Our paper "DOGE-Train: Discrete Optimization on GPU with End-to-end Training" accepted at AAAI 2024.
arXiv: https://arxiv.org/abs/2205.11638