Allen, J. (1983). Recognizing intentions from natural language utterances. In Brady, M., & Berwick, R. C. (Eds.), Computational Models of Discourse, pp. 107--166. MIT Press, Cambridge, MA.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....a role is missing from one side or the other, or cases in which input or task model has more than one value are also allowed) The core speech acts that are currently modelled include assert, info request, order, request and suggest. Unlike many accounts of the effects of these speech acts (e.g. [8, 1, 7, 13]) there are no direct effects on the beliefs, desires or intentions of the conversational participants. This allows for the possibility that participants are insincere in their utterances. Following [34] the direct effects involve social commitments, and one may then infer from these commitments ....

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. C. Berwick, editors, Computational Models of Discourse. MIT Press, 1983.

....[CPA82] that, in question answering systems, users expected the system to recognize their unstated goals in order to provide more helpful responses to questions. Cohen [Coh78, CP79] concentrated on using plan synthesis together with speech acts for natural language generation. Allen [All79, AP80, All83] on the other hand, used plan recognition of speech acts for natural language understanding. We will concentrate here only on Allen s work. Allen studied transcripts of actual interactions at an information booth in a Toronto train station. A typical exchange was something like this [All83] ....

....AP80, All83] on the other hand, used plan recognition of speech acts for natural language understanding. We will concentrate here only on Allen s work. Allen studied transcripts of actual interactions at an information booth in a Toronto train station. A typical exchange was something like this [All83] patron: When does the Montreal train leave clerk: 3:15 at gate 7. Note that, although the patron only requested the departure time, the clerk also volunteered information about the departure gate as well. Presumably, the clerk recognized the plan of the patron (to board the train) and ....

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. C. Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, 1983.

....user s competence in the domain. INTRODUCTION This paper presents a plan based consulUttion system for getting information on how to achieve a goal in a restricted domain. 1 The main purpo of the system is to recognize the user s plans and goals to build cooperative answers in a flexible way [Allen, 83] Carberry, 90] The system is composed of two parts: hypotheses construction and response generation. The construction of hypotheses is based on Context Models (CMs) Carberry, 90] Carberry uses default inferences [Carberry, 90b] to select a single hypothesis for building tile final answer of ....

J.F.Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick editors, Computational Models of Discourse. 107-166. MIT Press, 1983.

....of particular features. For example, Smith and Guinn provide a richer model of mixed initiative than that provided here [Guinn, 1994; Smith et al. 1992] and there are precise models of how hearers infer the utterance act type or the inten tion underlying a particular communicative action [Allen, 1983; Sidner, 1985; Litman and Allen, 1990; Traum, 1994] However, to my knowledge no previous work has included a specification of the agent architecture, the relationship of the architecture to language behavior, the role of resource limits, and the plan evaluation process. The remainder of this ....

James F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational Models of Discourse. MIT Press, 1983.

....one of those others is being invoked. Thus, expectations are one of the primary mechanisms needed for tracking the conversation as it jumps from subdialog to subdialog. This is known elsewhere as the plan recognition prob lem and it has received much attention in re cent years. See, for example, [Allen, 1983], Litman and Allen, 1987] Pollack, 1986] and [Carberry, 1990] Systems capable of all of the above behaviors are rare as has been observed by [Allen el al. 1989] no one knows how to fit all of the pieces together. One of the contributions of the current work is to present an architecture ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational Models of Discourse, pages 107-166. MIT Press, Cambridge, Mass., 1983.

....(PI) in models of conversation has been widely noted in the computational linguistics literature. Incorporating PI capabilities into systems that answer users questions has enabled such systems to handle indirect speech acts [13] supply more information than is actually requested in a query [2], provide helpful information in response to a yes no query answered in the negative [2] disambiguate requests [17] resolve certain forms of intersentential ellipsis [6,11] and handle such discourse phenomena as clarification subdialogues [11] and correction or debugging subdialogues The ....

....literature. Incorporating PI capabilities into systems that answer users questions has enabled such systems to handle indirect speech acts [13] supply more information than is actually requested in a query [2] provide helpful information in response to a yes no query answered in the negative [2], disambiguate requests [17] resolve certain forms of intersentential ellipsis [6,11] and handle such discourse phenomena as clarification subdialogues [11] and correction or debugging subdialogues The research reported in this paper has been made possible in part by an IBM Graduate ....

