Bolanos D.,Boulder Language Technologies |
Cole R.A.,Boulder Language Technologies |
Ward W.H.,Boulder Language Technologies |
Ward W.H.,University of Colorado at Boulder |
And 3 more authors.
Speech Communication | Year: 2013
We investigated the automatic assessment of expressive children's oral reading of grade level text passages using a standardized rubric. After a careful review of the reading literature and a close examination of the rubric, we designed a novel set of prosodic and lexical features to characterize fluent expressive reading. A number of complementary sources of information were used to design the features, each of them motivated by research on different components of reading fluency. Features are connected to the child's reading rate, to the presence and number of pauses, filled-pauses and word-repetitions, the correlation between punctuation marks and pauses, the length of word groupings, syllable stress and duration and the location of pitch peaks and contours. The proposed features were evaluated on a corpus of 783 one-minute reading sessions from 313 students reading grade-leveled passages without assistance (cold unassisted reading). Experimental results show that the proposed lexical and prosodic features provide complementary information and are able to capture the characteristics of expressive reading. The results showed that on both the 2-point and the 4-point expressiveness scales, computer-generated ratings of expressiveness agreed with human raters better than the human raters agreed with each other. The results of the study suggest that automatic assessment of expressive oral reading can be combined with automatic measures of word accuracy and reading rate to produce an accurate multidimensional estimate of children's oral reading ability. © 2012 Elsevier B.V. All rights reserved.
Ward W.,Boulder Language Technologies |
Bolanos D.,Boulder Language Technologies |
Cole R.,Boulder Language Technologies
13th Annual Conference of the International Speech Communication Association 2012, INTERSPEECH 2012 | Year: 2012
My Science Tutor (MyST) is an intelligent tutoring system designed to improve science learning by elementary school students through conversational dialogs with a virtual science tutor in an interactive multimedia environment. Marni, a lifelike 3-D character, attempts to elicit self-expression from students, process their spoken explanations to assess understanding, and scaffold learning by asking open-ended questions accompanied by illustrations, animations or interactive simulations. MyST uses automatic speech recognition, natural language processing and dialog modeling technologies to interpret student responses and manage the dialog.
Bolan-Os D.,Boulder Language Technologies |
Cole R.A.,Boulder Language Technologies |
Ward W.,Boulder Language Technologies |
Borts E.,Boulder Language Technologies |
Svirsky E.,Boulder Language Technologies
ACM Transactions on Speech and Language Processing | Year: 2011
We present initial results of FLORA, an accessible computer program that uses speech recognition to provide an accurate measure of children's oral reading ability. FLORA presents grade-level text passages to children, who read the passages out loud, and computes the number of words correct per minute (WCPM), a standard measure of oral reading fluency. We describe the main components of the FLORA program, including the system architecture and the speech recognition subsystems. We compare results of FLORA to human scoring on 783 recordings of grade level text passages read aloud by first through fourth grade students in classroom settings. On average, FLORA WCPM scores were within 3 to 4 words of human scorers across students in different grade levels and schools. © 2011 ACM.
Wang C.,Boulder Language Technologies |
Xue N.,Boulder Language Technologies |
Pradhan S.,Boulder Language Technologies
ACL-IJCNLP 2015 - 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing of the Asian Federation of Natural Language Processing, Proceedings of the Conference | Year: 2015
We report improved AMR parsing results by adding a new action to a transitionbased AMR parser to infer abstract concepts and by incorporating richer features produced by auxiliary analyzers such as a semantic role labeler and a coreference resolver. We report final AMR parsing results that show an improvement of 7% absolute in F1 score over the best previously reported result. Our parser is available at: https://github.com/Juicechuan/AMRParsing. © 2015 Association for Computational Linguistics.
Thompson C.K.,Northwestern University |
Choy J.W.J.,Northwestern University |
Holland A.,University of Arizona |
Cole R.,Boulder Language Technologies
Aphasiology | Year: 2010
Background: Treatment of underlying forms (TUF) is a linguistically based treatment for improving agrammatic sentence deficits. It enjoys a substantial database attesting to its efficacy for improving both sentence comprehension and production in agrammatic aphasia. However, TUF requires considerable linguistic background to administer, and administration time can exceed the number of treatment sessions allotted in toto for reimbursement by third-party payors in the United States. Thus Sentactics®, an interactive computer system that enables delivery of TUF by a virtual clinician, was developed. Aims: This study tested the effects of Sentactics® on the acquisition and generalised production and comprehension of complex sentences. Additionally, a direct comparison of the results of computer-delivered Sentactics® and clinician-delivered TUF was undertaken. Methods & Procedures: A total of 12 agrammatic aphasic speakers participated in the study, with 6 receiving Sentactics® and 6 serving as experimental controls who received no treatment. All participants were administered pre- and post-treatment sentence comprehension and production tests and other measures to evaluate the effects of Sentactics®. Performance of the Sentactics® group was also compared to eight agrammatic patients who previously received clinician-delivered TUF treatment identical to that delivered via Sentactics®, but with a human clinician. Outcomes & Results: Sentactics® significantly improved all six aphasic speakers' ability to comprehend and produce both trained and untrained, linguistically related, complex sentences as compared to six agrammatic control participants who did not receive Sentactics®. In addition, comparing the results of Sentactics® to clinician-delivered TUF revealed no significant differences between approaches with regard to acquisition or generalisation patterns. Conclusions: These data provide further support for the efficacy of TUF and demonstrate the viability of computer-delivered therapies in the field of aphasia treatment. © 2010 Psychology Press.
