S. Feiner and K. McKeown. Coordinating Text and Graphics in Explanation Generation. In Proceedings of AAAI-90, pages 442--449, 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....communication mode, such as either text or graphics. Therefore, the automatic generation of multimodal presentations tailored to the individual user has become necessary. Current research projects in artificial intelligence like SAGE ( Roth e al. 1990] FN ANDD ( Marks and ReiteL 1990] COMET ([Feiner and McKeown, 1990]) and WIP ( Wahlster et al. 1991a] reflect the growing interest in this topic. For the knowledge based presentation system WIP, the task is the generation of a multimodal document according to the formal description of the communicative intent of the planned presentation and a set of generation ....

S. K. Feiner and K. R. McKeown. Coordinating Text and Graphics in Explanation Generation. In Proc. 8lb AAAI, pages 442-449, 1990.

....model, whereas the FN system generates natural language expressions describing certain attributes of a particular object shown in the diagrams (see [Marks Reiter90] Fig.4. Current Research on Combining Natural Language, Graphics and Pointing The WIP (see [Wahlster et al. 89] and COMET (see [Feiner McKeown 89] projects share a strong research interest in the coordination of text and graphics. They differ from the rest of the systems in that they deal with physical objects (espresso machine, radio vs. forms, maps, charts, diagrams) that the user can access directly. For example, in the WIP project we ....

Feiner, S. and McKeown, K.: Coordinating Text and Graphics in Explanation Generation. In: DARPA Speech and Natural Language Workshop (1989)

....7o, P ( o ) Figure 1: A model for the Request Communicative Act media. The specification by means of rules will make future modifications easier. The second mode of merging concerns the dialogue acts labelled as requests. It is based on a representation of CA as plan operators [5, 15, 16, 21]. Dialogue acts as well as the tactile acts are considered as low level plan operatom. Whilst merging, we solve tactile and linguistic co references [3, 8] For example, Figure 1 shows a model of the CA REQUEST for the completive mode where a tactile event and a dialogue act have to be merged. ....

Feiner S.K. and McKeown K.R. Coordinating Text and Graphics in Explanation Generation. Proceedings of the AAA[-90, July 30-August 3, 1990.

....WIP has access to an analogical representation of the geometry of the machine in the form of a wireframe model. The automatic design of multimodal presentations has only recently received significant attention in artificial intelligence research (cf. the projects SAGE (Roth et al. 89) COMET (Feiner McKeown 89) FN ANDD (Marks Reiter 90) and WIP (Wahlster et al. 89) The WIP and COMET projects share a strong research interest in the coordination of text and graphics. They differ from systems such as SAGE and FN ANDD in that they deal with physical objects (espresso machine, radio vs. charts, ....

Steven Feiner and Kathleen McKeown. Coordinating Text and Graphics in Explanation Generation. In: DARPA Speech and Natural Language Workshop, 1989.

....7o, P ( o ) Figure 1: A model for the Request Communicative Act media. The specification by means of rules will make future modifications easier. The second mode of merging concerns the dialogue acts labelled as requests. It is based on a representation of CA as plan operators [5, 15, 16, 21]. Dialogue acts as well as the tactile acts are considered as low level plan operators. Whilst merging, we solve tactile and linguistic co references [3, 8] For example, Figure i shows a model of the CA REQUEST for the completive mode where a tactile event and a dialogue act have to be merged. ....

Feiner S.K. and McKeown K.R. Coordinating Text and Graphics in Explanation Generation. Proceedings of the AAAI-90 July 30-August 3 1990.

....we use the Figure Understander only as a tool for processing an input problem description, it solves a significant research issue in its own right. The issue of integrating the differing information in diagram and text input is a non trivial problem that has been attacked by many researchers [55, 59, 18, 32, 63, 64, 80, 83, 85, 84]. The basic issue is that of establishing coreference [63] between object references in text with objects in a diagram. Solutions to the coreference problem are important because the unique information in each form of input, diagrams and text, can be integrated only after the mappings between ....

