L.A. Ramshaw. A three-level model for plan exploration. In Proceedings of ACL-91, pages 39--46, Berkeley, CA, June 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....city of xxx is considering filing for bankruptcy. U2: One of your mutual funds owns xxx bonds. Their separation of discourse plans allows them to be able to recognize such speech act like phenomena at a multi utterance level. Ramshaw During the same period of Lambert, Ramshaw [Ram89b, Ram89a, Ram91] proposed a di#erently separated model (which was later extended by Ardissono [ABL96] Instead of a discourse level, Ramshaw proposed an exploration level. The intuition is that some utterances are made simply in the attempt to explore a possible course of action, whereas others are explicitly ....

....may request that actions be executed without first formulating a specific plan (such as sending an ambulance to a heart attack victim) Future plan recognition systems will need to be able to recognize and deal with these kinds of domains. Although Ramshaw and Ardissono s work [Ram89b, Ram89a, Ram91, ABL96] provide hints at how one might begin to structure such a solution, this problem has not yet been addressed by the literature. 6.4 Incremental Understanding Nearly all current dialogue systems operate at an utterance level of granularity. Human dialogue is very di#erent [Cla97] In ....

[Article contains additional citation context not shown here]

Lance A. Ramshaw. A three-level model for plan exploration. In Proceedings of the 29th ACL, pages 39--46, Berkeley, CA, June 1991.

....city of xxx is considering ling for bankruptcy. U2: One of your mutual funds owns xxx bonds. Their separation of discourse plans allows them to be able to recognize such speech act like phenomena at a multi utterance level. Ramshaw During the same period of Lambert, Ramshaw [Ram89b, Ram89a, Ram91] proposed a di erently separated model (which was later extended by Ardissono [ABL96] Instead of a discourse level, Ramshaw proposed an exploration level. The intuition is that some utterances are made simply in the attempt to explore a possible course of action, whereas others are explicitly ....

....user may request that actions be executed without rst formulating a speci c plan (such as sending an ambulance to a heart attack victim) Future plan recognition systems will need to be able to recognize and deal with these kinds of domains. Although Ramshaw and Ardissono s work [Ram89b, Ram89a, Ram91, ABL96] provide hints at how one might begin to structure such a solution, this problem has not yet been addressed by the literature. 6.4 Incremental Understanding Nearly all current dialogue systems operate at an utterance level of granularity. Human dialogue is very di erent [Cla97] In ....

[Article contains additional citation context not shown here]

Lance A. Ramshaw. A three-level model for plan exploration. In Proceedings of the 29th ACL, pages 39-46, Berkeley, CA, June 1991.

....of discourse plans allows the recognition of speech act like phenomena such as warnings and surprise at a multi utterance level. The focus of this work was on the discourse level; no new work was done at the metalevel. Ramshaw At the same time as Lambert, Ramshaw ( Ramshaw, 1989b; Ramshaw, 1989a; Ramshaw, 1991] proposed a di#erently separated three level model. Instead of a discourse level, Ramshaw proposed an exploration level. The intuition is that some utterances are made simply in the attempt to explore a possible course of action, whereas others are explicitly made to attempt to execute a plan. ....

....and Acting The second main shortcoming of these models is the same as that we discussed above about SharedPlans: modeling collaboration about planning and acting. Previous intention recognition work has had the implicit assumption that either only planning was taking place in the dialogue (e.g. [Ramshaw, 1991; Chu Carroll and Carberry, 2000] with the plan to be executed at some later time or that only execution of some previously created plan was occurring (e.g. Ardissono et al. 1996] As we discussed in Section 2.1, in certain domains, agents collaborate about acting and planning, often ....

[Article contains additional citation context not shown here]

Lance A. Ramshaw, "A Three-Level Model for Plan Exploration," In Proceedings of the 29th ACL, pages 39--46, Berkeley, CA, June 1991.

