Computer Science
5 years ago

Dense Convolutional Networks revolutionize object recognition with fewer parameters.

arXiv (Cornell University)
Convolutional Networks with Dense Connectivity
Gao Huang, Zhuang Liu, Geoff Pleiss, Laurens van der Maaten, Kilian Q. Weinberger

Paper Summary

Related papers

Theme background image
Engineering
8 years ago

Groundbreaking Benchmark Revolutionizes Dynamic Substructuring, Unlocking Transformative Possibilities.

Health Monitoring
Nonlinear Dynamics
Uncertainty Quantification
3 views
Theme background image
Engineering
2 years ago

Optimized Support Structures Revolutionize Manufacturing, Boost Productivity and Reduce Costs

Information Modeling
Structural Optimization
3 views
Theme background image
Computer Science
7 years ago

Faster Disk Speeds Unlock New Era of Computing Efficiency

Network Coding
1 view

Convolutional Networks with Dense Connectivity, or DenseNets, connect each layer to every other layer in a feed-forward way. This helps solve the vanishing-gradient problem, encourages feature reuse, and improves parameter efficiency. DenseNets outperform other networks on object recognition tasks like CIFAR-10, CIFAR-100, SVHN, and ImageNet, while using fewer parameters and less computation.