....input in raster form. 5.2.2 Related Research Issues The Figure Understander focuses on a very specific problem providing an efficient method for using text and diagrams as computer input. Some related research issues include the generation of explanations using both text and graphical input [18], generating a pictorial description from a text description of a scene [29, 49, 65] and generating a text description from a visual image of a scene [53] The solutions to these issues focus less on associating pictorial objects with text objects, and instead, concentrate on the modeling and ....

S. Feiner and K. McKeown. Coordinating text and graphics in explanation generation. In Proc. 8th National Conf. on Artificial Intelligence, Cambridge, MA, 1990. AAAI/MIT Press.

....the communication, the system must select media for realizing each piece of information and must determine how to coordinate the presentation of information in different media. Research in this area has raised interesting representation and architecture issues. In Feiner McKeown s COMET system [33], a content planner produces a representation of the information to be communicated in a media independent logical form (LF) A media coordinator annotates the LF to specify which information should be in text, which in graphics, and how text and graphics should be coordinated. Text and graphics ....

Feiner, S K and McKeown, K R `Coordinating Text and Graphics in Explanation Generation' Proc. National Conference on Artificial Intelligence, AAAI'90 (1990) pp 442-449

....software construction [TS94] and the construction of petri nets [Sto94] 5.4 Multi media approaches The most successful use of multi media presentations of information in KBS interfaces have been in application speci c rather then general purpose systems. APEX [Fei85, Fei91] and later COMET [FM90] is the main example of early work. These systems utilise a frame based database of objects together with a set of rules to determine how much detail is needed for each frame. The display of information is limited only to that which is needed and this permits a quite sophisticated combination of ....

....for WIP speci cally to address multi media layout. This framework represents three 6 CONCLUSIONS 19 types of communicative acts in an independent manner and uses a planner to coordinate textual and graphical displays. Of nal note for layout of multi media information is the COMET system [FM90] mentioned previously. COMET uses interactively generated text and 3D graphics to explain repair and maintenance tasks and uses communicative goals to govern presentation of graphics. While COMET is application speci c it takes a very interesting approach to the planning of information ....

S Feiner and K R McKeown. Coordinating text and graphics in explanation generation. In Proceedings AAAI'90, pages 442-449, Boston, MA, 1990.

....MATS. Status: Fully implemented prototype. Impact: Enabling technology for automatic or interactive response to temporal patterns of events. 1 Introduction Terminological knowledge representation (tkr) systems are used to represent and reason with conceptual knowledge in many application areas [21, 2, 22, 6, 9, 17]. However, the applicability of contemporary tkr systems is limited by restrictions on both their representation language and their inferences [8] Our work extends the scope and utility of tkr by representing complex constraint networks in a terminological framework, and by extending the range ....

....3 Metric arc M1 i1F ( Gammam1; Q1; R1; n1)j1G w metric arc M2 i2F ( Gammam2; Q2; R2; n2)j2G iff (1) Gammam1 Gammam2, 2) Q1 wQ2, 3) R1 wR2, and (4) n1 n2. Whenever M2 describes a metric relationship, M1 must also describe it. The constraint i1F (6; 9)j1G describes 6 but 9 units, i. e, [6,9], and subsumes i2F (7; 8)j2G describing 7 but 8 units, i.e. 7,8] Node i1 will be mapped to i2, and j1 to j2, as discussed below. Definition 4 Equality arc E1 rname i (c1) rname j (d1) w equality arc E2 rnamek (c2) rname l (d2) iff rname i = rnamek rname j = rname l . Whenever E2 ....

S. Feiner and K.R. McKeown. Coordinating text and graphics in explanation generation. In AAAI-90, pages 442--449, Boston, MA, 1990.

....criteria. 14 Where a multimodal information presentation system can choose between realizing some piece of information via, for example, natural language or a map, the issues we have discussed should arise. Computational methods for synthesising multimodal presentations have been investigated by Feiner and McKeown (1990), Marks and Reiter (1990) Badler et al. 1991) Maybury (1991) Roth et al. 1991) and Wahlster et al. 1991) issues concerning declarative knowledge about presentation modalities have been investigated by Arens and Hovy (1990) and Hovy and Arens (1991) In this area, is acknowledged that the ....

