Abstract ― This project aims to develop an automatic self-observation software for teachers, allowing them to gain rapid insight on the quality of their teaching in Mathematics. This software records the class audio from external microphones in a standard notebook every time the teacher actives it during a lecture. After the class, the software processes the teacher’s and students’ speech, detecting the contents of the curriculum taught and pedagogical strategies performed. The teacher receives a personal and confidential report about the development of the lectured class, if it went as planned and matched the sequence of classes in the unit. The system can also suggest increasing or decreasing some teaching strategies based in the past recorded lessons.
We expect teachers could become more aware of their practices using this system, allowing them to improve their pedagogical strategies and thereby also contribute to improved students’ learning results. Finally, if students use online homework or assessment systems, the system can provide an analysis of the impact of the teaching strategies used on students’ attainment.
Keywords: Artificial Intelligence, semantic analysis, speech to text, automatic classifiers, education, ICT, learning results, pedagogical practices
Funding source: CONICYT - FONDEF under grantt D11I1099