简介: 这篇论文主要介绍了一个新的模型DIFUSCO : 这个model 文中实现的可以解决的问题是TSP and MIS ,model 的主要实现方法是先根据diffusion model 进行向前传播-进行加噪声,然后进行先后传播进行去噪声。去噪声用的是12层的Anisotropic Graph Neural Networks,Graph节点数量为256.(网络是监督学习的)
该模型的效果非常的好:
code is available at here
下面是详细介绍:
Introduction 里面介绍了三种求解NPC问题的方法 - 自回归,非自回归,启发式方法。
分别介绍了三种方法的缺点:
自回归: Those methods typically suffer from the costly computation in their sequential decoding parts and hence are difficult to scale up to large problems
非自回归:unavoidably limits the capability of those methods to capture the multimodal nature of the problems, for example, when multiple optimal solutions exists for the same graph.
启发式方法 : have also suffered from the difficulty in scaling up and the latency in inference.partly due to the sparse rewards and sample efficiency issues at RL learning.
然后介绍了DIFUSCO model 对于三者的问题都有改善:
Firstly, DIFUSCO can perform inference on all variables in parallel with a few (≪ N ) denoising steps (Sec. 3.3), avoiding the sequential generation problem of autoregressive constructive solvers.
Secondly, DIFUSCO can model a multimodal distribution via iterative refinements, which alleviates the expressiveness limitation of previous non-autoregressive constructive models.
Last but not least, DIFUSCO is trained in an efficient and stable manner with supervised denoising (Sec. 3.2), which solves the training scalability issue of RL-based improvement heuristics methods.
文章主要采用了graph-based diffusion (可以用“constructive heuristics solvers” 和 “improvement heuristics solvers”),model 的实现有两种 : continuous diffusion with Gaussian noise and discrete diffusion with Bernoulli noise , 其中 discrete 的要好一点 。
相关工作:主要是 Diffusion Models for Discrete Data,包含了(Typical diffusion models ,Another line of work studies diffusion models for discrete data ,Discrete diffusion models)
model 设计部分:
foreard:
reverse:
加快inference的方法:
1.One way to speed up the inference of denoising diffusion models is to reduce the number of steps
2.取子集
ps:cosine schedule更好一点
然后介绍了一下网络结构
然后是Decoding Strategies
generate instance 的热力图 , 采用了Bernoulli sampling , 会discards the comparative information, 为了保留信息:
TSP Decoding :
1. Greedy decoding + 2-opt + all the edges are ranked by (Aij + Aji)/∥ci − cj∥
2. guided by the heatmap scores + k-opt + guided by the heatmap scores(Aij)