For the final project, I work on reading one classical paper in speech recognition: `Machine Learning Paradigms for Speech Recognition: An Overview`, which gives me a clearer big picture of what researches have been done before. Some known solutions include: generative learning, discriminative learning, supervised learning, unsupervised learning, semi-supervised learning, active learning, transfer learning, and deep learning.
https://hogansung.com/wp-content/uploads/2021/02/undergrad-Junior2013-DigitalSpeechProcessing_Last-report.pdf