Feiner, S. K. and McKeown, K. R. (1990). Coordinating text and graphics in explanation generation. In Proceedings of Eighth National Conference on Artificial Intelligence, Boston, July 29--August 3, 1990.

.... behavior models, the explanations typically follow from hard coded labeling of causal influence [21] or component function [14] Much of the work in explanation has concentrated on the generation of high quality presentations in natural language based on discourse planning and user modeling [7,17,18,19]. These presentation techniques are independent of the modeling method or technique for determining causal influence, and so could be adapted for the explanation approach presented in this paper. The task of giving a causal interpretation to device behavior has been addressed from several ....

S. K. Feiner & K. R. McKeown. Coordinating text and graphics in explanation generation. AAAI-90, pp. 442-449, 1990.

....individual components would be played back making use of the visualization of layer one. This is akin to a video of the individual parts being dismantled being played back to the user. This approach communicates causal information quite easily and is related somewhat to the concepts discussed in [Feiner90]. The underlying link between the three layers comes in the form of explanation. Explanation of why the physical components in a subsystem have caused it to fail are much easier to generate with a visual aid such as those available in layers one and two. An explanation of failure is much easier to ....

Feiner S. and McKeowen K. Coordinating Text and Graphics in Explanation Generation. Proceedings of the Eighth National Conference on Artificial Intelligence, Boston, Massachusetts. July 29-August 3, 1990. pp. 442-449.

.... a plausible explanation for a faulty input be used for continuing an interaction ) the input interpreter (what can be abducted from an input consisting of errors and incomplete information ) the generator (is it possible to transmit the same information using another mode of communication [FM90] In section 2 we will describe the two headed architecture de ned in order to handle the above mentioned multimodal re ective interactions. The interactional decision center consists of an interactional handler and several low level executive modules (see section 3) The re ectional level, ....

Steven Feiner and Kathleen McKeown. Coordinating text and graphics in explanation generation. In Proceedings of the AAAI'90, pages 442-449, 1990.

....recent work might be to say that he has been concerned with semantics, where Mackinlay s approach has been more or less syntactic. Others have explored the use of user specified constraints to improve automatic graph layout [BP90] 20 Extensions to non standard displays have also been proposed [FM90] and some early implementations have been developed. 16 Bertin includes much detail about media and printing techniques that need not concern us here. 17 Cleveland and McGill developed the ranking of quantitative tasks given here as the first column of table 2; the other two rankings in the ....

Steven K. Feiner and Kathleen R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings AAAI, pages 442--449, Boston, MA, July 1990.

....user in whatever modality is most suitable for the task at hand. It is useful to distinguish between three kinds of multimodality: Graphics Output: Much research has been done on generating diagrams (and other graphical presentations of data) and text from a single domain knowledge base, eg, Feiner and McKeown, 1990; Wahlster et al. 1991 ] Graphics output requires a different low level realisation rendering module than text output, but in some cases high level content oriented modules can be be used for both text and graphics output [ Andr e and Rist, 1993 ] Visual Formatting: Text can be much more ....

....is a definite weakness of the system, and we suspect that an automatic technical documentation generation system that is used in real applications may need to possess more sophisticated graphic abilities. There has been a fair amount of research on combining text and graphics generation (eg, Feiner and McKeown, 1990; Wahlster et al. 1991 ] but this work has tended to stress very principled ways of doing this, which may be too costly (in terms of both the amount of domain knowledge and the amount of compute time required) to be practical in realistically sized systems; further research probably needs to ....

Steve Feiner and Kathleen McKeown. Coordinating text and graphics in explanation generation. In Proceedings of the 8th National Conference on Artificial Intelligence (AAAI-1990), volume 1, pages 442--449, 1990.

....different types of documents needed for a complex system (user manual, maintenance guide, design description, etc. could be generated from a single integrated domain KB. Multimodal documents: Generation systems can automatically produce pictures as well as texts from a domain knowledge base [ Feiner and McKeown, 1990; Wahlster et al. 1991 ] and can embed mousable hypertext links in the textual portion of a message [ Stock, 1991; Reiter et al. 1992 ] Multilingual Generation: An NL generation system can in principle generate texts in multiple languages from the same domain knowledge base [ Hovy et al. ....

