the messages (agent communication language (ACL) like FIPA ACL [13] or KQML [32, 33] The order of exchanging messages of a certain ....

P. Maes, and R. Kozierok, Learning Interface Agents. In Proceedings of the 11th Nat Conf on Artificial Intelligence, AAAI, MIT-Press/AAAI-Press, 1993.

....may provide this capability. 4 Adapting to User Behavior The AI community has a long standing interest in adaptive interfaces. The Calendar s Apprentice [Dent et al. 1992] used machine learning to predict meeting location and durations. This research and similar work on email classification [Maes and Kozierok, 1993] led to an important principle about the incorporation of imperfect behavioral predictions in an interface: defaults are an effective way to minimize the cost to the user of (inevitable) poor predictions. Horvitz s decision theoretic framework [Horvitz, 1999] resulted in additional principles: ....

Pattie Maes and Robyn Kozierok. Learning interface agents. AAAI-93, pages 459--465, 1993.

....does not change often, whereas information filtering assumes a constant stream of time sensitive documents. Lastly, filtering is the act of removing irrelevant items from this stream, whereas IR is the act of finding relevant items in the database. hfformation filtering as been applied to e mail [Maes and Kozierok, 1993] and Usenet news groups [Sheth and Maes, 1993; Yan and Garcia Molina, 1995] using relevance feedback to build user profiles. Several commercial services now offer simple filtering systems for proprietary collections of documents (e.g. ZiffDavis Personal View, the San Jose Mercury News ....

....a constant stream of time sensitive documents. Lastly, filtering is the act of removing irrelevant items from this stream, whereas IR is the act of finding relevant items in the database. hfformation filtering as been applied to e mail [Maes and Kozierok, 1993] and Usenet news groups [Sheth and Maes, 1993; Yan and Garcia Molina, 1995] using relevance feedback to build user profiles. Several commercial services now offer simple filtering systems for proprietary collections of documents (e.g. ZiffDavis Personal View, the San Jose Mercury News NewsHound) This kind of application is inherently ....

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proccedings of the 11 tn National Conference on Artificial Intelligence, pages 459-465, Washington, DC, July 1993.

....These agents coordinate the transfer and processing of information among people and other agents. Application agents include: a) Time schedulers whichschedule group or individual meetings among a set of people by negotiating among their personal agents to maximize some convenience measure. [Maes and Kozierok, 1993, Dent et al. 1992, Kautz et al. 1994] b) Text processing systems whichallow complex processing of documents involving many people at different sites. c) Receptionist agents which accept requests and determine their appropriate destinations byinteracting with other agents (and perhaps people ....

....agents can do it for them without requiring that a person bother with the intricate constraint balancing inherent in meeting scheduling (and perhaps without anyone s feelings getting hurt) In fact, meeting scheduling is the most popular software agent negotiation application. Kozierok, 1993, Maes and Kozierok, 1993] schedules group meetings, Kautz et al. 1994] schedules meetings between individuals, and [Dent et al. 1992] does both (and more) The backbone of all of these systems could be implemented in SodaBot. However, the actual scheduling processes would require external applications. For example, ....

Maes, Pattie and Kozierok, Robyn. Learning interface agents. In Proceedings of the Eleventh National ConferenceonArtificial Intelligence, AAAI-93,Washington D.C., p459-464. 1993.

....such a Society of Agents. We can classify the existing agent research mainly into two groups. The first group concentrate on providing computer based assistant to users, helping them perform tedious, repetitive, or time consuming tasks more easily and efficiently (also known as Interface Agents) [10, 4, 5] (Fig.1 a) The second group focus on how agents attached to providers of information or services can work cooperatively in order to facilitate access to what they pro vide (mainly studied by DAI researches) 11, 14, 7, 17] Fig. lb) In this work, our aim is to integrate the existing ....

....operators may be needed to allow search of related topics in parallel. For (iii) the presence of Manager and replicas going to the different places in parallel could realize this feature easily. 5. 2 Related Works Some related works regarding agents include those regarding Interface Agents [10, 4, 5] which correspond to User Agent (UA) of our model. Another group of related works are those concerning Distributed Artificial Intelli gence (DAI) which mainly study the cooperation between different Machine Agents [11, 14, 7, 17] One additional group of works include Migrating Agents (e.g ....

P. Maes and R. Kozierok. Learning interface agents. In Proc. 11th National Conference on Artificial Intelligence, 1993.

....learning techniques, rules can be induced. This will result in a personalised rulebase with no additional work from the user. This final approach is being used in a number of interface agents to help the user in performing organisational tasks, such as calendar management [Dent et al. 1992; Maes Kozierok 1993], exploring newsgroups for interesting articles [Sheth 1994] and Email filters [Metral 1993] CAP (Calendar APprentice) is a personal learning apprentice which assists in managing a meeting calendar [Dent et al. 1992] The calendar manager is used by filling in parameters for a given meeting. ....

....of classes the ruleset can cover. It was noted that as the coverage of the rules increased, the overall accuracy decreased. This could indicate that by making rules cover more possible classes, they can become over generalised and hence less accurate. The work at MIT by Kozierok [Kozierok Maes 1993; Maes Kozierok 1993] used the machine learning method Memory Based Reasoning [Stanfill Waltz 1986] An unusual aspect of this work was the use of caricatures to provide feedback to the user of the current state of the agent. They addressed the problem of trusting the agent by generating a ....