....COLLAGEN, as it works on a subset of the SharedPlans formalism, also does not explicitly model most of our problem solving acts. Several dialogue systems have divided intention recognition into several di erent layers, although these layerings are at much di erent levels than our own. Ramshaw [14] analyzes intentions on three levels: domain, exploration, and discourse. Domain level actions are similar to our own domain level. The discourse level deals with communicative actions. The exploration level supports a limited amount of evaluations of actions and plans. These, however, cannot be ....

L.A. Ramshaw. A three-level model for plan exploration. In Proceedings of ACL-91, pages 39-46, Berkeley, CA, June 1991.

....script, a birthday party script, a football game script : ibid] each situation has its own script. Scripts can be used more than once. The use of scripts or schemata is pervasive in the field of planning and natural language understanding, both for plan recognition [Carberry, 1988; Ramshaw, 1991] and for plan synthesis [Allen, 1987; Rosenschein, 1981; Stefik, 1981b; Dean et al. 1988; Vere, 1983; Tate, 1977; Wilkins, 1984] among others) A good discussion of the sociological motivations for scripts is found in [Suchman, 1987] ffl Hierarchical decomposition. Researchers such as ....

Lance A. Ramshaw. A Three-Level Model for Plan Exploration. In ACL91, Proceedings of the 29th Meeting of the Association for Computational Linguistics, pages 39--46, 1991.

.... Several research efforts are investigating the nature and role of participants beliefs and intentions [Pollack 86, Cohen Levesque 90, Grosz Sidner 90, Lochbaum 91] and much effort is focused on the types of plans that underlie this type of discourse (see [Litman 85, Lambert Carberry 91, Ramshaw 91] Most of these theories postulate several levels of plans, each level handling a distinct phenomenon (discourse management, domain knowledge, etc. 4 Theory (DRT) Kamp 81] and that of [Polanyi 88, Reichman 85, Cohen 83, Heim 83] Extending beyond dialogue length discourse, Van Dijk 72] ....

Ramshaw, L.A. 1991. A Three-Level Model for Plan Exploration. Proceedings of the 29th Annual Conference of the Association of Computational Linguistics, Berkeley (39--46).

....turn taking, do not reflect face to face conversation. 2.3.1 Characterizing the Task Effects. Most studies on collaboration in computational linguistics have concentrated on dialogues that involve planning or scheduling tasks (among many others, see for example (Lochbaum, Grosz, and Sidner 1990; Ramshaw 1991; Sidner 1994; Walker 1996; Chu Carroll and Carberry 1998) It is not unprecedented to select a simple task taken out of a larger context in order to control the situation and potentially allow for a more objective analysis. In the case of the COCONUT task, we too opted for a simple ....

....collaborative dialogues. Whereas some researchers have taken grounding into account (Novick and Ward 1993; Traum 1994; Heeman and Hirst 1995) most of them, like us, have focused on level 4 in Clark s ladder of joint actions, i.e. proposals and their disposition in terms of acceptance or rejection (Ramshaw 1991; Lambert and Carberry 1992; Sidner 1994; Walker 1993; Walker 1996; Chu Carroll and Carberry 1998) Some of the work just mentioned models negotiation by means of discourse planners that represent actions at different levels. 22 For example, Lambert and Carberry (1991; 1992) postulate a problem ....

[Article contains additional citation context not shown here]

Ramshaw, Lance A. 1991. A Three-Level Model for Plan Exploration. In ACL91, Proceedings of the 29th Meeting of the Association for Computational Linguistics, pages 39--46.

....independently of the refashioning plan. However, with this approach, the hearer would not be able to recognize the new referring expression plan by recognizing the refashioning plan, and hence we would need a way to account for how this would be done. See Lambert and Carberry (1991) and Ramshaw (1991) for a division between discourse plans and problem solving plans. 56 versational participants, and so is available to be judged and refashioned in the same was as the initial referring expression is. So, as Clark and Wilkes Gibbs propose, the same process for judging and refashioning the ....

Ramshaw, L. A. (1991). A three-level model for plan exploration. In Proceedings of the 29 th Annual Meeting of the Association for Computational Linguistics, pages 39--46.

