Research Overview
My research work involves the application of machine learning and artificial intelligence (AI) to natural language processing (NLP) and computational linguistics.
I have worked and am currently working on several topics related to NLP and computational linguistics such as natural language dialogue systems (spoken, text-based, multi-modal; including tutorial dialogue systems), reinforcement learning of dialogue system policies, speech recognition and synthesis, speaker recognition, speech disfluency detection and generation, linguistic and behavioral differences between older and young people, verbal indicators of psychological distress, entrainment of natural language dialogue system users, machine learning models of negotiation, incremental spoken language processing, personalization of dialogue systems, differences in interaction styles, and automated scoring of non-native speech for second language learning.
My most recent work has been on natural language dialogue system evaluation, aligning large language models (LLMs) to human values (e.g., safety, transparency, trust) using reinforcement learning from human and artificial intelligence feedback, using natural language dialogue processing to understand and assess team effectiveness, dialogue embeddings (vector representations of dialogue context), speech recognition, and reinforcement learning of policies for intelligent tutoring systems.
List of Research Interests
Natural language dialogue systems and modeling (including spoken language and multi-modal dialogue, and two-party and multi-party dialogue).
Natural language processing (including spoken language processing).
Speech recognition.
Speech synthesis.
Large language model (LLM) alignment and safety using reinforcement learning from human feedback (RLHF) and reinforcement learning from AI feedback (RLAIF).
Machine learning (including reinforcement learning and deep learning).
Human-machine interaction and collaboration.
Human-robot interaction and collaboration.
Multi-agent modeling (including modeling teams of agents and human-agent teaming).
Trustworthy AI agents.
Intelligent tutoring systems.