Steve Feiner and Kathleen McKeown. Coordinating text and graphics in explanation generation. In Proceedings of the 8th National Conference on Artificial Intelligence (AAAI-1990), volume 1, pages 442--449, 1990.

.... are more easily understood than complex digitised images; # combine visual techniques with contrast only when appropriate : the combination of techniques increases the visual impact of the frame, but uncareful combination may destroy the intended purpose; Automatic Layout Generation GRIDS [8,9,10] is a good example of a system for automatically generating a layout. It allows automated layout of text, pictures and virtual interaction devices according to a layout grid. From the architecture point of view, Feiner [9] recommends to build a media layout component for laying out IO. This ....

....application from specifications As a matter of fact, our knowledge about multimedia IO is insufficient. Nevertheless, we are able to say some things : # multimedia IO could be qualified with attributes [28] # reasonable proportions for displaying images and video sequences are already known [9,10]; # visualisation techniques for special IO (e.g. 3D objects, maps) are well studied [12] # the layout frame should be governed by these attributes and information [2] It is sure that the introduced visual techniques could be part of such a layout generator, but one cannot assume that these ....

Feiner, S., McKeown, K. (1990). Coordinating Text and Graphics in Explanation Generation. In Proceedings of AAIA-90, pp. 442-449.

.... when appropriate : the combination of techniques increases the visual impact of the frame, but uncareful combination may destroy the intended purpose; multimedia IO could be qualified with attributes [28] reasonable proportions for displaying images and video sequences are already known [9,10]; Automatic Layout Generation . visualisation techniques for special IO (e.g. 3D objects, maps) are well studied [12] GRIDS [8,9,10] is a good example of a system for automatically generating a layout. It allows automated layout of text, pictures and virtual interaction devices according to a ....

....the intended purpose; multimedia IO could be qualified with attributes [28] reasonable proportions for displaying images and video sequences are already known [9,10] Automatic Layout Generation . visualisation techniques for special IO (e.g. 3D objects, maps) are well studied [12] GRIDS [8,9,10] is a good example of a system for automatically generating a layout. It allows automated layout of text, pictures and virtual interaction devices according to a layout grid. From the architecture point of view, Feiner [9] recommends to build a media layout component for laying out IO. This ....

Feiner, S., McKeown, K. (1990). Coordinating Text and Graphics in Explanation Generation. In Proceedings of AAIA-90, pp. 442-449.

....in an illustration. The evaluators can be thought of as psychophysically motivated constraints on the presentations. 10 10 See perceptual salience in Section 5.4.1. 55 Still more work explores the coordination of different media (text and graphics in particular) in single presentations [73]. WIP The WIP project at DFKI under the direction of Wolfgang Wahlster is similar to Feiner s IBIS system in that it too uses underlying static models of objects and tasks to determine the layout of a presentation. The WIP researchers say of their approach that it should be understood as a ....

Steven K. Feiner and Kathleen R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings AAAI, pages 442--449, Boston, MA, July 1990.

.... In addition, such a system could should learn from and adapt to its environment and improve its interaction with the user [Thies94, Wittburg94] Coordinating text, graphics, and images in a hypermedia environment is very important in effective presentation and management of the subject matter [Feiner90, Augenstein93, Hekmat94]. Some researchers have investigated the automatic generation of coordinated multimedia [Feiner91, Maybury91] Others have studied controlling and managing the different elements composing a 2 Hekmatpour 5 30 95 E 16 multimedia document [Merlet93, Hekmat94] And still others have proposed ....