James F. Allen. Recognizing intentions from natural language utterances. In Michael Brady and Robert C. Berwlck, editors, Computational Models of Discourse, pages 107-166, MIT Press, Cambridge, Mass., 1983.

....Interpretation and repair attempt to retrace this selection process abductively when a hearer attempts to interpret an observed utterance, he tries to identify the goals, expectations, or misunderstandings that might have led the to produce it. Previous plan based ap proaches [Allen, 1979; Allen, 1983; Litman, 1985; Carberry, 1985] have had difficulty constraining this inference from only a germ of content, potentially a tremendous number of goals could be inferred. A key assumption of our approach, which follows from in sights provided by Conversation Analysis [Garfinkel, 1967; ....

James F. Allen. Recognizing inten- tions from natural language utterances. In Michael Brady, Robert C. Berwick, and James F. Allen, editors, Computational Models of Discourse, pages 107-166. The MIT Press, 1983.

....[CPA82] that, in question answering systems, users expected the system to recognize their unstated goals in order to provide more helpful responses to questions. Cohen [Coh78, CP79] concentrated on using plan synthesis together with speech acts for natural language generation. Allen [All79, AP80, All83] on the other hand, used plan recognition of speech acts for natural language understanding. We will concentrate here only on Allen s work. Allen studied transcripts of actual interactions at an information booth in a Toronto train station. A typical exchange was something like this [All83] ....

....AP80, All83] on the other hand, used plan recognition of speech acts for natural language understanding. We will concentrate here only on Allen s work. Allen studied transcripts of actual interactions at an information booth in a Toronto train station. A typical exchange was something like this [All83] patron: When does the Montreal train leave clerk: 3:15 at gate 7. Note that, although the patron only requested the departure time, the clerk also volunteered information about the departure gate as well. Presumably, the clerk recognized the plan of the patron (to board the train) and ....

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. C. Berwick, editors, Computational Models of Discourse, pages 107-166. MIT Press, 1983.

....by means of the implementation of a dedicated agent architecture it has been shown that the defined notions and criteria are operational and provide a basis to develop applications of agents that monitor and interpret the behaviour of other agents. In research on plan recognition, such as [Allen, 1983; Konolige and Pollack, 1989] based on observed actions of an actor agent the observing agent ascribes intentions and plans to the actor that are probable. Plan recognition is performed using data on the actions from a single, ongoing interaction of the agent, and uses domain knowledge on actions ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, eds., Computational Models of Discourse. MIT Press, Cambridge, Ma., 1983.

....to recognize their unstated goals in order to provide more helpful responses to questions ( Cohen et al. 1982] Cohen ( Cohen, 1978; Cohen and Perrault, 1979] concentrated on using plan synthesis together with speech acts for content planning. Allen ( Allen, 1979; Allen and Perrault, 1980; Allen, 1983] on the other hand, used plan recognition and speech act theory for intention recognition. We concentrate here only on Allen s work. Allen studied transcripts of actual interactions at an information booth in a Toronto train station. A typical exchange was something like this ( Allen, 1983] ....

....1980; Allen, 1983] on the other hand, used plan recognition and speech act theory for intention recognition. We concentrate here only on Allen s work. Allen studied transcripts of actual interactions at an information booth in a Toronto train station. A typical exchange was something like this ([Allen, 1983]) Patron: When does the Montreal train leave Clerk: 3:15 at gate 7. Note that although the patron only requested the departure time, the clerk also volunteered information about the departure gate as well. Presumably, the clerk recognized the plan of the patron (to board the train) and ....

James Allen, "Recognizing Intentions from Natural Language Utterances," In M. Brady and R. C. Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, 1983.

....of a speaker s planning process, an addressee will be able to deduce not only a speaker s immediate goal in talking, but the higher level goal for which this goal was adopted, and other subgoals which may be required for the same objective. Thus, for instance, in the model described by Allen [6], the station employee, when asked by the anxious passenger at what time her train will depart, replies with the time and platform number. The employee supplies the additional information as a result of being helpfully disposed to passengers and consequently forming the goal of communicating in ....

