Timezone: »

 
Poster
Tiled convolutional neural networks
Quoc V. Le · Jiquan Ngiam · Zhenghao Chen · Daniel Jin hao Chia · Pang Wei Koh · Andrew Y Ng

Tue Dec 07 12:00 AM -- 12:00 AM (PST) @ None #None

Convolutional neural networks (CNNs) have been successfully applied to many tasks such as digit and object recognition. Using convolutional (tied) weights significantly reduces the number of parameters that have to be learned, and also allows translational invariance to be hard-coded into the architecture. In this paper, we consider the problem of learning invariances, rather than relying on hard-coding. We propose tiled convolution neural networks (Tiled CNNs), which use a regular “tiled” pattern of tied weights that does not require that adjacent hidden units share identical weights, but instead requires only that hidden units k steps away from each other to have tied weights. By pooling over neighboring units, this architecture is able to learn complex invariances (such as scale and rotational invariance) beyond translational invariance. Further, it also enjoys much of CNNs’ advantage of having a relatively small number of learned parameters (such as ease of learning and greater scalability). We provide an efficient learning algorithm for Tiled CNNs based on Topographic ICA, and show that learning complex invariant features allows us to achieve highly competitive results for both the NORB and CIFAR-10 datasets.

Author Information

Quoc V. Le (Google)
Jiquan Ngiam (Stanford University)
Zhenghao Chen (Stanford University)
Daniel Jin hao Chia (Stanford University)
Pang Wei Koh (Stanford University)
Andrew Y Ng (Baidu Research)

More from the Same Authors