Feiner S. and McKeown K., "Coordinating Text and Graphics in Explanation Generation," Proc. AAAI--90, pp. 442--449, Boston, MA, July 29--August 3, 1990. 20 Hekmatpour 5/30/95<E-16>

....from an analysis of over 30 user guides manuals for consumer appliances and discuss some of the implications. Introduction Instructional texts have been the object of many studies recently, with an emphasis on methods for integrating graphics and text, as in WIP (Wahlster et al. 1993) and COMET (Feiner and McKeown, 1990), for tailoring to the user (Peter and R osner, 1994) for generating purpose expressions in English (Vander Linden, 1993) for producing multilingual instructions, e.g. R osner and Stede, 1991) and for planning the appropriate referring expressions, e.g. Dale, 1992) Most of This work is ....

Feiner, S. K. and McKeown, K. R. (1990). Coordinating text and graphics in explanation generation. In Proceedings of the National Conference on Artificial Intelligence, pages 442--449, Boston, MA.

.... leaving technical communicators out of the loop, e.g. 20, 2] Similarly, systems developed to support the production of technical documentation (in one language) have also been fully automated systems, bypassing the technical communicator totally, e.g. Wahlster et al. [25] Feiner and McKeown [9], Peter and Rosner [17] Vander Linden [24] Dale [6] Kosseim and Lapalme [15] and Reiter et al. [19] In contrast, our goal within the drafter project [16] is to provide tools to support rather than replace the translator, designed with the benefit of interviews with professional linguists to ....

Steve Feiner and Kathleen McKeown. Coordinating Text and Graphics in Explanation Generation. In Proceedings of the Eighth National Conference on Artificial Intelligence, pages 442--449, Boston, Massachusetts, July 29 - August 3 1990.

....analysis is performed by Debrief, information is presented to the user. This presentation is performed via a hierarchical presentation planning process, initiated in the Present problem space shown in Figure 2. The planning process is similar to that of other multimedia generation systems [6, 1], although its ability to coordinate text and graphics is somewhat limited. 8.1 Selecting information to present The first step in the presentation process is selecting what information should be presented. If the question that was asked involved evaluating a decision or belief, this step is ....

S.K. Feiner and K.R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings of the Eighth National Conference on Artificial Intelligence, pages 442--449, Anaheim, CA, August 1990. MIT Press.

....shown in Figure 8. 3 Communication with the Reasoning System The internal representation for the purposes of reasoning is in terms of logical predicates. The external form is in terms of spoken sentences, graphical objects (pointers or highlighting, in our case) and displayed text. A translation [4, 8] is needed between the two forms so that the user can communicate with the Attempt proof of top level goal. If success then halt. While top level goal is not achieved: Select failed subgoal If system has initiative, its own scoring will govern choice. If user has initiative, the choice will ....

S.K. Feiner and K.R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings of the 8th National Conference on Artficial Intelligence, volume I, pages 442--449. AAAI Press/The MIT Press, 1990.

....text quality. Moreover, Knight currently employs very rudimentary pronominalization techniques. Including more sophisticated methods [12] should result in a significant increase in text quality. Finally, one of the most fruitful area for future work is research on multimedia explanation generation [17, 36, 65, 51, 10]. A multimedia explanation generator can combine the explanatory capabilities of Knight with the powerful display technologies of multimedia devices. By tightly coupling multimedia object indices to a formal representation of domain knowledge, a multimedia explanation generator should be able to ....

S. K. Feiner and K. R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings of the Eighth National Conference on Artificial Intelligence, pages 442--449, Boston, Massachusetts, 1990.

.... Natural language researchers have studied reference generation, e.g. Dale s classic work on referring expressions [Dale, 1992] scene description generation [Novak, 1987] and spatial layout description generation [Sibun, 1992] Work on intelligent multimedia systems [Andr e et al. 1993, Feiner and McKeown, 1990, Maybury, 1991, Roth et al. 1991, Mittal et al. 1995] has produced techniques for dynamically incorporating highlights, underlines, 1 The same study employed pre and post tests to evaluate learning effectiveness and found statistically significant gains in students performance [Lester et ....

Feiner, S. K. and McKeown, K. R. (1990). Coordinating text and graphics in explanation generation. In Proceedings of the Eighth National Conference on Artificial Intelligence, pages 442--449, Boston, MA.