....of which is to find out the train s departure time, this plan does not explicitly include the station employee forming the intention to help her. The driving force behind an agent s reasoning to determine a speaker s higher level goals can be described as part of being helpfully disposed, as in [6], or as part of general checking for contradictions. It would be infelicitous not to mention it if, for example, one knew that the train had been diverted, and a system that can model the generation of a useful response in such a case is described by Pollack in [101] Helpful responses must always ....

[Article contains additional citation context not shown here]

James F. Allen. Recognizing intentions from natural language utterances. In Brady and Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, Cambridge, MA, 1983.

.... elaboration of each speech act type in other tagging schemes [24] It is tempting to also consider this dimension as a means of inferring discourse structure on the basis of utterance level labels, since it is widely believed that models of task structure drive the behavior of dialogue systems [23, 3, 22], and the relationship between discourse structure and task structure has been a core topic of research since Grosz s thesis [15] However, we leave the inference of discourse structure as a topic for future work because the multifunctionality of many utterances suggests that the correspondence ....

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. Berwick, editors, Computational Models of Discourse. MIT Press, 1983.

....1990, Pelavin 1990] nearly all work on the inverse problem of plan recognition has focused on specific kinds of recognition in specific domains. This includes work on story understanding [Bruce 1981, Schank 1975, Wilensky 1983] psychological modelling [Schmidt 1978] natural language pragmatics [Allen 1983, Carberry 1983, Litman 1987, Grosz Sidner 1987] and intelligent computer system interfaces [Genesereth 1979, Huff Lesser 1982, Goodman Litman 1990] In each case, the recognizer is given an impoverished and fragmented description of the actions performed by one or more agents and expected ....

....describe it as the result of applying unsound rules of inference that are created by reversing normally sound implications. From the fact that a particular plan entails a particular action, one derives the unsound rule that that action may Kautz Plan Recognition 09 09 97 page 3 imply that plan [Allen 1983]. Such unsound rules, however, generate staggering numbers of possible plans. The key problems of deciding which rules to apply and when to stop applying the rules remain outside the formal theory. By contrast, the framework presented in this chapter specifies what conclusions are absolutely ....

[Article contains additional citation context not shown here]

Allen, James (1983) Recognizing Intentions From Natural Language Utterances, in Computational Models of Discourse, eds. Michael Brady & Robert Berwick, The MIT Press, Cambridge.

....by means of the implementation of a dedicated agent architecture it has been shown that the defined notions and criteria are operational and provide a basis to develop applications of agents that monitor and interpret the behaviour of other agents. In research on plan recognition, such as [Allen, 1983; Konolige and Pollack, 1989] based on observed actions of an actor agent the observing agent ascribes intentions and plans to the actor that are probable. Plan recognition is performed using data on the actions from a single, ongoing interaction of the agent, and uses domain knowledge on actions ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, eds., Computational Models of Discourse. MIT Press, Cambridge, Ma., 1983.

....and thus avoid redundancy and incomprehensibility in its answers and explanations; detect wrong beliefs of the user and inform the user about them. As an example, let us consider the following question answer pair: 1a) User: When does the Montreal train leave (b) System: 3:15 at gate 7. (Allen 1983) From the user s question in (1) the system (in this example, a system specifically designed for train information) can conclude (a) that the user has the goal (or better, already has the plan) to take the next train to Montreal; and (b) that s he does not know the departure time of this train. ....

....to him her and thereby revise its initial assumptions. Rich (1989) discusses techniques for resolving contradictions between stereotypes and direct observation, and for automatic stereotype modification. The work of Allen, Cohen and Perrault (e.g. Cohen Perrault 1979, Allen Perrault 1980, Allen 1983) concentrates on inferring probable user plans from a formal description of a user s dialog contribution (such as, in English paraphrase, the description Who is x such that p(x) On the basis of these assumptions their system tries to detect obstacles in the executability of the hypothesized ....

Allen, J. F. (1983): Recognizing Intentions from Natural Language Utterances. In: M. Brady and R. C. Berwick, eds.: Computational Models of Discourse. Cambridge, MA: MIT Press.

