Recently I noticed that Google and Apple have really high quality speech-recognition services. I was wondering about the state-of-the-art methods and techniques they are/might be using to achieve such quality. I already know that Hidden Markov Models can be used for speech recognition, but I was wondering if it's the used technique now. Because as I know HMM are old.
It would be great if you can link to some important recent papers that could include/descripe the used new methods.