A meta-rule approach to flexible plan recognition in dialogue

Although a number of researchers have demonstrated that reasoning on a model of the user's plans and goals is helpful in language understanding and response generation, current models of plan inference cannot handle naturally occurring dialogue. This paper argues that model building from less than ideal dialogues has a great deal in common with processing ill-formed input. It defines well-formedness constraints for information-seeking dialogues and contends that strategies for interpreting ill-formed input can be applied to the problem of modeling the user's plan during an ill-formed dialogue. It presents a meta-rule approach for hypothesizing the cause of dialogue ill-formedness, and describes meta-rules for relaxing the plan inference process and enabling the consideration of alternative hypotheses. The advantages of this approach are that it provides a unified framework for handling both well-formed and ill-formed dialogue, avoids unnatural interpretations when the dialogue is proceeding smoothly, and facilitates a nonmonotonic plan recognition system.Rhonda Eller is a Ph.D. candidate in Computer Science at the University of Delaware. She received her B.S. in Computer Science from Millersville University of Pennsylvania in 1987, and her M.S. degree in the same field from the University of Delaware. Her primary interests lie in the areas of natural language processing, plan recognition, and user modelling. This paper summarizes the current state of her thesis work on repair of an incorrectly inferred plan for a user.Sandra Carberry is an associate professor of computer science at the University of Delaware. Her research interests include discourse understanding, user modeling, planning and plan recognition, and intelligent natural language interfaces. This paper describes work that is part of an ongoing research project to develop a robust model of plan recognition.     

Representing discourse for collaborative interactive generation

The author's work has focused on one mode of collaborative natural language interaction. In this mode, the system is the primary speaker and the user is the primary listener. The Interactive Discourse Planner (IDP) is responsible for planning text to describe and/or justify a domain plan. The user collaborates by providing feedback that lets the system know how to extend the discussion in a way that satisfies the user and thereby achieves the system's goal. As a testbed for the model, IDP discusses driving routes as the domain plans. The primary insight that this work contributes is that it may be useful to identify modes of collaboration. Collaborative interactive modes are defined by the assignment of conversational roles and responsibilities to a computational system and a human. When building a model for an interactive mode, a designer can make assumptions that lead to simpler, computationally tractable systems that provide for efficient and effective interaction.     

Recognizing the agent’s goals incrementally: planning graph as a basis

Plan recognition, the inverse problem of plan synthesis, is important wherever a system is expected to produce a kind of cooperative or competitive behavior. Most plan recognizers, however, suffer the problem of acquisition and hand-coding a larger plan library. This paper is aims to show that modern planning techniques can help build plan recognition systems without suffering such problems. Specifically, we show that the planning graph, which is an important component of the classical planning system Graph-plan, can be used as an implicit, dynamic planning library to represent actions, plans and goals. We also show that modern plan generating technology can be used to find valid plans in this framework. In this sense, this method can be regarded as a bridge that connects these two research fields. Empirical and theoretical results also show that the method is efficient and scalable.     

Recognizing the agent''s goals incrementally:planning graph as a basis

Plan recognition, the inverse problem of plan synthesis, is important wherever a system is expected to produce a kind of cooperative or competitive behavior. Most plan recognizers, however, suffer the problem of acquisition and hand-coding a larger plan library. This paper is aims to show that modern planning techniques can help build plan recognition systems without suffering such problems. Specifically, we show that the planning graph, which is an important component of the classical planning system Graph-plan, can be used as an implicit, dynamic planning library to represent actions, plans and goals. We also show that modern plan generating technology can be used to find valid plans in this framework. In this sense, this method can be regarded as a bridge that connects these two research fields. Empirical and theoretical results also show that the method is efficient and scalable.     

The intelligent help system COMFOHELP

The paper is concerned with the question of whether and under what conditions active help systems with plan recognition components that have been developed in the environment of artificial intelligence research are able to prove their value in the real context of commercial application programs. The question is investigated using the development of the COMFOHELP intelligent help system as an example. COMFOHELP supports the COMFOTEX graphical text processing program and has been developed by the Linguistic Information Science Group at the University of Regensburg since 1988. The system recognizes erroneous and suboptimal plans pursued by the user by analyzing the dialog history and comparing them with the correct plan for achieving the user's goal.Section 2 discusses the research situation and elaborates on those problems which up to now prevented research concepts for plan recognition and intelligent help systems from being practically applied. Testing error situations empirically is a first prerequisite since potential erroneous plans can only be established in real-world tests. The second prerequisite is a special system architecture which counteracts the problem of ambiguities in plan recognition. Section 3 introduces a first still restricted prototype version of COMFOHELP whose efficiency was verified in a statistical hypothesis test. The users performing their text processing tasks with the support of COMFOHELP came off significantly better than members of a reference group working without the intelligent help. Section 4 shows that the proposed COMFOHELP system architecture is reconfirmed by the results of extensive empirical investigations (with more than 100 users) of erroneous plans when using a more complex version of COMFOTEX. The architecture still proves to be worthwhile even when functionality is increased by a factor of three to four.     

