Elizabeth Bratt

Roku · Voice

Intelligent tutoring systems are becoming ever more ubiquitous in training environments that are simulation-based. The U.S. Navy shiphandling training program at the Surface Warfare Officers School (SWOS) in Newport, RI is one such environment. In this program, students are placed into a multi-instrumented, virtual ship’s bridge where they assume t...

The goal of an Intelligent Tutoring System (ITS) is to improve training efficiency by monitoring student performance and providing automated tutoring advice with the goal of increasing student learning and throughput. Traditional ITS development has focused on static problems, such as math and physics (Koedinger, Anderson, Hadley, & Mark, 1997; Van...

This paper describes the role of simulation-based training in the military. Interviews and observations of military instructors in the damage control and shiphandling domains provide examples of how the instructors extend the student;s training beyond the well-defined simulated world with qualitative reasoning about context, hypothetical variants,...

In designing and building tutorial dialogue systems it is important not only to understand the tactics employed by human tutors but also to understand how tutors decide when to use various tactics. We argue that these decisions are based not only on student problem-solving steps and the content of student utterances, but also on the meta-communicat...

The goal of the Conversational Interfaces project at CSLI is to develop a general purpose architecture which supports multi-modal dialogues with complex devices, services, and applications. We are developing generic dialogue management software which supports collaborative activities between a human and devices. Our systems use a common software ba...

In this paper we describe the ways that SCoT, a Spoken Conversational Tutor, uses flexible and adaptive planning as well as multimodal task modeling to support the contextualization of learning in reflective dialogues. Past research on human tutoring has shown reflective discussions (discussions occurring after problem-solving) to be effective in h...

The Navy is shifting its training and education from traditional methods, such as on-site instruction, texts, and observing students during drills, to computer-supported learning such as web-based instruction and computer simulations in lieu of live drills. This transition presents the challenge of keeping the best parts of traditional methods of i...

The ability to lead collaborative discussions and appropriately scaf- fold learning has been identified as one of the central advantages of human tu- torial interaction (6). In order to reproduce the effectiveness of human tutors, many developers of tutorial dialogue systems have taken the approach of identi- fying human tutorial tactics and then i...

We describe SCoT, a Spoken Conversational Tutor, which has been implemented in order to investigate the advantages of natural language in tutoring, especially spoken language. SCoT uses a generic architecture for conversational intelligence which has capabilities such as turn management and coordination of multi-modal input and output. SCoT also in...

This demonstration shows a flexible tutoring system for studying the effects of different tutoring strategies enhanced by a spoken language interface. The hypothesis is that spoken language increases the effectiveness of automated tutoring. The domain is Navy damage control.

Contextualizing learning in an intelligent tutoring system is difficult for many reasons. Goals such as presenting material in an understandable manner, minimizing confusion and frustration, and helping the student reason about their actions all need to be balanced. Previous research has shown reflective discussions (with human tutors) occurring af...

Scalability and reusability of tutorial dialogue systems is a function of the corresponding characteristics in their component tutoring and dialogue technologies.

Speculative execution of information gathering plans can dramatically reduce the effect of source I/O latencies on overall performance. However, the utility of speculation is closely tied to how accurately data values are predicted at runtime. Caching ...

Shipboard damage control displays (at least) two features that distinguish it from other domains like math: (i) non-determinism (e.g., actions have unexpected outcomes) and (ii) a dynamic problem state (e.g., a problem increases in complexity over time). These features impact how human-to-human tutoring is conducted; e.g., human tutors must teach s...

The CommandTalk military spoken dialogue system uses Semantic Head-Driven Generation in Gemini grammars, which have compositional semantics, permitting the iden- tification of parse tree nodes corresponding to logical form subexpressions. This method of generation gives the Com- mandTalk prosody agent access to appropriate syntactic and semantic in...

This paper describes the evaluation methodology and results of the 2001 DARPA Communicator evaluation. The experiment spanned 6 months of 2001 and involved eight DARPA Communicator sys- tems in the travel planning domain. It resulted in a corpus of 1242 dialogs which include many more dialogues for complex tasks than the 2000 evaluation. We describ...

The goal of the Conversational Interfaces project at CSLI is to develop a general purpose architecture which supports multi-modal dialogues with devices. Our systems use a common software base consist- ing of the Open Agent Architecture, Nuance speech recogniser, Gemini (SRI's parser and generator), Fes- tival speech synthesis, and CSLI's "Conversa...

