AI engineering glossary · Reinforcement learning
What is softmax exploration?
Softmax exploration (also called Boltzmann exploration): An exploration method that picks each choice with probability proportional to exp(Q / temperature). High temperature makes choices nearly uniform; low temperature makes it nearly greedy. Unlike random-with-fixed-chance methods, it favours better-looking options over clearly bad ones.
Softmax exploration: an example
With Q values 1 and 2 and temperature 1, probabilities are about 0.27 and 0.73.
Related terms
More Reinforcement learning terms
- Sparse reward
- State
- Stochastic policy
- Target network
- TD error
- TD(0)
- TD(lambda)
- Temporal-difference learning
Where you learn it
Softmax exploration is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.