Agency: NSF | Branch: Standard Grant | Program: | Phase: INFORMATION TECHNOLOGY RESEARC | Award Amount: 282.39K | Year: 2011
This collaborative research investigates a new class of dialog-based, home robotic healthcare assistants to facilitate a new level of in-home, real-time care to elderly and depressed patients, providing lower total costs and higher quality of life. An emotive, physical avatar, called a companionbot, which possesses the ability to engage humans in a way that is unobtrusive and suspends disbelief will be built in this project. The companionbot will be an integration of human language technology, vision, other sensory processing and emotive robotic technology to proactively recognize and dialog with isolated and elderly patients suffering from depression. The companionbot will utilize proactive or companionable dialog based on the context with users suffering from depression. This will require the first multimodal integration of a user model, environment model, and temporal processing with spoken dialog understanding and generation to produce dynamic dialog and emotive interaction, beyond the traditional scripted dialog and emotion. Object recognition, facial expression recognition, and human activity recognition will augment natural language processing to provide current and historical context important to dynamic dialog.
A team of skilled researchers, assembled from the University of Colorado Boulder, University of Denver, CU Anschutz Medical Campus, and Boulder Language Technologies, will work together to achieve the project goals. The investigators will use the companionbots as a tool to run clinical trials to monitor and dialog with their partners to detect signs of physical and emotional deterioration. The companionbots can then notify remote caregivers, as necessary, provide warnings, reminders, life coaching and therapeutic dialog, extending independence and quality of life, and even saving lives. The other benefits of such a system include continuous, annotated data to improve doctor-patient interaction and analysis, real-time monitoring of mental state for remote healthcare providers and, ultimately, real-time intervention as part of a comprehensive treatment strategy.
In addition, this research will promote both STEM practice and research education at the graduate and the undergraduate levels of the affiliated institutions. The companionbots are ideal for teaching the next generation of engineers and scientists in critical emerging technologies, as they permit either a deep focus on specific topics or an interdisciplinary perspective while providing a simple high-level interface to manage everything else. Furthermore, the project will develop related educational material to support others and will provide public outreach to K-12 classes in the area.
Agency: NSF | Branch: Standard Grant | Program: | Phase: PROGRAM EVALUATION | Award Amount: 249.57K | Year: 2012
This exploratory proposal addresses a significant goal in the New Science Framework and one that was previously highlighted in the NRC report: Taking Science to School: Learning and Teaching Science in Grades K-8. The goal of the project is the teaching and learning of the norms of scientific argument, explanation, and the evaluation of evidence for productive participation in the discourses of science, starting in the early grades. The investigators assert that even though some students do not read well, there may be significant scientific understanding which will go unrecognized unless alternative methods of evaluation are developed such as oral assessments. Oral assessments by pedagogical agents which can elicit explanations from students and engage them in serious dialog are an alternate to oral assessments by humans.
The project has two goals: 1) assess the benefits of incorporating spoken prompts into written assessments, and 2) investigate spoken dialogs between a student and a virtual or human teacher testing if they can produce detailed and accurate assessments of science understanding as well as the ability to verbalize complete and accurate science explanations. The site for this project is the Boulder Valley School Districts low and mid-performing schools and its Summer Science Camp program for language-minority students. FOSS, the Full Option Science System, a curriculum used by over 100,000 teachers and 2 million students, and the FOSS Summative ASK assessments will be used for the curriculum and to measure learning. This effort builds on an earlier successful development by the same team of My Science Tutor (MyST), an intelligent tutoring system to improve science learning by third, fourth, and fifth grade students through spoken dialogs with Marni, a virtual science tutor in multimedia environments. Elementary science will be the content area for this project as well. Primary organizations involved are Boulder Language Technologies, the Boulder Valley School District, and the University of Colorado. It is expected that these exploratory studies will show that alternative forms of assessment are feasible and should be used more widely to raise the achievement level of both language-minority students and elementary students in general. The evaluation will include an external expert review and regular critical review of the project?s methods and progress, analysis procedures, and interpretation of data into findings.
The proposed studies will expand the options of assessment conditions available for students in early grades, particularly those who are weak readers and writers and English language learners. The project will expand the pool of students who can demonstrate their knowledge and understanding of science, addressing a critical need to expand the pool of students engaged in science learning.
