Speech and Language processing (SLP) is essentially a series of sequence-to-sequence learning problems. Conventional SLP systems map input to output sequences through module-based architectures where each module is independently trained. 

Neurotechnology is one of the hottest areas of engineering, and the technological achievements sound miraculous: Paralyzed people have controlled robotic limbs with their brains, while blind people are receiving implants that send signals to their visual centers.

Within scientific research, engineering, and consumer electronics, there is a multitude of new discrete sensor-interfaced devices. Maintaining high accuracy in signal quantization while staying within the strict power-budget of these devices is a very challenging problem.