....B is an assertion related to another subgoal on the theorem proving tree as shown in Figure 1. The user may initiate such a change in subdialogue in an attempt to pursue another path to the global goal. Here the machine first must track the user s intention (in a process called plan recognition [15, 16, 17, 18, 19]) and then evaluate whether to follow the move or not. This decision is based upon the current level of the initiative of the system as described below. If the system follows the user s initiative, it will apply its internal theorem proving system to the subgoal E and pursue voice interactions ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, Cambridge, Mass., 1983.

....plan generation requires an ability to solve certain aspects of the planning problem automatically, while deferring other functions to the user. Complex query response dialogue abilities allow the planner to interact with the user to clearly establish intent and mutual understanding (Allen 1983). Informative feedback from the planner to the human is essential to the mixed initiative planning process. This suggests that results and approaches from mixed initiative planning may map directly to our problems of AMP CSP collaborative planning. And, while originally conceived of as useful for ....

Allen, J. 1983. Recognizing intentions from natural language utterances. In Brady, J. M., and Berwick, R. C., eds., Computational Models of Discourse. Cambridge, Mass.: M. I. T. Press.

....the importance of plans as mental attitudes; after all, inferring another agent s plan means figuring out what actions he has in mind. Consider Allen s model, which was one of the earliest accounts of plan inference in conversation and inspired much of the subsequent work in the field [1, 2]. Allen writes AW(P) where P is some proposition, to express the fact that A has a plan to achieve P; AW(ACT) where ACT names some action, expresses the fact that A has a plan to perform ACT. A typical plan inference rule, then, is expressed as SBAW(P) SBAW(ACT) if P is a precondition of ....

James F. Allen. Recognizing intentions from natural language utterances. In Michael Brady and Robert C. Berwick, editors, Computational 50 Models of Discourse, pages 107--166, MIT Press, Cambridge, Ma., 1983.

....These representation and reasoning problems are among the hardest in AI and cognitive science. 2.1 From Plan Recognition to Collaborative Planning In the late 1970s, James Allen, Philip Cohen, and C. Raymond Perrault were pioneers in developing models of language based on AI planning theory [12, 13, 14, 15]. They drew on work in the philosophy of language [1, 2, 3, 4] that conceptualized language as goal directed action and examined the intentions and beliefs necessary for successful communication. Speech acts, such as informing, requesting, asserting, and promising, are produced through a planning ....

....and intentions of other agents from observations of their actions, including speech acts. People tend to be helpful; in particular, when they recognize obstacles in another agent s plans, they will take action (e.g. provide information) to help overcome the obstacles. A simple exchange from Allen[12] illustrates these points. patron: When does the Montreal train leave clerk: 3:15 at gate 7 4 The patron wants to get on the Montreal train and constructs a plan to achieve this goal. However, he does not know the departure time. One way of finding out a piece of information is to have ....

Allen, J F `Recognizing Intentions from Natural Language Utterances' in Brady, M and Berwick, R C (Eds.) Computational Models of Discourse MIT Press, USA (1983) pp 107-166

....predicts the user using the resources by referring previously learned behavior. Either cooperative or self interested agents try to improve their ability to recognize the likely actions of other agents in multiagent systems. The recognition of other agents intentions is an important task [1][2] 3] It is particularly so when an agent is expected to produce some useful information in an interactive computing environment, where acquiring knowledge of the current world serves as a basis for immediate or future actions of the system. Examples of such systems include intelligent help ....

J. Allen. Recognizing intentions from natural language utterances, In Computational Models of Discourse M. Brady and R. Berwick eds, The MIT Press. 1983.

....to progress. 1. Dialogue analysis based on models of individual agents beliefs and knowledge structures, usually presented within an AI Derived theory of plans, inference and possibly speech dialogue acts and truth maintenance, many using a space metaphor to represent individuals (e.g. Allen [8], Traum [9] Kobsa [10] Ballim and Wilks [11] 2. AI derived models of dialogue based on more linguistic notions and not primarily based on models of individuals; the representation is often in terms of partitioned semantic nets to represent domains but uses concepts of focus, failure, repair ....

J Allen. Recognizing Intentions From Natural Language Utterances. In: Computational Models of Discourse Chapter 2 (M Brady and R Berwick eds.), Cambridge, MA: MIT Press, 1982.

