M
Matthew W. Moskewicz
Researcher at University of California, Berkeley
Publications - 39
Citations - 10330
Matthew W. Moskewicz is an academic researcher from University of California, Berkeley. The author has contributed to research in topics: Software portability & Convolutional neural network. The author has an hindex of 17, co-authored 39 publications receiving 8617 citations. Previous affiliations of Matthew W. Moskewicz include Cadence Design Systems & Princeton University.
Papers
More filters
Posted Content
SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
TL;DR: This work proposes a small DNN architecture called SqueezeNet, which achieves AlexNet-level accuracy on ImageNet with 50x fewer parameters and is able to compress to less than 0.5MB (510x smaller than AlexNet).
Proceedings ArticleDOI
Chaff: engineering an efficient SAT solver
TL;DR: The development of a new complete solver, Chaff, is described which achieves significant performance gains through careful engineering of all aspects of the search-especially a particularly efficient implementation of Boolean constraint propagation (BCP) and a novel low overhead decision strategy.
Posted Content
DenseNet: Implementing Efficient ConvNet Descriptor Pyramids
Forrest Iandola,Matthew W. Moskewicz,Sergey Karayev,Ross Girshick,Trevor Darrell,Kurt Keutzer +5 more
TL;DR: DenseNet is presented, an open source system that computes dense, multiscale features from the convolutional layers of a CNN based object classifier.
Proceedings ArticleDOI
FireCaffe: Near-Linear Acceleration of Deep Neural Network Training on Compute Clusters
TL;DR: FireCaffe is presented, which successfully scales deep neural network training across a cluster of GPUs, and finds that reduction trees are more efficient and scalable than the traditional parameter server approach.