Embedding Principle of Loss Landscape of Deep Neural Networks
–Neural Information Processing Systems
Understanding the structure of loss landscape of deep neural networks (DNNs) is obviously important. In this work, we prove an embedding principle that the loss landscape of a DNN "contains" all the critical points of all the narrower DNNs.
Neural Information Processing Systems
Aug-15-2025, 09:50:00 GMT
- Country:
- Asia > China
- Europe > France (0.04)
- North America > Canada
- British Columbia > Vancouver (0.04)
- Oceania > Australia
- New South Wales > Sydney (0.04)
- Technology: