Boolean Variation and Boolean Logic BackPropagation
This addresses the challenge of building efficient deep models for hardware or applications requiring Boolean operations, though it appears incremental as it adapts backpropagation to a new domain.
The paper tackles the problem of training deep models using Boolean numbers and logic instead of real arithmetic, introducing Boolean variation and Boolean logic backpropagation to enable direct training in the Boolean domain without latent weights or gradients.
The notion of variation is introduced for the Boolean set and based on which Boolean logic backpropagation principle is developed. Using this concept, deep models can be built with weights and activations being Boolean numbers and operated with Boolean logic instead of real arithmetic. In particular, Boolean deep models can be trained directly in the Boolean domain without latent weights. No gradient but logic is synthesized and backpropagated through layers.