
Equinormalization of Neural Networks
Modern neural networks are overparametrized. In particular, each rectif...
read it

And the Bit Goes Down: Revisiting the Quantization of Neural Networks
In this paper, we address the problem of reducing the memory footprint o...
read it

3D Semantic Segmentation with Submanifold Sparse Convolutional Networks
Convolutional networks are the defacto standard for analyzing spatiote...
read it

LargeScale 3D Shape Reconstruction and Segmentation from ShapeNet Core55
We introduce a largescale 3D shape understanding benchmark using data a...
read it

Confusing Deep Convolution Networks by Relabelling
Deep convolutional neural networks have become the gold standard for ima...
read it

Submanifold Sparse Convolutional Networks
Convolutional network are the defacto standard for analysing spatiotem...
read it

LowPrecision BatchNormalized Activations
Artificial neural networks can be trained with relatively lowprecision ...
read it

Fractional MaxPooling
Convolutional networks almost always incorporate some form of spatial po...
read it

Spatiallysparse convolutional neural networks
Convolutional neural networks (CNNs) perform well on problems such as ha...
read it

Sparse arrays of signatures for online character recognition
In mathematics the signature of a path is a collection of iterated integ...
read it

Unsupervised learning with sparse spaceandtime autoencoders
We use spatiallysparse two, three and four dimensional convolutional au...
read it

C3DPO: Canonical 3D Pose Networks for NonRigid Structure From Motion
We propose C3DPO, a method for extracting 3D models of deformable object...
read it

The iisignature library: efficient calculation of iteratedintegral signatures and log signatures
Iteratedintegral signatures and log signatures are vectors calculated f...
read it
Benjamin Graham
is this you? claim profile