Evolution Strategies
📰 Lilian Weng's Blog
Evolution Strategies (ES) can be used for optimizing model parameters when gradients are unknown or cannot be computed directly
Action Steps
- Identify problems where gradients are unknown or cannot be computed directly
- Explore classic ES methods such as gradient-free optimization
- Apply ES to deep reinforcement learning tasks
Who Needs to Know This
Machine learning engineers and researchers can benefit from ES, especially in deep reinforcement learning scenarios where traditional gradient descent methods may not be applicable
Key Insight
💡 ES can be used as an alternative to gradient descent in certain scenarios
Share This
🤖 Evolution Strategies: optimize model params without gradients!
Key Takeaways
Evolution Strategies (ES) can be used for optimizing model parameters when gradients are unknown or cannot be computed directly
Full Article
<!-- Gradient descent is not the only option when learning optimal model parameters. Evolution Strategies (ES) works out well in the cases where we don't know the precise analytic form of an objective function or cannot compute the gradients directly. This post dives into several classic ES methods, as well as how ES can be used in deep reinforcement learning. --> <p>Stochastic gradient descent is a universal choice for optimizing deep learning models. However, it is not the only option. With bl
DeepCamp AI