....reason with conceptual knowledge required by intelligent software applications. Examples include database querying [Beck and Gala, 1989] financial marketing [Apte et al. 1992] software information systems [Devanbu et al. 1991] and multimedia explanation of repair and maintenance procedures [Feiner and McKeown, 1990]. However, contemporary terminological systems are limited by their inability to handle complex compositions of concepts. Therefore, we propose to extend their scope and utility via terminological constraint networks, whose nodes are described by associated concepts. Noting that much artificial ....

....technologies in turn. 3. 1 Terminological Knowledge Representation There is ample evidence that systems of the KL ONE family are well suited for representing the classes of actions which make up plans and, in turn, the objects that are acted upon, e.g. Apte et al. 1992, Devanbu et al. 1991, Feiner and McKeown, 1990, Heinsohn et al. 1992, Wellman, 1990] The system we propose will build upon such a knowledge representation system and extend its capabilities to reason with structured plan descriptions. 3.1.1 Concept Languages and their Semantics Terminological knowledge representation, which originated with ....

S. Feiner and K.R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings of AAAI-90, pages 442--449, Boston, MA, 1990.

....also builds upon work in information science, psychology and linguistics. In particular, the design of MERIT addressed problems of cooperative graphical or multimodal interfaces [cf. Hayes 1987, Neal Shapiro 1989; Cohen et al. 1989; Allgayer et al. 1989; Arens Hovy 1990; Moore Swartout 1990; Feiner McKeown 1990; Bandyopadhyay 1990; Stock 1991] Another body of related work aims at the plan based or task driven automatic generation of graphics [e.g. Wahlster et al. 1991; Oei et al. 1992] and the case based composition of video presentations [MacNeil 1991] However, these approaches are not interactive in ....

Feiner, S.K. & McKeown, K.R.: Coordinating Text and Graphics in Explanation Generation. In: AAAI '90, Proc. of the 8th National Conference on Artificial Intelligence, Vol. I. Menlo Park et al.: AAAI Press / The MIT Press, 1990, pp. 442-449.

....automated generation of multimedia presentations have been proposed. The clarity of the presentations has generally been taken into account in that the media allocation relies on rules or constraints based on common sense or psychological research findings see, e.g. BOZ (Casner, 1991) COMET (Feiner McKeown, 1990), APT (Mackinlay, 1986) SAGE (Roth, Mattis, Mesnard, 1991) WIP PPP (Andr e, 1995; Wahlster et al. 1993) and the approach presented by Arens, Hovy, van Mulken (1993) However, that aspect of multimedia presentation understanding that seems to be one of the central criteria for the ease of ....

Feiner, S., and McKeown, K. 1990. Coordinating text and graphics in explanation generation. In Proceedings of the Eighth Conference of the American Association of Artificial Intelligence, 442--449.

.... [MacGregor, 1991b] and sb one [Kobsa, 1991] It has seen a correspondingly diverse range of applications, e.g. financial marketing expertise [Apt e et al. 1992] software information retrieval [Devanbu et al. 1991] knowledge based presentation of multimedia information for equipment repair [Feiner and McKeown, 1990] and for equipment operation [Wahlster et al. 1993] conflict resolution in production systems [Yen et al. 1991] and system configuration [Owsnicki Klewe, 1988; Searls and Norton, 1990; Wright et al. 1993; Weida, 1996] k rep is currently being used in a clinical information system to ....

S. Feiner and K.R. McKeown. Coordinating text and graphics in explanation generation. In Proceedings of AAAI-90, pages 442--449, Boston, MA, 1990.

....data and required inputs as well as manipulations. Also, unlike other help generation systems, both Cartoonist and H3 generate help that depends on the context of an application and its interface. Techniques based on AI technology have been proposed to produce help and documentation systems, [1, 6]. The AI systems have more sophisticated natural language generation facilities, and also include planners to construct help texts with better discourse structures than those in H3. However, the AI systems are typically difficult to integrate into applications because they require much more ....

Feiner, S., and McKeown, K.: Coordinating Text and Graphics in Explanation Generation. Proceedings AAAI-90. Boston, MA, 1990.

