Scalable Interconnect Learning in Boolean Networks
This work addresses scalability and efficiency for DBNs on resource-constrained hardware, representing an incremental improvement over earlier learnable-interconnect designs.
The authors tackled the problem of scaling learned Differentiable Boolean Logic Networks (DBNs) to wider layers by introducing a trainable, differentiable interconnect with constant parameter growth, and they reduced model size through two pruning stages: an SAT-based logic equivalence pass and a similarity-based data-driven pass that outperforms a greedy baseline.
Learned Differentiable Boolean Logic Networks (DBNs) already deliver efficient inference on resource-constrained hardware. We extend them with a trainable, differentiable interconnect whose parameter count remains constant as input width grows, allowing DBNs to scale to far wider layers than earlier learnable-interconnect designs while preserving their advantageous accuracy. To further reduce model size, we propose two complementary pruning stages: an SAT-based logic equivalence pass that removes redundant gates without affecting performance, and a similarity-based, data-driven pass that outperforms a magnitude-style greedy baseline and offers a superior compression-accuracy trade-off.