Joseph Bethge

Binary Neural Networks (BNNs) are neural networks which use binary weights and activations instead of the typical 32-bit floating point values. Expand

We propose a new BNN architecture BinaryDenseNet, which significantly surpasses all existing 1-bit CNNs on ImageNet without tricks. Expand

Binary Neural Networks (BNNs) show promising progress in reducing computational and memory costs, but suffer from substantial accuracy degradation compared to their realvalued counterparts on large-scale datasets, e.g., ImageNet. Expand

We present a simple training strategy for binary models without using a pretrained full-precision model. Expand

We evaluate different network architectures and hyperparameters to provide useful insights on how to train a binary neural network and show how to create efficient models. Expand

We present a topology-based metric for layout stability in treemaps—the Relative Direction Change (RDC). Expand

This paper presents a hybrid treemap layout approach that optimizes layout-quality metrics by combining state-of-the-art treemAP layout algorithms. Expand

We introduce a new model for scene text recognition that only consists of off-the-shelf building blocks for neural networks. Expand

Binary neural networks are a promising approach to execute convolutional neural networks on devices with low computational power. Expand

Binary and quantized neural networks are a promising technique to run convolutional neural networks on mobile or embedded devices. Expand