Abstract Historically, the coverage bias from excluding the United States cell-only population from survey samples has been minimal due to the relatively small size of this group. However, the unrelenting growth of this segment has... more
Abstract Historically, the coverage bias from excluding the United States cell-only population from survey samples has been minimal due to the relatively small size of this group. However, the unrelenting growth of this segment has sparked growing concern that telephone surveys of the general public in the United States will become increasingly subject to coverage bias. While there is evidence that demographic weighting can be used to eliminate this bias, the availability of the weights lag behind the rapidly changing cell-only population. To explain the extent of the problem, we propose a reliable model to forecast cell-only population size and demographics. This model posits that a stable behavioral process, the rate of habit retention, can be estimated from prior wireless lifestyle adoption in the United States and may also describe adoption of the cell-only lifestyle. Using measures of incentive and habituation, we test this assumption by predicting changes in the cell-only population size and changes in age demographics. The accuracy of predictions confirms the two adoption behaviors are similar. We then develop forecasts of age demographics through 2009, and show how cell-only lifestyle adoption leads to potential coverage bias that is better addressed through this type of modeling rather than weighting from historical data. For some time, investigators have noted that the emergence of a population that uses only cellular phones may lead to coverage bias in telephone surveys of American residents, since random-digit dial survey practices in the United States do not usually include cell phones in their sample frames (Kuusela 2003;
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
ABSTRACT Next-generation multimodal systems designed for use in mobile environments present challenges to the task of data collection that are not faced by speech-based systems. We discuss some established data collection and evaluation... more
ABSTRACT Next-generation multimodal systems designed for use in mobile environments present challenges to the task of data collection that are not faced by speech-based systems. We discuss some established data collection and evaluation methods and their limitations in the context of a mobile multimodal system. These limitations are addressed by the "on-board" multimodal data collection method developed for MATCH, a multimodal mobile city guide. Our approach exploits MATCH's component architecture in that each component can be redeployed in evaluation and annotation tools, allowing user test sessions to be replayed with a high degree of fidelity without the use of recorded video. Instead, the components themselves perform a dynamic re-enactment of test sessions directed by the script of a comprehensive log file. This method enabled continual user testing and piloting to inform the iterative development process for MATCH.
Research Interests:
Research Interests:
Research Interests:
<p>Experimental design and procedure.</p
Research Interests:
<p>Disclosure effects for each question.</p
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
ABSTRACT Speak4it℠ uses a multimodal interface to perform mobile search for local businesses. Users combine simultaneous speech and touch to input queries or commands, for example, by saying, &amp;quot;gas stations,&amp;quot;... more
ABSTRACT Speak4it℠ uses a multimodal interface to perform mobile search for local businesses. Users combine simultaneous speech and touch to input queries or commands, for example, by saying, &amp;quot;gas stations,&amp;quot; while tracing a route on a touchscreen. This demonstration will exhibit an extension of our multimodal semantic processing architecture from a one-shot query system to a multimodal dialogue system that tracks dialogue state over multiple turns and resolves prior context using unification-based context resolution. We illustrate the capabilities and limitations of this approach to multimodal interpretation, describing the challenges of supporting true multimodal interaction in a deployed mobile service, while offering an interactive demonstration on tablets and smartphones.
Research Interests:
Research Interests:
<p>Multitasking across the four modes.</p
Research Interests:
Research Interests: Psychology and PLoS one
Research Interests:
Research Interests: Computer Science, Communication, Knowledge Management, Semantics, Augmented Reality, and 12 moreSurvey Research, Computational Semantics, Digital Communication, Literature Review, Instructional Strategies, Project Planning, Originality, Skills, Design Methodology, Semantic Information, Just in Time, and Information media
Research Interests:
The current issue and full text archive of this journal is available at www.emeraldinsight.com/1753-8378.htm
Research Interests:
@Book{ACTS:2006, editor = {Eduard Hovy and Klaus Zechner and Liang Zhou}, title = {Proceedings of the Analyzing Conversations in Text and Speech}, month = {June}, year = {2006}, address = {New York City, New York}, publisher =... more
@Book{ACTS:2006, editor = {Eduard Hovy and Klaus Zechner and Liang Zhou}, title = {Proceedings of the Analyzing Conversations in Text and Speech}, month = {June}, year = {2006}, address = {New York City, New York}, publisher = {Association for Computational Linguistics}, url = {http://www.aclweb.org/anthology/W/W06/W06-18} } @InProceedings{murray-taboada-renals: 2006:ACTS, author = {Murray, Gabriel and Taboada, Maite and Renals, Steve}, title = {Prosodic Correlates of Rhetorical Relations}, booktitle = {Proceedings of the Analyzing ...