语音识别中声学模型研究综述

智能语音技术包含语音识别、自然语言处理、语音合成三个方面的内容,其中语音识别是实现人机交互的关键技术,识别系统通常需要建立声学模型和语言模型。神经网络的兴起使声学模型数量急剧增加,基于神经网络的声学模型与传统识别模型相结合的方式,极大地推动了语音识别的发展。语音识别作为人机交互的前端,具有许多研究方向,文中着重对语音识别任务中的文本识别、说话人识别、情绪识别三个方向的声学模型研究现状进行归纳总结,尽可能对语音识别技术的演化进行细致介绍,为以后的相关研究提供有价值的参考。同时对目前语音识别的主流方法进行概括比较,介绍了端到端的语音识别模型的优势,并对发展趋势进行分析展望,最后提出当前语音识别任务中面临的挑战。     

The SINIX Consultant – Towards a Theoretical Treatment of Plan Recognition

We have realized the help system SINIX Consultant (SC) for SINIX¹ users. The system is capable of answering – in German – natural language questions concerning SINIX commands, objects, and concepts. But not only does this help system react to inquiries – additionally, the system is capable of activating itself. If the user employs a sequence of SINIX commands (a plan) in order to reach a specific goal, the help system proposes a sequence which reaches the same goal, but, with fewer commands. In this paper, a brief survey of the SINIX Consultant and the realized plan recognizer REPLIX is first given. Then, an initial attempt of a theoretical treatment of plan recognition is presented. This is done within the logical framework. We show how we can use an interval-based logic of time to describe actions, atomic plans, non-atomic plans, action execution, and simple plan recognition. We also show that the recognition of inserted sub-plans managed by REPLIX can be handled as well. Then, we present a problem which cannot be treated in the formalism. Thus, in this paper, we don't present a full developed theory but nevertheless, a step towards it. This revised version was published online in August 2006 with corrections to the Cover Date.     

Deterministic planning in the fifth international planning competition: PDDL3 and experimental evaluation of the planners

The international planning competition (IPC) is an important driver for planning research. The general goals of the IPC include pushing the state of the art in planning technology by posing new scientific challenges, encouraging direct comparison of planning systems and techniques, developing and improving a common planning domain definition language, and designing new planning domains and problems for the research community. This paper focuses on the deterministic part of the fifth international planning competition (IPC5), presenting the language and benchmark domains that we developed for the competition, as well as a detailed experimental evaluation of the deterministic planners that entered IPC5, which helps to understand the state of the art in the field.We present an extension of pddl, called pddl3, allowing the user to express strong and soft constraints about the structure of the desired plans, as well as strong and soft problem goals. We discuss the expressive power of the new language focusing on the restricted version that was used in IPC5, for which we give some basic results about its compilability into pddl2. Moreover, we study the relative performance of the IPC5 planners in terms of solved problems, CPU time, and plan quality; we analyse their behaviour with respect to the winners of the previous competition; and we evaluate them in terms of their capability of dealing with soft goals and constraints, and of finding good quality plans in general. Overall, the results indicate significant progress in the field, but they also reveal that some important issues remain open and require further research, such as dealing with strong constraints and computing high quality plans in metric-time domains and domains involving soft goals or constraints.     

A STRENGTHENED ALGORITHM FOR TEMPORAL REASONING ABOUT PLANS

Allen's interval algebra has been shown to be useful for representing plans. We present a strengthened algorithm for temporal reasoning about plans, which improves on straightforward applications of the existing reasoning algorithms for the algebra. This is made possible by viewing plans as both temporal networks and hierarchical structures. The temporal network view allows us to check for inconsistencies as well as propagate the effects of new temporal constraints, whereas the hierarchical view helps us to get the strengthened results by taking into account the dependency relationships between actions.
We further apply our algorithm to the process of plan recognition through the analysis of natural language input. We show that such an application has two useful effects: the temporal relations derived from the natural language input can be used as constraints to reduce the number of candidate plans, and the derived constraints can be made more specific by combining them with the prestored constraints in the plans being recognized.     