....On the other hand, the utterances may 2 The majority of the examples in this paper are taken from (Gross et al. 1993) user165.tex; 8 06 1998; 11:38; no v. p. 6 Tracking Initiative in Collaborative Dialogue Interactions 7 propose problem solving actions (Allen, 1991; Lambert and Carberry, 1991; Ramshaw, 1991) that contribute not directly to the agents domain goal, but to how they will go about constructing a plan to achieve this goal, such as Let s look at the first [problem] first. I think they are separate. An agent is said to have the dialogue initiative if she takes the conversational lead in ....

Ramshaw, L. A.: 1991, `A Three-Level Model for Plan Exploration'. In: Proceedings of the 29th Annual Meeting of the Association for Computational Linguistics. pp. 36--46.

....the different levels (at least for task oriented dialogs) with links between trees. With this model, it is possible to recognize discourse intentions that span several utterances, and to allow processes such as focusing to operate independently at all three levels. In concurrent work, Ramshaw [55] hypothesized three slightly different levels: domain execution, exploration, and discourse. The domain execution level is the set of plans and goals that the agent has adopted. The objects of planning at the exploration level are domain plans or nested exploration level plans; example strategies ....

L. A. Ramshaw. A three-level model for plan exploration. In Proceedings of the Twenty-Ninth Annual Meeting of the Associaton for Computational Linguistics (ACL-91), pages 18--21, 1991.

....from the definition of communicative actions in terms of team formation. Other intention based approaches to dialogue modeling have discussed and analyzed the importance of considering also the agents metalevel activity for modeling their interaction with other agents (Litman Allen 1987, Ramshaw 1991, Carberry et al. 1992) this metalevel activity is important to understand behaviors like plan exploration. 3 An architecture based on plans In our architecture, the actions which an agent can execute are of three types and, correspondingly, are defined as belonging to three different sets ....

.... a) Picking out the domain or linguistic actions having the desired goal g as an effect ( Find actions ) 18 b) Exploring these actions in order to obtain a set of (possibly partial) plans for g ( Explore actions ) In the exploration task, which recalls the exploration strategies described in Ramshaw (1991) and Carberry et al. 1992) the feasibility of the actions is checked and they are partially expanded, if they are complex; moreover, the preferred plans are selected. 2. Cycle do action describes the execution of the selected plans, until either the de18 Some of the actions appearing as ....

Ramshaw, L.A. (1991). A three-level model for plan exploration. In Proc. 29th Annual Meeting of ACL, pages 39--46, Berkeley,CA.

....comments made at the domain level and comments made at the discourse level. Further, as the first example question above shows, the two levels may be intermixed in a single question. These referential problems are beyond the capabilities of current natural language understanding systems (but see [16, 11] for research aimed at solving this problem in task oriented dialogues) However, if students are allowed to point to the portion of the previous explanation they wish to ask about, these difficult referential problems can be avoided. Moreover, in a study comparing direct manipulation to natural ....

Ramshaw, L. A Three-Level Model for Plan Exploration. In Proceedings of the Twenty-Ninth Annual Meeting of the Association for Computational Linguistics, (Berkeley, CA, June), 1991, pp. 39--46.

.... those recipes describe the actions that an agent focuses on during his activity, they cannot represent all the actions which may be observed: in fact, agents carry on several activities, like evaluating different recipes for obtaining his goals, checking the success of their actions, and so forth; [19, 9] used metalevel plans to represent and recognize problem solving activity and distinguish dialog behavior deriving from different attitudes towards actions (e.g. commitment vs. exploration) moreover, 11] introduced an explicit, plan based representation of intentions, in order to model ....

L.A. Ramshaw. A three-level model for plan exploration. In Proc. 29th Annual Meeting of ACL, pages 39--46, Berkeley,CA, 1991.

....the different levels (at least for task oriented dialogs) with links between trees. With this model, it is possible to recognize discourse intentions that span several utterances, and to allow processes such as focusing to operate independently at all three levels. In concurrent work, Ramshaw [84] hypothesized three slightly different levels: domain execution, exploration, and discourse. The domain execution level is the set of plans and goals that the agent has adopted. The objects of planning at the exploration level are domain plans or nested exploration level plans; example strategies ....