Research Interests:
Respondents in standardized surveys tend to assume that their definitions of everyday terms such as "bedroom" or "job" must match those of the survey designers, even though we know that they often differ substantially.... more
Respondents in standardized surveys tend to assume that their definitions of everyday terms such as "bedroom" or "job" must match those of the survey designers, even though we know that they often differ substantially. Even when they are offered clarification, they often do not request it because they do not think that it is needed. In our earlier studies of telephone interviews, we found that respondents answer more accurately when they receive clarification about question meaning (Schober & Conrad 1997, Conrad & Schober 2000) This is also true for web survey interfaces (Schober & Conrad, 1998) and applies whether the respondent requests the clarification (by clicking to get official definitions) or the system offers unsolicited clarification.
Research Interests:
Research Interests:
Research Interests:
This paper describes a multimodal application architecture that facilitates rapid prototyping of multimodal interfaces with flexible input and adaptive output. Our testbed application MATCH (Multimodal Access To City Help) provides a... more
This paper describes a multimodal application architecture that facilitates rapid prototyping of multimodal interfaces with flexible input and adaptive output. Our testbed application MATCH (Multimodal Access To City Help) provides a mobile multimodal speech-pen interface to restaurant and subway information for New York City. Finite-state methods for multimodal language understanding are employed to enable users to interact using pen, speech, or dynamic combinations of the two. A speech-act based multimodal dialogue manager provides support for mixed-initiative multimodal dialogue. Multimodal generation and text planning components enable the system to respond using synchronized multimodal presentations that combine dynamic graphics with text-to-speech and are tailored to the user's individual preferences. The development of this architecture is driven by ongoing scenario-based evaluation and data collection with users
Research Interests:
Interfaces for mobile information access need to allow users flexibility in their choice of modes and interaction style in accordance with their preferences, the task at hand, and their physical and social environment. This paper... more
Interfaces for mobile information access need to allow users flexibility in their choice of modes and interaction style in accordance with their preferences, the task at hand, and their physical and social environment. This paper describes the approach to multimodal language processing in MATCH (Multimodal Access To City Help), a mobile multimodal speech-pen interface to restaurant and subway information for New York City. Finite-state methods for multimodal integration and understanding enable users to interact using pen, speech, or dynamic combinations of the two, and a speech-act based multimodal dialogue manager enables mixedinitiative multimodal dialogue.
Research Interests:
Computational models of dialog context have primarily focused on unimodal spoken dialog or text where the primary source of context is the (verbal) dialog itself. As we move from spoken interaction to multimodal interaction on mobile... more
Computational models of dialog context have primarily focused on unimodal spoken dialog or text where the primary source of context is the (verbal) dialog itself. As we move from spoken interaction to multimodal interaction on mobile platforms supporting a combination of spoken dialog with graphical interaction, touchscreen input, and access to the spatial context of the device itself through GPS, accelerometers, and cameras, there is a growing need for more sophisticated context models that capture the interplay of different modalities in shaping the dialog context, and determining the influence these modalities have on semantic interpretation and dialog flow. In this paper, we report on an empirical investigation of this issue in the context of a deployed multimodal system for local search and explore a range of different candidate techniques for multimodal context modeling.
Research Interests:
We present a system for extracting useful information from multi-party meetings and presenting the results to users via a browser. Users can view automatically extracted discussion topics and action items, initially seeing high-level... more
We present a system for extracting useful information from multi-party meetings and presenting the results to users via a browser. Users can view automatically extracted discussion topics and action items, initially seeing high-level descriptions, but with the ability to click through to meeting audio and video. Users can also add value by defining and searching for new topics and editing, correcting, deleting, or confirming action items. These feedback actions are used as implicit supervision by the understanding agents, retraining classifier models for improved or user-tailored performance.
We describe a process for automatically detecting decision-making sub-dialogues in transcripts of multi-party, human-human meetings. Extending our previous work on action item identification, we propose a structured approach that takes... more
We describe a process for automatically detecting decision-making sub-dialogues in transcripts of multi-party, human-human meetings. Extending our previous work on action item identification, we propose a structured approach that takes into account the different roles utterances play in the decisionmaking process. We show that this structured approach outperforms the accuracy achieved by existing decision detection systems based on flat annotations, while enabling the extraction of more fine-grained information that can be used for summarization and reporting. 1
Research Interests:
Meeting participants can experience cognitive overload when they need both to verbally contribute to ongoing discussion while simultaneously creating notes to promote later recall of decisions made during the meeting. We designed two... more
Meeting participants can experience cognitive overload when they need both to verbally contribute to ongoing discussion while simultaneously creating notes to promote later recall of decisions made during the meeting. We designed two novel cuing tools to reduce the cognitive load associated with note-taking, thus improving verbal contributions in meetings. The tools combine real-time automatic speech recognition (ASR) with lightweight annotation to transform note-taking into a low overhead markup process. To create lightweight notes, users do not generate the notes ’ content themselves. Instead they simply highlight important phrases in a real-time ASR transcript (Highlighter tool), or press a button to indicate when they heard something important (Hotspots tool). We evaluated these markup tools against a traditional pen-and-paper baseline with 26 users. Hotspots was highly successful: compared with handwritten notes, it increased participants’ conversational contributions and reduc...