.... for generation [ Moore and Paris, 1989 ] in top down prediction of utterance function and structure [ Alexandersson et al. 1994 ] and most importantly, to provide a representation of natural language utterances that is uniform with that used in general facilities for planning and action [ Allen, 1983 ] We follow [ Traum and Hinkelman, 1992 ] in regarding speech acts as fully joint actions between conversational participants. Not only are joint speech acts co operatively undertaken, but they have at least nominally joint effects: if they complete but still fail to result in shared goals or ....

....inference allows an individual to be represented as a member of several different groups holding conflicting beliefs, and inheriting only those beliefs consistent with those represented as being held privately. The results are thus substantially different from those obtained in classical logics [ Allen, 1983; Kraus and Lehmann, 1988; Appelt, 1985; Cohen and Levesque, 1990 ] They differ from other path based algorithms [ Ballim and Wilks, 1991 ] in the provision of semantic relevance conditions and in addressing the need for shared attitudes by ascribing them directly to groups, rather than by ....

James Allen. Recognizing intentions from natural language utterances. In Michael Brady and Robert C. Berwick, editors, Computational Models of Discourse. MIT Press, 1983.

....to reason further from that entire set; instead it should ascribe only the single, less specific belief that can be strongly supported by the evidence. In fact, this is the insight inherent in several strategies that have been proposed to control plan recognition, such as Allen s forking heuristic [All83] and Sidner s single branch strategy [Sid85] The single branch strategy, for instance, applies when a plan recognition system reaches a point at which it has assumed that the actor intends to perform some action #,and 3 We naturally assume that the definition of an assumption set incorporates ....

....that if an agent believes that p is an e#ect of performing #, and he intends to do #, then he both intends to do # in order to make p true (that is, he intends the TO fragment) and plans to achieve p. Such rules correspond closely to the classical plan recognition rules in a system such as Allen s [All83]. As noted earlier, the second important set of rules in the system are the defeat rules, which express the relative strength of arguments. One example is Purposeful Action Defeat. This rule encodes the presumption that agents engage in purposeful actions: they do not typically intend actions ....

J. F. Allen (1983): Recognizing intentions from natural language utterances. In: M. Brady and R. C. Berwick, eds.: Computa32 tional Models of Discourse. Cambridge, MA: MIT Press.

....in the domain in which the interaction takes place#; various techniques have been developed for plan and intention recognition, i.e. to understand the agent s goals whichhave led him to act. This kind of reasoning is particularly useful in case of communication: as it has been noticed since #Allen, 1983#, when a user asks a question to a dialog system, he does not expect just a literal answer, but he wants also that his domain goals are taken into account: for example, from a question concerning the location of a library it is possible to infer that #probably# the user wants to go to the library ....

Allen, J. #1983#. Recognizing intentions from natural language utterances. In Brady, M. and Berwick, R., editors, Computational models of discourse, pages 107#166. MIT Press.

....in the domain in which the interaction takes place) various techniques have been developed for plan and intention recognition, i.e. to understand the agent s goals which have led him to act. This kind of reasoning is particularly useful in case of communication: as it has been noticed since [Allen, 1983], when a user asks a question to a dialog system, he does not expect just a literal answer, but he wants also that his domain goals are taken into account: for example, from a question concerning the location of a library it is possible to infer that (probably) the user wants to go to the library ....

Allen, J. (1983). Recognizing intentions from natural language utterances. In Brady, M. and Berwick, R., editors, Computational models of discourse, pages 107--166. MIT Press.

....the mistake. We model the treatment of misunderstandings as a goal directed, rational behavior: our notion of coherence in communication is based on the idea that the goals identified when interpreting a turn have to be related with the previously expressed or inferred goals of the interactants [2, 8]. When an agent receives an incoherent turn (or a turn which only partially matches his expectations) he adopts the further goal of realigning the interactants diverging interpretations. To do so, he reasons to understand which agent misunderstood the other one; then, he can restructure his own ....

....Goal adoption: g is one of the goals that B has inferred A is going to aim at; e.g. in: T1: A: I need a book. Where is the library T2: B: It is over there, but it is closed. B provides A with an extra helpful information which satisfies his next goal of checking whether the library is open [2]. 3. Plan continuation: g contributes to a plan that B is carrying on. 2 E.g: T1 B: Where is the library T2 A: It s over there. T3 B: Is it open today Goal adherence and adoption refer to a complete satisfaction of the hearer s pending intentions. So, when the partner satisfies only ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational models of discourse, pages 107--166. MIT Press, 1983.

