Towards Provable and Scalable Training of Quantized Neural Networks with Ising Optimization
This work introduces an exact Quadratic Constrained Binary Optimization (QCBO) framework with provable guarantees and describes the stratified topology of network zero-loss level sets, demonstrating resilience against low-bit representational collapse.
Wenxin Li, Chuan Wang, Hongdong Zhu et al.
· 5 citations