This paper describes the evaluation methodology and results of the DARPA Communicator spoken dialog system evaluation experiments in 2000 and 2001. Nine spoken dialog systems in the travel planning domain participated in the experiments resulting in a total corpus of 1904 dialogs. We describe and compare the experimental design of the 2000 and 2001...

This paper describes extensions to CommandTalk to support spoken dialogue. While we make no theoretical claims about the nature and structure of dialogue, we are influenced by the theoretical work of (Grosz and Sidner, 1986) and will use terminology from that tradition when appropriate. We also follow (Chu-Carroll and Brown, 1997) in distinguishing...

The intelligent tutoring system based on the DC-TRAIN simulation requires specific information of various sorts about the student's performance during a session, in order to provide a focussed instructional review. The flexible dialogue architecture allows for dynamic system adaptation to student performance during the dialogue.

This paper describes results of an experiment with 9 different DARPA Communicator Systems who participated in the June 2000 data collection. All systems supported travel planning and utilized some form of mixed-initiative interaction. However they varied in several critical dimensions: (1) They targeted different back-end databases for travel infor...

We describe robustness techniques used in the CommandTalk system at the recognition level, the parsing level, and the dialogue level, and how these were influenced by the lack of domain data. We used interviews with subject matter experts (SME's) to develop a single grammar for recognition, understanding, and generation, thus eliminating the need f...

Systems now exist which are able to compile unification grammars into language models that can be included in a speech recognizer, but it is so far unclear whether non-trivial linguistically principled grammars can be used for this purpose. We describe a series of experiments which investigate the question empirically,by incrementally constructing...

Systems now exist which are able to compile unification grammars into language models that can be included in a speech recognizer, but it is so far unclear whether non-trivial linguistically principled grammars can be used for this purpose. We describe a series of experiments which investigate the question empirically, by incrementally constructing...

This paper will focus on how two representations of context are used in CommandTalk to correctly interpret the user's spoken utterances: situational context represents the current state of the simulation, and linguistic context represents the history of the user's linguistic acts.

. In this paper, we're showing how we are improving the interface to a graphical simulation by adding natural inputs, such as Speech and Gesture Recognition, and Natural Language Understanding using an Agent based Architecture. 1. Overview ModSAF provides both 2D and 3D (CommandVu, Fig.1) representations of a field where the users, the commanders,...

This thesis explores the properties of Korean lexical and periphrastic causatives as a key to issues of constituent structure, case marking, complementation, and the organization of the grammar into the lexicon and syntactic structure. The analysis, set within the framework of Head-Driven Phrase Structure Grammar, treats both forms of causative as...

The Navy is shifting its training and education from traditional methods, such as on-site instruction, texts, and observing students uring drills, to computer-supported learning such as web-based nstruction and computer simulations in lieu of live drills. This transition presents the challenge of keeping the best parts of traditional methods of ins...

SCoT is a tutorial dialogue system that engages students in natural language discussions through a speech interface. The current instantiation, SCoT-DC, is applied to the domain of shipboard damage control—the task of containing the effects of crises (e.g. fires) that occur aboard Navy vessels. This paper describes a recent evaluation of SCoT-DC an...

Situations where a spoken dialogue system cannot interpret a user's utterance are a major source of frustration in human- computer spoken dialogue. Current spoken dialogue sys- tems generally respond with an unhelpful "I'm sorry, I didn't understand," or something similarly uninformative. Recent work on Targeted Help has shown that giving users mor...

Scalability and reusability of tutorial dialogue systems is a function of the corresponding characteristics in their component tutoring and dialogue technologies. This paper discusses an architecture for a scalable, reusable spoken conversational tutor, SCoT. With this design we hope to minimize the efforts needed to reuse the components for implem...

Systems now exist which are able to con:pile unification gralmnars into language models that can be included in a speech recognizer, but it is so far unclear whether non-trivial linguisti- cally principled gralnlnars can be used for this purpose. We describe a series of experiments which investigate the question empirica.lly, by incrementally const...

Training is increasingly becoming technology-based, swapping classroom time and live instructors for distance learning, serious games, and simulation exercises. This presents both challenges and opportunities for tailoring training to accommodate differences in trainees' backgrounds, prior knowledge, and abilities. Our interest is in building compr...

VTT Working Papers 101 Global skill shortages are reported in many occupations. Existing strategies for addressing skill shortages are not successful and, as a result, skill shortages are an intractable problem. A new strategy, real-time communication of skill knowledge without human instructors, has the potential to bring about radical reductions...