A framework for multinational information systems planning

The purpose of this study was to develop and identify formal strategic planning and control approaches and requisites to better manage and integrate the information systems resources (e.g. information systems, telecommunications and office automation) more effectively with the business plans and processes in a large multidivisional, multiproduct and transnational corporate environment. Additionally, an examination of certain external and internal environmental factors and pressures was conducted to assist business and information systems executives, and professionals to focus on those factors which influence the successful management and direction of the multinational information systems resources and functions (MNISR) within their environments.Twenty-five U.S. based multinational organizations participated in the research. Most of the organizations were selected from the manufacturing (e.g. energy, pharmaceuticals, consumer products and electronic) industry with some representation from the banking and financial services industry. In addition, four case studies were conducted to provide in-depth analysis, comparisons and verifications of actual MNISR planning and control practices.A pragmatic strategic planning framework and structure consisting of multiple but interrelated components and activities is suggested as the formal end product of this research. In addition, five general conclusions stand out from the research. First, the MNISR plan structure must accomodate both the similarities and differences between the domestic and international MNISR environments and the key factors which influence them. Second, business and MNISR plan linkages must be established by addressing strategic business and MNISR support issues and strategies at several organizational levels based on the involvement of multiple personnel across business and geographic unit boundaries. Third, the utilization of a uniform MNISR planning language, process, cycle and structure similar to the business plan language, process, cycle and structure is critical. Fourth, the role of the corporate (central MNISR) staff should be well defined and limited to only those activities where its expertise is either clearly recognized and/or established by corporate policy. Fifth, corporate MNISR should provide funding for MNISR strategies that are either mandated by corporate headquarters or have a demonstrable comparative advantage to the individual multinational business components as well as the corporation as a whole.     

Proof systems for planning under 0-approximation semantics

In this paper we propose Hoare style proof systems called PR0Dand PRKW0Dfor plan generation and plan verification under 0-approximation semantics of the action language AK.In PR0D(resp.PRKW0D),a Hoare triple of the form{X}c{Y}(resp.{X}c{KWp})means that all literals in Y become true(resp.p becomes known)after executing plan c in a state satisfying all literals in X.The proof systems are shown to be sound and complete,and more importantly,they give a way to efficiently generate and verify longer plans from existing verified shorter plans by applying so-called composition rule,provided that an enough number of shorter plans have been properly stored.The idea behind is a tradeoff between space and time,we refer it to off-line planning and point out that it could be applied to general planning problems.     

General Temporal Knowledge for Planning and Data Mining

We consider the architecture of systems that combine temporal planning and plan execution and introduce a layer of temporal reasoning that potentially improves both the communication between humans and such systems, and the performance of the temporal planner itself. In particular, this additional layer simultaneously supports more flexibility in specifying and maintaining temporal constraints on plans within an uncertain and changing execution environment, and the ability to understand and trace the progress of plan execution. It is shown how a representation based on single set of abstractions of temporal information can be used to characterize the reasoning underlying plan generation and execution interpretation. The complexity of such reasoning is discussed.     

CNLs for the semantic web: a state of the art

One of the core challenges for building the semantic web is the creation of ontologies, a process known as ontology authoring. Controlled natural languages (CNLs) propose different frameworks for interfacing and creating ontologies in semantic web systems using restricted natural language. However, in order to engage non-expert users with no background in knowledge engineering, these language interfacing must be reliable, easy to understand and accepted by users. This paper includes the state-of-the-art for CNLs in terms of ontology authoring and the semantic web. In addition, it includes a detailed analysis of user evaluations with respect to each CNL and offers analytic conclusions with respect to the field.     

On the interaction between plan recognition and intelligent interfaces

Plan recognition is an active research area in automatic reasoning, as well as a promising approach to engineering interfaces that can exploit models of user's plans and goals. Much research in the field has focused on the development of plan recognition algorithms to support particular user/system interactions, such as found in naturally occurring dialogues. However, two questions have typically remained unexamined: 1) exactly what kind of interface tasks can knowledge of a user's plans be used to support across communication modalities, and 2) how can such tasks in turn constrain development of plan recognition algorithms? In this paper we present a concrete exploration of these issues. In particular, we provide an assessment of plan recognition, with respect to the use of plan recognition in enhancing user interfaces. We clarify how use of a user model containing plans makes interfaces more intelligent and interactive (by providing an intelligent assistant that supports such tasks as advice generation, task completion, context-sensitive responses, error detection and recovery). We then show how interface tasks in turn provide constraints that must be satisfied in order for any plan recognizer to construct and represent a plan in ways that efficiently support these tasks. Finally, we survey how interfaces are fundamentally limited by current plan recognition approaches, and use these limitations to identify and motivate current research. Our research is developed in the context of CHECS, a plan-based design interface.     