....i.e. Would you be so kind as to open the window ) 1 and to produce the same request in appropriate circumstances. Note that also the plan formation activity must be properly recognized by a hearer, as is shown by the classical case of applicability condition check ( Is the library open , Allen (1983)) Therefore, also knowledge about how to build plans is necessary for both the phases of understanding and producing behavior. A consequence of the choice of assimilating dialogue as general cooperative interaction is that we model dialogue avoiding specific knowledge about the admissible ....

....goal, belonging to the speaker or to his partner. If the goal belongs to the partner, the speaker may have identified it because it was communicated explicitly by the partner (by means of an illocutionary act) or may have inferred it although it was not stated overtly (Allen Perrault 1980, Allen 1983)) Following the notion of cooperation in terms of goal delegation and goal adoption defined in Castelfranchi Parisi (1980) and Castelfranchi Falcone (1997) we consider an utterance coherent with the previous context if its receiver A can interpret it as the means of the speaker B to achieve ....

[Article contains additional citation context not shown here]

Allen, J.F. (1983). Recognizing intentions from natural language utterances. In Brady, M. & Berwick, R.C., Eds. Computational models of discourse, pages 107--166. MIT Press.

....like the domain level plans are. On the other hand, intention recognition requires an explicit, declarative representation of this hard wired activity of agents. In the dialog modeling literature, plan recognition was considered necessary to improve the interaction with the user since the work of [1]. Initially, the speakers activity has been represented by means of plan recipes: those recipes described the sequence of domain and linguistic actions to be performed in order to achieve a goal. However, it was soon clear that, while those recipes describe the actions that an agent focuses on ....

....plan; if it is so, the process expands the interpretation structure by adding to it an instance of the higher level hypothesis which explains why the observed action has been performed. 9 The plan recognition process is inspired to the heuristic plan recognition techniques described in [9, 1]; the use of AM plans makes explicit the rationale behind these traditional heuristics, by representing the reasons underlying the execution of object level actions. The object level plans represent the activities typically performed in the domain: being pre compiled recipes, they are useful to ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational models of discourse, pages 107--166. MIT Press, 1983.

....dialog phenomena like the extra helpful behavior, the displays of behavioral and conversational cooperation among agents, the presequences, subdialogs, notifications and acknowledgments for grounding purposes. Overanswering is the phenomenon modeled most frequently in plan based dialogue systems [2]. In our model, it represents the simplest example of goal adoption: 4) A: Could you give me a pass for the CS labs B: Sorry, today the labs are closed. B has identified A s plan to work in the labs and he adopts the goal that A knows that they are closed (in fact, this is a constraint in A s ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational models of discourse, pages 107--166. MIT Press, 1983.

....where a hearer can deduce the speaker s plan and provide help even when it was not requested. For instance, providing the next step in a deduced plan, or detecting obstacles in the speaker s plan that may be unknown to the speaker, and providing help. This work was later expanded and improved in [3], where the speech act operators in Figure 2 were proposed. While the Allen and Perrault model nicely handled the role of plan and intention recognition in recognizing indirect speech acts, it was not particularly helpful in relating the role of various aspects of the surface form. In fact, some ....

....p. 202] He defines mutual knowledge between two agents A and S of a proposition p, K SA p as [86, p. 30] KS p KA p KSKA p KAKS p KSKAKS p KAKSKA p Delta Delta Delta It is thus an infinite conjunction of nested beliefs. This approach has since been adopted by many others, including Allen [3] and Perrault, who provides an elegant default logic theory of how to obtain each of these beliefs given prior knowledge and a conversational setting [77] Barwise credits Harman with the fixed point approach. Harman formulates mutual knowledge as knowledge of a self referential fact: A group of ....

James [F.] Allen. Recognizing intentions from natural language utterances. In Michael Brady and Robert C. Berwick, editors, Computational Models of Discourse. MIT Press, 1983.

