Learning Deep ReLU Networks Is Fixed-Parameter Tractable
2021 IEEE 62nd Annual Symposium on Foundations of Computer Science (FOCS)(2021)
摘要
We consider the problem of learning an unknown ReLU network with respect to Gaussian inputs and obtain the first nontrivial results for networks of depth more than two. We give an algorithm whose running time is a fixed polynomial in the ambient dimension and some (exponentially large) function of only the network's parameters. Our results provably cannot be obtained using gradient-based methods a...
更多查看译文
关键词
Geometry,Computer science,Computational modeling,Neural networks,Lattices,Filtering algorithms,Approximation algorithms
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要