XOR-CD: Linearly Convergent Constrained Structure Generation
XOR-CD: Linearly Convergent Constrained Structure Generation
复制标题
DOI：
--
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
Fan Ding;Jianzhu Ma;Jinbo Xu;Yexiang Xue
中科院分区：
文献类型：
--
作者：
Fan Ding;Jianzhu Ma;Jinbo Xu;Yexiang Xue
我们提出了XOR-对比发散学习（XOR-CD），这是一种用于约束结构生成的可证明方法，对于最先进的神经网络和约束推理方法来说仍然很困难。XOR-CD利用XOR采样从CD学习中的模型分布生成样本，并保证生成有效的结构。此外，XOR-CD在学习指数族模型时，在一个消失常数内具有向似然函数的全局最大值的线性收敛速度。XOR-CD实现的约束满足也提高了其学习性能。我们在数据驱动的实验设计、调度路径生成和基于序列的蛋白质同源性检测方面的真实实验证明了XOR-CD在生成有效结构以及捕获训练集中的诱导偏差方面比基线方法具有上级性能。
We propose XOR-Contrastive Divergence learning (XOR-CD), a provable approach for constrained structure generation, which remains dif-ficult for state-of-the-art neural network and constraint reasoning approaches. XOR-CD harnesses XOR-Sampling to generate samples from the model distribution in CD learning and is guaranteed to generate valid structures. In addition, XOR-CD has a linear convergence rate towards the global maximum of the likelihood function within a vanishing constant in learning exponential family models. Constraint satisfaction enabled by XOR-CD also boosts its learning performance. Our real-world experiments on data-driven experimental design, dispatching route generation, and sequence-based protein homology detection demonstrate the superior performance of XOR-CD compared to baseline approaches in generating valid structures as well as capturing the inductive bias in the training set.