Towards Better Automatic Speech Recognition Systems For Children