Timezone: »

Maximum-and-Concatenation Networks
Xingyu Xie · Hao Kong · Jianlong Wu · Wayne Zhang · Guangcan Liu · Zhouchen Lin

Tue Jul 14 07:00 AM -- 07:45 AM & Tue Jul 14 07:00 PM -- 07:45 PM (PDT) @

While successful in many fields, deep neural networks (DNNs) still suffer from some open problems such as bad local minima and unsatisfactory generalization performance. In this work, we propose a novel architecture called Maximum-and-Concatenation Networks (MCN) to try eliminating bad local minima and improving generalization ability as well. Remarkably, we prove that MCN has a very nice property; that is, every local minimum of an (l+1)-layer MCN can be better than, at least as good as, the global minima of the network consisting of its first l layers. In other words, by increasing the network depth, MCN can autonomously improve its local minima's goodness, what is more, it is easy to plug MCN into an existing deep model to make it also have this property. Finally, under mild conditions, we show that MCN can approximate certain continuous function arbitrarily well with high efficiency; that is, the covering number of MCN is much smaller than most existing DNNs such as deep ReLU. Based on this, we further provide a tight generalization bound to guarantee the inference ability of MCN when dealing with testing samples.

Author Information

Xingyu Xie (Peking University)
Hao Kong (Peking University)
Jianlong Wu (Peking University)
Wayne Zhang (SenseTime Research)
Guangcan Liu (Nanjing University of Information Science and Technology)
Zhouchen Lin (Peking University)

More from the Same Authors