Over the past decade, robust literature focused on teacher "talk moves" that promote student argumentation has emerged, especially in mathematics education. Teachers and students can use talk moves to construct conversations in which students share their thinking, actively consider the ideas of others, and engage in sustained reasoning. Providing teachers with detailed feedback about the talk moves utilized in their lessons requires considerable human expertise. These highly trained observers must hand-code transcripts of classroom recordings, analyze talk moves and provide one-on-one expert coaching, a time-consuming and expensive process. Our research team developed Talkback - an innovative application to address a significant challenge in education: providing teachers with immediate and actionable feedback on their use of effective classroom discourse strategies. My work is situated in the research and development of a cyberinfrastructure for TalkBack, including deep learning models for Natural Language Processing (NLP) for automated feedback. Starting with a bidirectional long short-term memory (bi-LSTM) network, I explore different state-of-the-art deep learning models, including transformers, to automatically analyze classroom recordings and generate information about classroom discourse strategies with F1 measures up to 78.92%. The TalkMoves dataset used for training and evaluating these models was curated by an interdisciplinary research team and comprised 500+ human-annotated classroom transcripts. The strong performance of both the student and the teacher talk moves models illustrates the reliability and robustness of artificial intelligence algorithms applied to noisy real-world classroom data. TalkBack application serves as an example to support a well-specified theory of learning (accountable talk), addresses a recognized challenge in education (teacher feedback), and has the potential to scale to large classrooms and teachers. The ability to better understand teachers' perceptions and use of the TalkBack application can provide structured professional learning opportunities that promote discourse-rich pedagogy. Results from a mixed-methods study with teachers highlight several emergent themes relating to the perceived utility of TalkBack as an AI-based tool and serving as a platform for research and innovations in NLP and education. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]