Negative Learning Rates and P-Learning
arXiv:1603.08253v31 citations
Originality Synthesis-oriented
AI Analysis
This addresses a problem in machine learning for researchers and practitioners by introducing a novel training approach, though it appears incremental as it builds on existing methods.
The paper tackles training differentiable function approximators for regression tasks using negative examples and negative learning rates, and extends the method to direct policy learning in reinforcement learning.
We present a method of training a differentiable function approximator for a regression task using negative examples. We effect this training using negative learning rates. We also show how this method can be used to perform direct policy learning in a reinforcement learning setting.