L. A. Ramshaw. A three-level model for plan exploration. In Proceedings of the Twenty-Ninth Annual Meeting of the Associaton for Computational Linguistics (ACL-91), pages 18--21, 1991.

....axiomatization of knowledge preconditions. In Section 4, we demonstrate the use of knowledge preconditions in accounting for informationseeking subdialogues, such as those in Figure 1. We then compare our approach to the alternative accounts [ Litman and Allen, 1987; Lambert and Carberry, 1991; Ramshaw, 1991 ] 1) E: First you have to remove the flywheel. 2) A: How do I remove the flywheel (3) E: First, loosen the two allen head setscrews holding it to the shaft, then pull it off. 4) A: OK. 5) I can only find one screw. Where s the other one (6) E: On the hub of the flywheel. Figure 1: ....

....satisfy a has:sat:descr requirement. Unlike Lambert and Carberry s approach, however, and Litman and Allen s as well, our approach actually recognizes this structure. The other approaches are essentially utteranceto utterance based and thus do not recognize discourse segments as separate units. Ramshaw [ 1991 ] has added a different third type of plan, exploration plans, to Litman and Allen s two types. Exploration plans are intended to model the process by which agents explore courses of actions. Although we have not yet incorporated such reasoning into our model, we hypothesize that the exploration ....

L. A. Ramshaw. A three-level model for plan exploration. In Proceedings of the 29th Annual Meeting of the ACL, pages 39--46, Berkeley, CA, 1991.

....and Request Ref. Second, there are domain actions (Litman and Allen, 1987) that specify the domain specific actions that the dialogue participants have chosen to satisfy their domain goal, such as Transfer Money. Finally, there are problem solving actions (Allen, 1991; Lambert and Carberry, 1991; Ramshaw, 1991) which are meta level actions describing how the dialogue participants are going about constructing their domain plan. Examples of problem solving actions include Introduce Plan, Evaluate Plan, and Instantiate Parameter. For a decade, researchers have developed models for discourse analysis that ....

....Instantiate Parameter. For a decade, researchers have developed models for discourse analysis that distinguish between domain and discourse actions (Litman and Allen, 1987) and more recently, models that further distinguish between domain and problem solving actions (Lambert and Carberry, 1991; Ramshaw, 1991). By distinguishing among these three types of actions, these models are able to apply action type19 specific heuristics to the recognition process at each level while maintaining the overall uniform plan inference process across all levels. In addition, by distinguishing between ....

Ramshaw, Lance A. 1991. A three-level model for plan exploration. In Proceedings of the 29th Annual Meeting of the Association for Computational Linguistics, pages 36--46.

....context tree which represents the plan refining and variable instantiating moves available to the user. In the case of ill formed input, heuristics are used to hypothesize how the input might relate to the user s inferred plan, thus allowing the system to interpret the input. In later work [ Ramshaw, 1991 ] Ramshaw modified his model to include a domain execution level, an exploration level, and a discourse level to represent commitment to a plan versus exploration of a plan. While Ramshaw s work represents the first substantive attempt to model the problem solving process, it has several ....

Lance A. Ramshaw. A Three-Level Model for Plan Exploration. In Proceedings of the 29th ACL, pages 36--46, Berkeley, California, 1991.

....are recognized on the grounds of linguistic information provided by the syntactic semantic analyzer [ Ahrenberg, 1988 ] If ordering is allowed it is important to know which task is currently being performed, exploring the database or ordering. This problem has been discussed by, for instance, Ramshaw [ 1991 ] and Lambert and Carberry [ 1991 ] They present models using three different, but interacting, levels of plans to know when users stop exploring different plans and instead commit themselves to one plan. However, a result emerging from the analysis of our dialogues [ Jonsson, 1993a ] is that ....

Lance A. Ramshaw. A three-level model for plan exploration. In Proceedings of the 29th Annual Meeting of the ACL, Berkeley, pages 39--46, 1991.