Text planning with opportunistic control

This report describes the current state of our central research thrust in the area of natural language generation. We have already reported on our text-level theory of lexical selection in natural language generation ([59, 60]), on a unification-based syntactic processor for syntactic generation ([73]) and designed a relatively flexible blackboard-oriented architecture for integrating these and other types of processing activities in generation ([60]). We have implemented these ideas in our prototype generator, Diogenes — a DIstributed, Opportunistic GENEration System — and tested our lexical selection and syntactic generation modules in a comprehensive natural language processing project — the KBMT-89 machine translation system ([15]). At this stage we are developing a more comprehensive Diogenes system, concentrating on both the theoretical and the system-building aspects of a) formulating a more comprehensive theory of distributed natural language generation; b) extending current theories of text organization as they pertain to the task of planning natural language texts; c) improving and extending the knowledge representation and the actual body of background knowledge (both domain and discourse/pragmatic) required for comprehensive text planning; d) designing and implementing algorithms for dynamic realization of text structure and integrating them into the blackboard style of communication and control; e) designing and implementing control algorithms for distributed text planning and realization. In this document we describe our ideas concerning opportunistic control for a natural language generation planner and present a research and development plan for the Diogenes project.Many people have contributed to the design and development of the Diogenes generation system over the last four years, especially Eric Nyberg, Rita McCardell, Donna Gates, Christine Defrise, John Leavitt, Scott Huffman, Ed Kenschaft and Philip Werner. Eric Nyberg and Masaru Tomita have created genkit, which is used as the syntactic component of Diogenes. A short version of this article appeared in Proceedings of IJCAI-89, co-authored with Victor Lesser and Eric Nyberg. To all the above many thanks. The remaining errors are the responsibility of this author.     

Generating tailored definitions using a multifaceted user model

This paper presents a computational strategy for reasoning on a multifaceted user model to generate definitions tailored to the user's needs in a task-oriented dialogue. The strategy takes into account the current focus of attention in the user's partially constructed plan, the user's domain knowledge, and the user's receptivity to different kinds of information. It constructs a definition by weighting both the strategic predicates that might comprise a definition and the propositions that might be used to fill them. These weights are used to construct a definition that includes the information deemed most useful, using information of lesser importance as necessary to adhere to common rhetorical practices. This strategy reflects our hypothesis that beliefs about the appropriate content of a definition should guide selection of a rhetorical strategy, instead of the choice of a rhetorical strategy determining content.Margaret Sarner is a Senior Consulting Software Specialist at the E.I. duPont Chemical Company. She received her Ph.D. in Computer and Information Sciences from the University of Delaware. Her research interests include user modeling and natural language generation. This paper is based on her dissertation work.Sandra Carberry is an associate professor of computer science at the University of Delaware. Her research interests include discourse understanding, user modeling, planning and plan recognition, and intelligent natural language interfaces. Much of her research has focused on developing robust models of plan recognition; this work is part of a long-term project to develop strategies for providing intelligent responses.     

TRUMP: A transportable language understanding program

Transportability has perpetually been the nemesis of natural language processing systems, in both the research and commercial sectors. During the last 20 years, the technology has not moved much closer to providing robust coverage of everyday language, and has failed to produce commercial successes beyond a few specialized interfaces and application programs. the redesign required for each application has limited the impact of natural language systems. Trump (TRansportable Understanding Mechanism Package) is a natural language analyzer that functions in a variety of domains, in both interfaces and text processing. While other similar efforts have treated transportability as a problem in knowledge engineering, Trump instead relies mainly on a “core” of knowledge about language and a set of techniques for applying that knowledge within a domain. the information about words, word meanings, and linguistic relations in this generic knowledge base guides the conceptual framework of language interpretation in each domain. Turmp uses this core knowledge to piece together a conceptual representation of a natural language input by combining generic and specialized inforamtion. the result has been a language processing system that is capable of performing fairly extensive analysis with a minimum of customization for each application.     

A Survey of Autonomous Human Affect Detection Methods for Social Robots Engaged in Natural HRI

In Human-Robot Interactions (HRI), robots should be socially intelligent. They should be able to respond appropriately to human affective and social cues in order to effectively engage in bi-directional communications. Social intelligence would allow a robot to relate to, understand, and interact and share information with people in real-world human-centered environments. This survey paper presents an encompassing review of existing automated affect recognition and classification systems for social robots engaged in various HRI settings. Human-affect detection from facial expressions, body language, voice, and physiological signals are investigated, as well as from a combination of the aforementioned modes. The automated systems are described by their corresponding robotic and HRI applications, the sensors they employ, and the feature detection techniques and affect classification strategies utilized. This paper also discusses pertinent future research directions for promoting the development of socially intelligent robots capable of recognizing, classifying and responding to human affective states during real-time HRI.