Absolutely, but this is not necessarily related to speech recognition. For instance, I’m currently working on “engagement” detection using the voice, but my system just analyzes how engagement “sounds”, without any understanding of what has been said.