Carberry, S.; Chu-Carroll, J.; and Elzer, S. 1999. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence Journal 15(3):185-- 217.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....[6, 3] Additionally, preference strength can be estimated depending on in what conversational circumstance a user states her preferences. For example, a preference given as a reason for rejecting a recommendation has higher preference strength than a preference given as answer to a system query [7]. 3 Adaptive Recommendation Strategies Every user has her own way of thinking about movie preferences. Some people are highly biased on genres, whereas others may want to watch all kinds of movies as long as they star their favorite actor, etc. This is highly individual, and providing a ....

....that allows for dialogues such as the one in Table 1. AdFilm gathers data about preferences to be used for content based recommendations, and title preferences to be used for cf recommendations. The system assesses preference strengths depending on the conversational circumstance, as outlined by [7]. AdFilm also gathers preferences about the dialogue strategy. In the example dialogue in Table 1 the user is reluctant to provide a favorite genre, and prefer to volunteer a favorite movie. In subsequent sessions, AdFilm will ask this user for more favorite movies, instead of bothering him with ....

S. Carberry, J. Chu-Carroll, and S. Elzer, "Constructing and utilizing a model of user preferences in collaborative consultation dialogues," Computational Intelligence, vol. 15, no. 3, pp. 185--217, 1999.

.... her needs, interests and preferences (see [48, 36, 19, 34] and of using strategies for adapting its behavior to each specific user (see [8, 10, 35] Adaptive systems have been designed and applied in different areas, such as intelligent tutoring [12, 15, 23, 28] access to information sources [40, 26, 16, 1, 18, 39, 49], electronic catalogues [42, 13, 38, 31, 20] health care assistance [29, 7, 17] information filtering and recommender systems [33, 4, 43, 6, 9, 5, 27, 21] These systems make use of different techniques as regards both user modeling and adaptation. For example, a model of a user can be generated ....

S. Carberry, J. Chu-Carroll, and S. Elzer. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence, 15(3):185--217, 1999.

.... user is to classes of users is not new; for instance, see the exploitation of user communities in (Orwant, 1995) The linear combination of the stereotypical predictions applied in our system has been exploited in other user modeling systems, such as those described in (Linden et al. 1997) and (Carberry et al. 1999). For instance, consider the Domain Expertise family and the preference for products ease of use. Figure 6 reports the information about the prediction of each stereotype on the user preference. If the classification process has produced the following matching degrees: Novice: 0.7; ....

....contributions carried by the attributes of the items. As noticed in other work, this approach is well suited to modeling the way humans rate alternative options; e.g. see (Linden et al. 1997; umuai per www.tex; 4 01 2001; 18:23; no v. p. 20 Tailoring the Interaction with Users in Web Stores 21 Carberry et al. 1999). However, when deciding how closely a user matches a stereotype, the contribution of each classification datum must have a stronger impact than that supported by this theory: as already discussed, although the match must take into account all the classification data, the presence of totally ....

Carberry, S., J. Chu-Carroll, and S. Elzer: 1999, `Constructing and utilizing a model of user preferences in collaborative consultation dialogues'. Computational Intelligence 15(3), 185--217.

.... Bauer s work was specifically designed for domains where repeated patterns of actions can be observed; it will be interesting to explore how a mechanism such as Bauer s might be extended to preferences that do not represent repeated patterns of actions, such as the preferences recognized by Elzer[ECCC94, CCCE99] from a collaborative planning dialogue. Lesh[Les97] furthered this user tailored approach by investigating how observation of an agent s behavior might be used for more general adaptation of the plan recognition process. His experiments considered two types of adaptation: 1) addition or removal ....

Sandra Carberry, Jennifer Chu-Carroll, and Stephanie Elzer. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence, 1999.

.... obtained from the system s knowledge base; the executing agent s preferences are extracted from the system s model of the executing agent which is constructed incrementally by reasoning about the agent s utterances, the agent s acceptance rejection of proposals, and stereotypical user preferences [Carberry et al. 1999] . In ranking candidate instantiations of a proposed action, the ranking advisor takes into account both the strengths of the executing agent s preferences, as well as the closeness of the matches between the preferred values and the actual values of the attributes. We utilize the weighted ....

.... very weak) where the identified strength of a preference is based on the surface form of an utterance used to convey a preference, the conversational circumstances in which the preference is conveyed, patterns of acceptance and rejection of proposals, and a model of stereotypical user preferences [Carberry et al. 1999] . The closeness of a match indicates how well the actual value of an attribute matches the value preferred by the executing agent; thus it corresponds to the score assigned to a candidate attribute pair in the weighted additive rule. The possible values for the closeness of a match are exact, ....

[Article contains additional citation context not shown here]

Sandra Carberry, Jennifer Chu-Carroll, and Stephanie Elzer. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence, 15(3), 1999.

No context found.

Carberry, S.; Chu-Carroll, J.; and Elzer, S. 1999. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence Journal 15(3):185-- 217.

No context found.

S. Carberry, J. Chu-Carroll, and S. Elzer. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence, 15(3):185--217, 1999.

No context found.

S. Carberry, J. Chu-Carroll, and S. Elzer. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence, 15(3):185--217, 1999.

No context found.

S. Carberry, J. Chu-Carroll, and S Elzer. Constructing and utilizing a model of user preferences in collaborative consultation dialogues. Computational Intelligence, 15(3):185--217, 1999.

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