.... is necessary in the plan recognition domain for the plan recognition process to begin (Grosz, Pollack, and Sidner, 1989) For example, Allen s Action Effect rule derives an agent H s belief that another agent S intended action E from H s belief that S intended action A, where E is an effect of A (Allen, 1983). This rule will apply in a context in which H observes S perform some action A only if the system assumes that S did A intentionally 2 . The Declarative rule, therefore, provides the background context in which the reasoning process will take place. From this rule, we derive that speakers ....

....standard sense, i.e. independent of mental states. For example, one approach that has been adopted to start the reasoning process is to assume a restricted set of possible goals. In Allen s model, the domain is a train station and the only possible goals are boarding a train and meeting a train (Allen, 1983). Hence, if a speaker asks Do you know when the Windsor train leaves , the system may assume that the speaker wants to board this train, and therefore answer appropriately. Another approach is to assume that speakers start by stating their goal. Pollack s model, for example, reasons on the basis ....

Allen, James F. 1983. Recognizing intentions from natural language utterances. In Michael Brady and Robert C. Berwick, editors, Computational Models of Discourse. MIT Press, Cambridge, MA, pages 107--166.

....one of those others is being invoked. Thus, expectations are one of the primary mechanisms needed for tracking the conversation as it jumps from subdialog to subdialog. This is known elsewhere as the plan recognition problem and it has received much attention in recent years. See, for example, Allen, 1983 ] Litman and Allen, 1987 ] Pollack, 1986 ] and [ Carberry, 1990 ] Systems capable of all of the above behaviors are rare as has been observed by [ Allen et al. 1989 ] no one knows how to fit all of the pieces together. One of the contributions of the current work is to present an ....

J.F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, Cambridge, Mass., 1983.

....obviously, plan recognition is the process of inferring an agent s plan from observation of the agent s actions. The agent s actions can be physical actions or speech actions. Four principal methods for plan recognition have been proposed in the literature. The methods are plausible inference [ Allen, 1983; Carberry, 1988; Litman and Allen, 1987; Sidner, 1985 ] parsing [ Vilain, 1990 ] circumscribing a hierarchical representation of plans and using deduction [ Kautz, 1987 ] and abduction [ Charniak and McDermott, 1985; Konolige and Pollack, 1989; Poole, 1989 ] Our particular interest is in ....

....interest is in the use of plan recognition in question answering systems, where recognizing the plan underlying a user s queries aids in the generation of an appropriate response. Here, the plan of the user, once recognized, has been used to: supply more information than is explicitly requested [ Allen, 1983; Luria, 1987 ] handle pragmatically ill formed queries [ Carberry, 1988 ] provide an explanation from the appropriate perspective [ McKeown et al. 1985 ] respond to queries that result from an invalid plan [ Pollack, 1984; Pollack, 1986 ] and avoid misleading responses and produce ....

[Article contains additional citation context not shown here]

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. C. Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, 1983.

....may be used to reduce the ambiguity on the user s intentions. Introduction When an automated system plays the role of a consultant in information seeking dialogues, the extent to which it provides cooperative responses is directly related to its ability to understand the user s plans and goals (Allen 1983), Carberry 1990b) In such kind of dialogues, the determination of the user s intentions is generally quite difficult because not always the user states them explicitly. More often, s he only asks general questions, or questions on particular aspects of the tasks. So, in order to reduce the ....

Allen, J. (1983). Recognizing intentions from natural language utterances. In Brady, M. and Berwick, R., eds., Computational models of discourse. MIT Press.

....Secondly, cooperating agents use plan recognition to enhance the effectiveness of their communication. Typically an addressee applies plan recognition to the speaker, and then supplements her reactive reply with an initial utterance addressing an unfulfilled goal in the recognized plan, e.g. Allen (1983). Since plan recognition is an imprecise process, and indeed most initial utterances could potentially form part of a number of plans, the beliefs that one agent ascribes to another in the process of plan recognition follow in part from assumptions and must be defeasible. Given the multiple ....

Allen, J. F. (1983). Recognizing intentions from natural language utterances.

.... for the semantics of anaphora (e.g. Kamp and Reyle, 1993] and on the other hand, the models of context proposed for intention recognition and dialogue management, whose emphasis is on capturing the effects of speech acts on the beliefs, intentions, and obligations of the participating agents [Allen, 1983; Carberry, 1990; Cohen and Levesque, 1990; Perrault, 1990; Traum and Allen, 1994] These traditions resulted in very detailed proposals about context and context update; 1 but the resulting models of context differ significantly. It is not possible to simply adopt one or the other model. While ....