[Article contains additional citation context not shown here]

P.Maes & R.Kozierok; Learning Interface Agents. In AAAI-93 Chapter 9. References 57 Proceedings, Eleventh National Conference on Artificial Intellegence, AAAI Press, 1993, 459-465

....interaction or use the current wrapper. Previous work on adaptive user interfaces provided some mechanisms for collaborating to re ne the learner s concept. Peridot [16] and Metamouse [14] are early PBD systems that request guidance from the user when generalizing actions. The mail clerk agent [13] learns by observing the user, from explicit feedback, and by being trained. In addition, it compares its con dence against two user set thresholds ( tell me and do it ) to decide whether to initiate action, make a suggestion, or remain quiet. None of these systems describes a general interface ....

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the Fourteenth National Conf. on Articial Intelligence, pages 459-465, 1993.

....our system, the collaborative task is to teach SMARTedit how to perform a particular repetitive edit. Although beyond the scope of this paper, we have also constructed a mixed initiative interface for SMARTedit that could reduce user e#ort during a training session [53] Researchers such as Maes [30], Schlimmer [45] and many others have considered complementary approaches to adaptive interfaces, using machine learning techniques to predict the user s next actions. Space precludes a detailed discussion of the considerable work in this vein. 7.4. Plan recognition We view programming by ....

Maes, P. and R. Kozierok: 1993, `Learning interface agents'. In: Proceedings of AAAI-93. pp. 459--465.

....1 Unfortunately, no suitable papers were submitted in the area of user modeling. 0004 3702 00 see front matter 2000 Published by Elsevier Science B.V. All rights reserved. PII: S0004 3702(00)00013 8 2 A.Y. Levy, D.S. Weld Artificial Intelligence 118 (2000) 1 14 processing [86], netnews filtering [16,108] Web search [75] book and music recommendations [85,110] intrusion dection [73] and Web browsing recommendations [63,84,97, 98,124] For example, when run on data from a faculty member s scheduling behavior, the Calendar Apprentice [31,88] might learn rules such as ....

P. Maes, R. Kozierok, Learning interface agents, in: Proc. AAAI-93, Washington, DC, 1993, pp. 459--465.

....Complex Systems, No.108) at 1997. users with high precision and recall, because information space is large and users can not appropriately express their interests by themselves. In order to make information gathering re ect interests of a user, it is useful to equip a personalized interface agent [8][2] that accumulates interests of a user, applies them at information gathering and adapts to interest changes of the user, with information gathering tools. This paper proposes an extraction method of user s interests based on genetic algorithms (GA) 6] 5] from web documents gathered by ....

Maes, P., Kozierok, R.: Learning Interface Agents. Proceedings of AAAI-93, 459-465 (1993)

....such a Society of Agents. We can classify the existing agent research mainly into two groups. The first group concentrate on providing computer based assistant to users, helping them perform tedious, repetitive, or time consuming tasks more easily and efficiently (also known as Interface Agents) [7, 2, 3]. The second group focus on how agents attached to providers of information or services can work cooperatively in order to facilitate access to what they provide (mainly studied by DAI researches) 8, 11, 5, 13] In this work, our aim is to integrate the existing heterogeneous agents to build a ....

....operators may be needed to allow search of related topics in parallel. For (iii) the presence of Manager and replicas going to the different places in parallel could realize this feature easily. 5. 2 Related Works Some related works regarding agents include those regarding Interface Agents [7, 2, 3] which correspond to User Agent (UA) of our model. Another group of related works are those concerning Distributed Artificial Intelligence (DAI) which mainly study the cooperation between different Machine Agents [8, 11, 5, 13] One additional group of works include Migrating Agents (e.g ....

P. Maes and R. Kozierok. Learning interface agents. In Proc. 11th National Conference on Artificial Intelligence, 1993.

No context found.

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th National Conference on Artificial Intelligence - AAAI-93, pages 459--464, Menlo Park, CA, USA, July 1993. AAAI Press.

No context found.

Maes, P., Kozierok, R., 1993. Learning interface agents. In: Proceedings of the 11th National Conference on Artificial Intelligence, Washington, DC, USA. MIT-Press, AAAI-Press, Cambridge, MA, pp. 459--465.

No context found.

Maes, P., Kozierok, R.: Learning interface agents. In: Proceedings of the Eleventh National Conference on Artificial Intelligence (AAAI-93), Washington, DC (1993) 459--465.

No context found.

R. Kozierok and P. Maes. Learning interface agents. In Proceedings of the 11th National Conference on Artificial Intelligence (AAAI-93), pages 459--465. AAAI Press, 1993.

No context found.

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of AAAI93, pages 459-465, 1993.

No context found.

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th Nat Conf on Artificial Intelligence. AAAI, MIT-Press/AAAIPress, 1993.

No context found.

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th Nat Conf on Artificial Intelligence. AAAI, MIT-Press/AAAI-Press, 1993.

No context found.

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th Nat Conf on Artificial Intelligence. AAAI, MIT-Press/AAAI-Press, 1993.

No context found.

Pattie Maes and Robyn Kozierok. Learning interface agents. In Proceedings of the 11th Nat Conf on Artificial Intelligence. AAAI, MIT-Press/AAAI-Press, 1993.

No context found.

P. Maes and R. Kozierok. Learning interface agents. In Proceedings of the Eleventh National Conference on Artificial Intelligence (AAAI93, pages 459--465. AAAI Press/The MIT Press, 1993.

No context found.

Maes, P., and Kozierok, R. 1993. Learning interface agents. In Proceedings of the Eleventh National Conference on Artificial Intelligence, 459--465. Menlo Park, CA: AAAI Press.

First 50 documents  Next 50

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