Research Interests:
We address the problem of identifying words and phrases that accurately capture, or contribute to, the semantic gist of deci-sions made in multi-party human-human meetings. We first de-scribe our approach to modelling decision discussions... more
We address the problem of identifying words and phrases that accurately capture, or contribute to, the semantic gist of deci-sions made in multi-party human-human meetings. We first de-scribe our approach to modelling decision discussions in spo-ken meetings and then compare two approaches to extracting information from these discussions. The first one uses an open-domain semantic parser that identifies candidate phrases for decision summaries and then employs machine learning tech-niques to select from those candidate phrases. The second one uses categorical and sequential classifiers that exploit simple syntactic and semantic features to identify words and phrases relevant for decision summarization. Index Terms: phrase extraction, human-human meetings, deci-sion detection and summarization
Research Interests:
In face-to-face meetings, assigning and agreeing to carry out future actions is a fre-quent subject of conversation. Work thus far on identifying these action item discussions has focused on extracting them from entire transcripts of... more
In face-to-face meetings, assigning and agreeing to carry out future actions is a fre-quent subject of conversation. Work thus far on identifying these action item discussions has focused on extracting them from entire transcripts of meetings. Here we investi-gate a human-initiative targeting approach by simulating a scenario where meeting par-ticipants provide low-load input (pressing a button during the dialogue) to indicate that an action item is being discussed. We compare the performance of categorical and sequential machine learning methods and their robustness when the point of user input varies. We also consider automatic summa-rization of action items in cases where indi-vidual utterances contain more than one type of relevant information. 1
Research Interests:
Abstract. This paper presents the results of initial investigation and experiments into automatic action item detection from transcripts of multi-party human-human meetings. We start from the flat action item annotations of [1], and show... more
Abstract. This paper presents the results of initial investigation and experiments into automatic action item detection from transcripts of multi-party human-human meetings. We start from the flat action item annotations of [1], and show that automatic classification performance is limited. We then describe a new hierarchical annotation schema based on the roles utterances play in the action item assignment process, and propose a corresponding approach to automatic detection that promises improved classification accuracy while also enabling the extraction of useful information for summarization and reporting. 1
Research Interests:
Integrating Multiple Representations of Spatial Knowledge for Mapping, Navigation, and Communication / 1 Patrick Beeson, Matt MacMahon, Joseph Modayil, Aniket Murarka, Benjamin Kuipers, and Brian Stankiewicz ... Bringing the User Back... more
Integrating Multiple Representations of Spatial Knowledge for Mapping, Navigation, and Communication / 1 Patrick Beeson, Matt MacMahon, Joseph Modayil, Aniket Murarka, Benjamin Kuipers, and Brian Stankiewicz ... Bringing the User Back into Scheduling: Two Case Studies of Interaction with Intelligent Scheduling Assistants / 10 Pauline Berry, Bart Peintner, and Neil Yorke-Smith ... Task Learning by Instruction: Benefits and Challenges for Intelligent Interactive Systems / 12 Jim Blythe, Prateek Tandon, and Mandar Tillu ... Supporting Interaction in the ...
@Book{NAACLHLTDemos:2007, editor = {Bob Carpenter and Amanda Stent and Jason D. Williams}, title = {Proceedings of Human Language Technologies: The Annual Conference of the North American Chapter of the Association for Computational... more
@Book{NAACLHLTDemos:2007, editor = {Bob Carpenter and Amanda Stent and Jason D. Williams}, title = {Proceedings of Human Language Technologies: The Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT)}, month = {April}, year = {2007}, address = {Rochester, New York, USA}, publisher = {Association for Computational Linguistics}, url = {http://www.aclweb.org/anthology/N/N07/N07-4} } @InProceedings{allen-EtAl:2007:NAACLHLTDemos, author = {Allen, James and Chambers ...
We explore the plausibility of using automated spoken dialog systems (SDS) for administering survey interviews. Because the goals of a survey dialog system differ from more traditional information-seeking and transactional applications,... more
We explore the plausibility of using automated spoken dialog systems (SDS) for administering survey interviews. Because the goals of a survey dialog system differ from more traditional information-seeking and transactional applications, different measures of task accuracy and success may be warranted. We report a large-scale experimental evaluation of an SDS that administered survey interviews with questions drawn from government and social scientific surveys. We compare two dialog confirmation strategies: (1) a traditional strategy of explicit confirmation on low-confidence recognition; and (2) no confirmation. With explicit confirmation, the small percentage of residual errors had little to no impact on survey data measurement. Even without confirmation, while there are significantly more errors, impact on the substantive conclusions of the survey is still very limited.