.... contrast, provides less flexibility because its output is typically constrained by the style and language of the source text (i.e. the original text) Most automated generation systems developed to date are intended to be used as stand alone tools, leaving technical writers out of the loop (e.g. [4, 12, 9]) They assume that an underlying knowledge base containing all the information necessary to produce instructions (or documentation) is already available, or can be easily obtained. This, however, is unlikely in the near future. Text generation is a very knowledge intensive task. Furthermore, the ....

Steve Feiner and Kathleen McKeown. Coordinating Text and Graphics in Explanation Generation. In Proceedings of the Eighth National Conference on Artificial Intelligence, pages 442--449, Boston, Massachusetts, July 29 - August 3 1990.

.... the bewildering amount of relevant information, we spent some time surveying several relevant fields, including some of the more theoretical literature on good presentation design [35, 5, 34] recent studies in Psychology [14, 36, 9, 12] Cognitive Science [30, 33, 27, 20] and Computer Science [21, 39, 11, 29, 26]. It was clear that the large number of parameters playing a role, and their complex interrelationships, makes it imperative that features be defined with care at the proper level of generality to ensure their successful interaction with one another. As a result of our reading, and after ....

....producing any output. The wip system [39, 1, 13] plans a text graphics description of the use of an espresso machine, starting with a database of facts about the machine and appropriate communicative goals, and using text and presentation plans to build a presentation tree. The comet system [10, 11] plans text graphic presentations of a military radio using text schemas and pictorial perspective presentation rules. The aimi system [26, 6] builds a tree to produce presentations using text, maps, and tables about database information of military operations and hardware. A question about which ....

Feiner, S. and McKeown, K.R. 1990. Coordinating Text and Graphics in Explanation Generation. In Proceedings of the 8th AAAI Conference (442--449).

....design. However, such recommendations for graphic layout are aimed to be a knowledge resource for the more experienced interface designers. Modelling Approaches Formalised methods to produce presentations have been developed since the late 1980s as part of automatic document synthesis systems [7] [17] 2] These algorithms plan a presentation by employing modelled semantic information about the involved objects and their interrelationships, as well as rhetorical information in this context mostly referred to as communicative goals. However, the driving force in this area has been the ....

Feiner, S.K. and McKeown, K.R. Coordinating text and Graphics in Explanation Generation. In: Proceedings of AAAI'90. Boston, MA, 1990

....application [Roth89] ffl Three Dimensional Illustrations. The APEX [Fein85] and IBIS [Seli89] systems produce illustrations that depict objects and actions in the physical world. The illustrations satisfy communicative goals generated automatically by a multimedia explanation generator [Fein90]. ffl User Interface Displays. Several systems have been built that are capable of generating automatically the graphical objects needed for user interface displays [Aren88, Kim90, Wiec90] ffl Network Diagrams. The ANDD system [Mark90a, Mark90b] designs and articulates network diagrams to ....

Feiner, S. and McKeown, K. 1990. "Coordinating text and graphics in explanation generation," Proceedings of AAAI '90, pages 442--449, August, Boston, Massachusetts.

....manuals coupled with the cost of technical writing, the time required to produce documentation, and the potential flexibility offered by the automatic generation of instructions. Researchers have concentrated on designing methods for integrating graphics and text, e.g. Wahlster et al. 1993; Feiner and McKeown, 1990] and for tailoring instructions to the user s level of expertise, e.g. Peter and R osner, 1994] At a more linguistic level of concern, others have studied various ways of realising purpose expressions in English, e.g. Vander Linden, 1993] and of generating appropriate referring ....

Feiner, Steve and McKeown, Kathleen 1990. Coordinating Text and Graphics in Explanation Generation. In Proceedings of the Eighth National Conference on Artificial Intelligence, Boston, Massachusetts. 442--449.

....to automate more general multimedia computer presentation design. Work on automatic presentation is evolving along two distinct lines: systems dealing only with graphics (APT [Mackinley 86b] BOZ, SAGE) and systems dealing with a combination of natural language and line drawings (II, COMET [Feiner McKeown 90] WIP [Wahlster et al. 91] Systems of the first type concentrate entirely (or primarily) on one medium of presentation, and hence do not develop generic models of media capabilities. As discussed in [ISAT 91] however, generic models (called there Virtual Devices) are necessary to ensure the ....

