Accented Speech Recognition: Benchmarking, Pre-training, and Diverse Data - 42Papers