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. Berwick, editors, ComputationalModels of Discourse, pages 107--166. MIT Press, Cambridge, MA, 1983.

....in here. b: Do you have a watch on In context, a) may be a request to close the window. Sentence (b) may be asking what time it is or requesting to borrow the watch. The idiom approach allows neither for context nor the reasoning connecting utterance and desired action. The plan based approach [Allen 83, McCafferty 86, Perfault 80, Sidner 81] presumes a mechanism modelling human problem solving abilities, including reasoning about other agents and inferring their intentions. The system has a model of the current situation and the ability to choose a course of action. It can relate uttered ....

....schemas for attributing plans: having observed that an agent wants an effect, believe that they may plan an action with that effect, and so on. For modelling communication, it is necessary to complicate these rules by embedding the antecedent and consequent in one sided mutual belief operators [Allen 83] In the Alien approach, our Spanish example hinges on the acts precondifions: Su7 anne will not attribute a qknUesfion to Mrs. de Prado if she believes she already ows the answer, but this knowledge could be the basis for a request. Sentences like It s cold in.here are also interpreted by ....

Allen, J., "Recognizing Intentions From Natural Language Ut(erances," in Computational Models of Discourse, Brady, M. and Betwick, B. (ed.), MIT Press, Cambridge, MA, 1983, 107-166.

....symptom. The system knows that aspirin can relieve headaches and infers that the user s higher level objective is possibly relieving a headache. However, the system is not sure about this inference, so it utters utterance 6 to verify it. Such plan recognition enables the system to be helpful (cf. Allen, 1983a] Later in the exchange the system, based on this inference, is able to take initiative and suggest that the user take Tylenol for his headache. Intention recognition is vital to dialogue, since it recognizes what the user intended for the system to understand. Even in a question answer system ....

James Allen, "Recognizing Intentions from Natural Language Utterances," In M. Brady and R. C. Berwick, editors, Computational Models of Discourse, pages 107--166. MIT Press, 1983.

No context found.

James F. Allen, "Recognizing Intentions From Natural Language Utterances," in M. Brady and R. Betwick, editors, Computational Models of Discourse. MIT Press, Cambridge, MA, 1983.

No context found.

Allen, J. (1983). Recognizing intentions from natural language utterances. In Brady, M., & Berwick, R. C. (Eds.), Computational Models of Discourse, pp. 107--166. MIT Press, Cambridge, MA.

No context found.

J. Allen, Recognizing intentions from natural language utterances, in: M. Brady, R.C. Berwick (Eds.), Computational Models of Discourse, MIT Press, Cambridge, MA, 1983, pp. 107--166.

No context found.

J. F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. C. Berwick, editors, Computational Models of Discourse. MIT Press, 1983.

No context found.

James F. Allen. Recognizing intentions from natural language utterances. In M. Brady and R. Berwick, editors, Computational Models of Discourse, pages 107--166. The MIT Press, Cambridge, Massachusetts, 1983.

No context found.

James Allen, \Recognizing Intentions from Natural Language Utterances," In M. Brady and R. C. Berwick, editors, Computational Models of Discourse, pages 107-166. MIT Press, 1983.

No context found.

James Allen, "Recognizing Intentions From Natural Language Utterances," In Michael Brady and Robert C. Betwick, editors, Computational Models of Discourse. MIT Press, 1983.

No context found.

James F. Allen. Recognizing intentions from natural language utterances. In ComputationalModels of Discourse [111], pages 107--166.

No context found.

Allen, J. (1983) Recognizing intentions from natural language utterances," in M. Brady, and R. Berwick, eds., Computational Models of Discourse, MIT Press, Cambridge, MA, 107-166.

No context found.

James F. Allen. 1983. Recognizing intentions from natural language utterances. In M. Brady and R.C. Berwick, editors, Computational Models of Discourse. MIT Press.

No context found.

J.F. Allen (1983). Recognizing intentions from natural language utterances. In M. Brady & R.C. Berwick, editors, Computational models of discourse, pages 107--166. MIT Press.

No context found.

Allen, J. (1983) "Recognizing intentions from natural language utterances," in M. Brady and R.

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