Feiner, S. and McKeown, K.R. 1990. Coordinating Text and Graphics in Explanation Generation. Proceedings of the 8th AAAI. (442--449).

....the system of [Wahlster et al. 92, Andr e Rist 92] and see this volume) plans a text graphics description of the use of an espresso machine, starting with a database of facts about the machine and appropriate communicative goals, and using text and presentation plans. The system [Feiner 88, Feiner McKeown 90] plans text graphic presentations of a military radio using text schemas and pictorial perspective presentation rules. The system [Maybury 91, Burger Marshall 91] and see this volume) plans text map tables presentations of database information about military operations two stage ....

Feiner, S. and McKeown, K.R. 1990. Coordinating Text and Graphics in Explanation Generation. In (442--449).

.... alone; that is one cannot just refer and predicate without making an assertion or asking a question or performing some other illocutionary acts [Searle,1969] Quite elaborated theories have been proposed for multimodal output (e.g. Wahlster et al. 1992] Maybury,1993b] Arens et al. 1993] [Feiner and McKeown,1990]) that, each in an autonomous way, have taken into account communicative intentions and illocutionary acts; prototypes have shown the validity of the results. As far as interactive multimodal dialogue systems are concerned, in the first place we can consider systems where language is used with a ....

S. Feiner and K. McKeown. Coordinating text and graphics in explanation generation. In Proceedings AAAI-90, 1990.

....interface, the content planner uses schemas [McKeown 85] to determine which information should be included from the underlying knowledge sources in the explanation. The explanation content, represented as a hierarchy of logical forms (LFs) McKeown et al. 90] is passed to the media coordinator [Feiner and McKeown 90b, Elhadad et al. 89] which refines the LFs by adding directives indicating which portions are to be produced by the text generator and which by the graphics generator [Seligmann and Feiner 89] Figure 1: COMET System Architecture The Lexical Chooser is part of the text generator [McKeown et al. ....

Feiner, S. and K.R. McKeown. Coordinating Text and Graphics in Explanation Generation. In Proceedings of the National Conference on Artificial Intelligence. Boston, Mass., August, 1990. 16

No context found.

S. Feiner and K. McKeown. Coordinating Text and Graphics in Explanation Generation. In Proceedings of AAAI-90, pages 442--449, 1990.

No context found.

FEINER, S., AND MCKEOWN, K. R. 1989. Coordinating text and graphics in explanation generation. In Proceedings of the Speech and Natural Language Workshop, 424--433.

No context found.

Steven Feiner and Kathleen McKeown. 1990. Coordinating text and graphics in explanation generation. In Proc. of AAAI-90, pages 442--449, Boston, MA.

No context found.

Feiner, S. andMcKeown, K.R., Coordinating text and graphics in explanation generation. In Proceedings of the AAAI-90, PP442-449, Boston, 1990

No context found.

Steven Feiner and Kathleen R. McKeown. Coordinating text and graphics in explanation generation. In Speech & NL Workshop [1092], pages 424--433.

No context found.

Feiner, S. and K. McKeown. Coordinating Text and Graphics in Explanation Generation. In: Proceedings Eighth National Conference on Artificial Intelligence (AAAI-90), 1990.

No context found.

Feiner, S. and McKeown, K. R. Coordinating Text and Graphics in Explanation Generation. In Proceedings of the Eleventh AAAI Conference, pages pp.442-449. American Association of Artificial Intelligence, 1990.

No context found.

Steve Feiner and Kathleen McKeown. 1990. Coordinating text and graphics in explanation generation.

No context found.

Feiner S. and McKeown K., "Coordinating Text and Graphics in Explanation Generation," Proc. AAAI--90, pp. 442--449, Boston, MA, July 29--August 3, 1990.

No context found.

Feiner, S. and K. McKeown, Coordinating Text and Graphics in Explanation Generation, Proceedings of AAAI-90, Boston, July 1990.

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