A5006905045- Intelligent Tutoring Systems and Adaptive Learning
- Topic Modeling
- Innovative Teaching and Learning Methods
- Natural Language Processing Techniques
- Speech and dialogue systems
- Online Learning and Analytics
- Text Readability and Simplification
- Teaching and Learning Programming
- AI-based Problem Solving and Planning
- Advanced Text Analysis Techniques
- Visual and Cognitive Learning Processes
- Language, Metaphor, and Cognition
- Reading and Literacy Development
- Semantic Web and Ontologies
- Music and Audio Processing
- Education and Critical Thinking Development
- Educational and Psychological Assessments
- Mind wandering and attention
- Educational Methods and Media Use
- Online and Blended Learning
- Speech Recognition and Synthesis
- Educational Assessment and Pedagogy
- Educational Games and Gamification
- Child Development and Digital Technology
- Statistics Education and Methodologies
University of Memphis
2015-2024
FedEx (United States)
2007
Robotics Research (United States)
2005
AutoTutor simulates a human tutor by holding conversation with the learner in natural language. The dialogue is augmented an animated conversational agent and three-dimensional (3-D) interactive simulations order to enhance learner's engagement depth of learning. Grounded constructivist learning theories tutoring research, achieves gains approximately 0.8 sigma (nearly one letter grade), depending on measure comparison condition. computational architecture system uses .NET framework has...
It is often assumed that engaging in a one-on-one dialogue with tutor more effective than listening to lecture or reading text. Although earlier experiments have not always supported this hypothesis, may be due part allowing the tutors cover different content noninteractive instruction. In 7 experiments, we tested interaction hypothesis under constraint (a) all students covered same during instruction, (b) task domain was qualitative physics, (c) instruction natural language as opposed...
Analyzing the quality of classroom talk is central to educational research and improvement efforts. In particular, presence authentic teacher questions, where answers are not predetermined by teacher, helps constitute serves as a marker productive discourse. Further, questions can be cultivated improve teaching effectiveness consequently student achievement. Unfortunately, current methods measure question authenticity do scale because they rely on human observations or coding To address this...
The Office of Naval Research (ONR) organized a STEM Challenge initiative to explore how intelligent tutoring systems (ITSs) can be developed in reasonable amount time help students learn topics. This competitive sponsored four teams that separately covered topics mathematics, electronics, and dynamical systems. After the shared their progress at conclusion an 18-month period, ONR decided fund joint applied project Navy integrated those on subject matter electronic circuits. University...
In this paper we present a question generation approach suitable for tutorial dialogues. The is based on previous psychological theories that hypothesize questions are generated from knowledge representation modeled as concept map. Our model automatically extracts maps textbook and uses them to generate questions. purpose of the study evaluate pedagogically-appropriate at varying levels specificity across one or more sentences. evaluation metrics include scales Question Generation Shared...
We present a method to automatically detect collaborative patterns of student and tutor dialogue moves. The identifies significant two-step excitatory transitions between moves, integrates the into directed graph representation, generates tests data-driven hypotheses from graph. was applied large corpus student-tutor moves expert tutoring sessions. An examination subset consisting lectures revealed consistent with information-transmission, information-elicitation, off topic-conversation,...
Current learning technologies have no direct way to assess students' mental effort: are they in deep thought, struggling overcome an impasse, or zoned out? To address this challenge, we propose the use of EEG-based cognitive load detectors during learning. Despite its potential, EEG has not yet been utilized as a optimize instructional strategies. We take initial step towards goal by assessing how experimentally manipulated (easy and difficult) sections intelligent tutoring system (ITS)...
We investigate automatic analysis of teachers' instructional strategies from audio recordings collected in live classrooms. a data set teacher and human-coded activities (e.g., lecture, question answer, group work) 76 middle school literature, language arts, civics classes eleven teachers across six schools. automatically segment to analyze speech vs. rest patterns, generate transcripts the extract natural features, compute low-level acoustic features. train supervised machine learning...
We focus on data collection designs for the automated analysis of teacher-student interactions in live classrooms with goal identifying instructional activities (e.g., lecturing, discussion) and assessing quality dialogic instruction questions). Our were motivated by multiple technical requirements constraints. Most importantly, teachers could be individually micfied but their audio needed to excellent automatic speech recognition (ASR) spoken utterance segmentation. Individual students not...
This study investigated learning outcomes and user perceptions from interactions with a hybrid intelligent tutoring system created by combining the AutoTutor conversational Assessment Learning in Knowledge Spaces (ALEKS) adaptive for mathematics. (ITS) uses service-oriented architecture to combine these two web-based systems. Self-explanation dialogs were used talk students through step-by-step worked examples algebra problems. These presented an isomorphic problem preceding that student...
We investigate automatic detection of teacher questions from audio recordings collected in live classrooms with the goal providing automated feedback to teachers. Using a dataset 11 teachers across 37 class sessions, we automatically segment into individual utterances and code each as containing question or not. train supervised machine learning models detect human-coded using high-level linguistic features extracted speech recognition (ASR) transcripts, acoustic prosodic recordings, well...
Abstract Since its global emergence in 2020, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused multiple epidemics the United States. When medical treatments for virus were still emerging and a vaccine was not yet available, state local governments sought to limit spread by enacting various social-distancing interventions, such as school closures lockdowns; however, effectiveness of these interventions unknown. We applied an established, semimechanistic Bayesian...
Researchers in the cognitive and affective sciences investigate how thoughts feelings are reflected bodily response systems including peripheral physiology, facial features, body movements. One specific question along this line of research is cognition affect manifested dynamics general Progress area can be accelerated by inexpensive, non-intrusive, portable, scalable, easy to calibrate movement tracking systems. Towards end, paper presents validates Motion Tracker, a simple yet effective...
This paper describes classification of typed student utterances within AutoTutor, an intelligent tutoring system. Utterances are classified to one 18 categories, including 16 question categories. The classifier presented uses part speech tagging, cascaded finite state transducers, and simple disambiguation rules. Shallow NLP is well suited the task: session log file analysis reveals significant eleven frozen expressions, assertions.
We investigate multi-sensor modeling of teachers' instructional segments (e.g., lecture, group work) from audio recordings collected in 56 classes eight teachers across five middle schools. Our approach fuses two sensors: a unidirectional microphone for teacher and pressure zone general classroom audio. segment analyze the streams with respect to discourse timing, linguistic, paralinguistic features. train supervised classifiers identify that collectively comprised majority data, achieving...