....specifies how domain and linguistic actions have been connected to each other. Introduction This paper aims at clarifying the role of problemsolving plans in modeling the process of plan and goal recognition of a speaker. A first step toward the introduction of problem solving plans was made by (Ramshaw 1991) with the main goal of characterizing the difference between the activity of plan development and plan execution. The idea is that in some cases an utterance does not aim at finding out the best way to reach a goal (exploration) but at pursuing an already adopted plan (execution) So, he ....

....be a domain subaction or can be a discourse action. Of course, nothing says a priori that when we engage in a dialogue we are carrying out an exploration activity: if I say Where is the library I am probably committed to a plan involving going to the library. This point is set out clearly in (Ramshaw 1991). But he has to introduce an explicit marker to keep apart exploration linguistic actions from execution linguistic actions. It seems that the problem solving level is able to explain the difference without adding extra information: in the first case it specifies that a given action (asking ....

Ramshaw, L. 1991. A three-level model for plan exploration.

.... In contrast, a number of researchers coming from linguistics, multi agent systems, and mixinitiative planning argue for a much richer view of the planning process, which I will call the intentional approach (Bratman et al. 1988; Cohen and Levesque, 1990; Pollack 1992; Grosz and Kraus, 1996; Ramshaw. 1991; Traum and Allen, 1994) Essentially, the intentional approach adds a layer atop standard plan representations that allows a system to reason about properties of different plans. For example, a having a plan doesn t change an agent s behavior unless it intends to perform the actions in the plan ....

....However, it is unclear how to relate the intentional approach to more traditional planning methods. Intentional approaches are typically described as normative formalisms rather than implemented systems (Grosz and Kraus, 1996) or when implemented, do not utilize conventional planning algorithms (Ramshaw, 1991). This article describes a research program that aims to bridge the gap between these two formalisms. Specifically, I make the following four claims: 1) Planning agents operating in complex multi agent environment must be able to reason about multiple plans and allow different plans to have ....

[Article contains additional citation context not shown here]

Ramshaw, L. A., 1991. "A three-level model for plan exploration," in Proceedings of ACL91.

....Carberry shows how it can be used to handle some complicated discourse phenomena such as ill formed queries and inter sentential ellipsis. In [Lambert and Carberry, 1991] this is extended to a three layer model involving plans at the domain level, the problem solving level, and the discourselevel. Ramshaw [1991] proposes a similar three layer model. My work on plans as arguments has so far concentrated on only the domain level, although the relationship between plans and discourse actions has been explored in the context of the TRAINS project described in the next chapter (see also [Traum and Allen, ....

Lance A. Ramshaw, "A three-level model for plan exploration," in Proceedings of the Twenty-Ninth Annual Meeting of the Association for Computational Linguistics, Berkeley, CA, 18--21 June 1991. University of California.

....the incremental recognition of communicative goals that cannot be recognized from a single utterance alone; 3. it differentiates between illocutionary effects and desired perlocutionary effects. Other work based on distinguishing between different types of plans in understanding discourse is by Ramshaw [1989a; 1989b; 1991]. 3.2.1.2 Pollack s model of plans as mental phenomena In her influential work [1986; 1990] Pollack challenges the traditional approach to plan inference in discourse understanding, which amounts to directly reasoning about the object of an actor s plan, and to constructing a recipe for action ....

Lance A. Ramshaw. A Three-Level Model for Plan Exploration. In ACL91, Proceedings of the 29th Meeting of the Association for Computational Linguistics, pages 39--46, 1991.

No context found.

L.A. Ramshaw. A three-level model for plan exploration. In Proceedings of ACL-91, pages 39--46, Berkeley, CA, June 1991.

No context found.

Lance A. Ramshaw, \A Three-Level Model for Plan Exploration," In Proceedings of the 29th ACL, pages 39-46, Berkeley, CA, June 1991.

No context found.

Ram91a. Ramshaw, L.A. A Three-Level Model for Plan Exploration. In Proc. 29th Annual Meeting of the ACL, 1991, pp. 39-46.

First 50 documents

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