The energy landscape of a simple neural network
arXiv:1706.07101v14 citations
Originality Synthesis-oriented
AI Analysis
This work provides incremental insights into implicit regularization in neural networks, relevant for researchers in machine learning theory.
The paper investigates the energy landscape of a simple neural network, building on prior findings that neural networks exhibit lower empirical complexity than expected from parameter counts, which aids generalization.
We explore the energy landscape of a simple neural network. In particular, we expand upon previous work demonstrating that the empirical complexity of fitted neural networks is vastly less than a naive parameter count would suggest and that this implicit regularization is actually beneficial for generalization from fitted models.