Agency: National Science Foundation | Branch: | Program: SBIR | Phase: Phase I | Award Amount: 150.00K | Year: 2014
This SBIR Phase I project's broader/commercial impact is the potential of the intelligent tutoring system to increase young learners' science motivation and achievement and to improve their fluent reading and comprehension skills. The innovation thus addresses a critical national need to improve students' science learning and reading proficiency in our nation's schools, since national assessments of educational progress indicate that the majority of fourth, eighth and twelfth grade students are not proficient in either science or reading on standardized tests. The innovation is thus designed to provide teachers with an accessible, affordable and highly effective learning tool that will improve their students' engagement, motivation and learning in both science and reading. The projects' commercialization plan is designed to support widespread distribution of the innovation to schools in the U.S. and globally through existing channels, thus increasing the potential for broad societal impact. The project aims to develop and demonstrate the feasibility of an intelligent tutoring system that will help children learn to reason about science and comprehend science texts through conversational interactions with a virtual tutor. Successful outcomes of the project will lead to new insights and contribute to scientific knowledge about how advanced human language technologies, character animation technologies and computer vision technologies can be combined and integrated into intelligent tutoring systems to optimize young learners' engagement, motivation, self-efficacy and learning. The project is based on theory and scientific research on how children learn through social interaction in multimedia environments, and prior research by the project team that improved children's motivation to learn science and their science achievement through conversations with a virtual science tutor. Successful outcomes of the SBIR Phase I project will extend this prior research and advance scientific knowledge by demonstrating the feasibility and promise of an intelligent tutoring system that tracks and interprets students? eye movements to improve science learning, oral reading fluency and reading comprehension.
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 150.00K | Year: 2014
This SBIR Phase I projects broader/commercial impact is the potential of the intelligent tutoring system to increase young learners science motivation and achievement and to improve their fluent reading and comprehension skills. The innovation thus addresses a critical national need to improve students science learning and reading proficiency in our nations schools, since national assessments of educational progress indicate that the majority of fourth, eighth and twelfth grade students are not proficient in either science or reading on standardized tests. The innovation is thus designed to provide teachers with an accessible, affordable and highly effective learning tool that will improve their students engagement, motivation and learning in both science and reading. The projects commercialization plan is designed to support widespread distribution of the innovation to schools in the U.S. and globally through existing channels, thus increasing the potential for broad societal impact.
The project aims to develop and demonstrate the feasibility of an intelligent tutoring system that will help children learn to reason about science and comprehend science texts through conversational interactions with a virtual tutor. Successful outcomes of the project will lead to new insights and contribute to scientific knowledge about how advanced human language technologies, character animation technologies and computer vision technologies can be combined and integrated into intelligent tutoring systems to optimize young learners engagement, motivation, self-efficacy and learning. The project is based on theory and scientific research on how children learn through social interaction in multimedia environments, and prior research by the project team that improved childrens motivation to learn science and their science achievement through conversations with a virtual science tutor. Successful outcomes of the SBIR Phase I project will extend this prior research and advance scientific knowledge by demonstrating the feasibility and promise of an intelligent tutoring system that tracks and interprets students? eye movements to improve science learning, oral reading fluency and reading comprehension.
Agency: NSF | Branch: Standard Grant | Program: | Phase: | Award Amount: 100.00K | Year: 2013
In this Cyberlearning: Transforming Education EXP project, the PIs focus on designing classrooms as collaborative workspaces and learning how such learning environments can foster learning well. They are addressing these issues in high school mathematics classrooms. Learners view videos and read textbooks as homework to begin to learn new content and to deepen their understanding of material already covered. The classroom is flipped; rather than the teacher lecturing, the teacher plays the role of mentor and facilitator as learners work in the classroom at making sense of what theyve read or heard, applying what they are learning, and deepening their understanding and capabilities. The hard work of learning is thus done along their peers as collaborators and the teacher available as a mentor. Based on cognitive and socio-cognitive theories of learning, the PIs have designed an ensemble of strategies and technological tools for promoting learning in such an environment. The tools include video for story telling in support of reflection, electronic pen-and-ink, and intelligent-tutoring type systems, but the innovation is in the integration of these tools into an ensemble. Research addresses how such an ensemble of technologies can foster deeply absorbing and effective learning experiences and important dynamics associated with learning when collaborative workspaces are in place in formal classrooms.
Many educators and educational theorists are experimenting with the idea of flipped classrooms, where learners read or view video lectures outside of class and spend classroom time working on problems together or working on projects, in effect, using classroom time for making sense together of what is being learned, applying what is being learned, deepening understanding, and mastering capabilities. While such an approach holds promise for promoting engagement and learning, little systematic research has been done about how, exactly, to design such learning environments to best promote deep and engaged learning. The PIs in this project address that issue, focusing specifically on students learning high school mathematics.