Matthew Marge

Learn More
We investigate whether Amazon's Mechanical Turk (MTurk) service can be used as a reliable method for transcription of spoken language data. Utterances with varying speaker demographics (native and non-native English, male and female) were posted on the MTurk marketplace together with standard transcription guidelines. Transcriptions were compared against(More)
Recent years have seen increasing interest in automatic metrics for the evaluation of generation systems. When a system can generate syntactic variation, automatic evaluation becomes more difficult. In this paper, we compare the performance of several automatic evaluation metrics using a corpus of automatically generated paraphrases. We show that these(More)
How should a robot represent and reason about spatial information when it needs to collaborate effectively with a human? The form of spatial representation that is useful for robot navigation may not be useful in higher-level reasoning or working with humans as a team member. To explore this question, we have extended previous work on how children and(More)
We describe an approach to improving the naturalness of a social dialogue system, Talkie, by adding disfluencies and other content-independent enhancements to synthesized conversations. We investigated whether listeners perceive conversations with these improvements as natural (i.e., human-like) as human-human conversations. We also assessed their ability(More)
Spoken language interaction between humans and robots in natural environments will necessarily involve communication about space and distance. The current study examines people’s close-range route instructions for robots and how the presentation format (schematic, virtual or natural) and the complexity of the route affect the content of instructions. We(More)
We describe an empirical study that crowdsourced human-authored recovery strategies for various problems encountered in physically situated dialogue. The purpose was to investigate the strategies that people use in response to requests that are referentially ambiguous or impossible to execute. Results suggest a general preference for including specific(More)
Human-robot interaction could be improved by designing robots that engage in adaptive dialogue with users. An adaptive robot could estimate the information needs of individuals and change its dialogue to suit these needs. We test the value of adaptive robot dialogue by experimentally comparing the effects of adaptation versus no adaptation on information(More)
Creating a human-robot interface is a daunting experience. Capabilities and functionalities of the interface are dependent on the robustness of many different sensor and input modalities. For example, object recognition poses problems for state-of-the-art vision systems. Speech recognition in noisy environments remains problematic for acoustic systems.(More)
Due to its complexity, meeting speech provides a challenge for both transcription and annotation. While Amazon’s Mechanical Turk (MTurk) has been shown to produce good results for some types of speech, its suitability for transcription and annotation of spontaneous speech has not been established. We find that MTurk can be used to produce highquality(More)
Our overall program objective is to provide more natural ways for soldiers to interact and communicate with robots, much like how soldiers communicate with other soldiers today. We describe how the Wizard-of-Oz (WOz) method can be applied to multimodal human-robot dialogue in a collaborative exploration task. While the WOz method can help design robot(More)