English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present a novel definition of the reinforcement learning state, actionsand reward function that allows a deep Q-network (DQN) to learn to control anoptimization hyperparameter. Using Q-learning with experience replay, we traintwo DQNs to accept a state representation of an objective function as input andoutput the expected discounted return of rewards, or q-values, connected to theactions of either adjusting the learning rate or leaving it unchanged. The twoDQNs learn a policy similar to a line search, but differ in the number ofallowed actions. The trained DQNs in combination with a gradient-based updateroutine form the basis of the Q-gradient descent algorithms. To demonstrate theviability of this framework, we show that the DQN's q-values associated withoptimal action converge and that the Q-gradient descent algorithms outperformgradient descent with an Armijo or nonmonotone line search. Unlike traditionaloptimization methods, Q-gradient descent can incorporate any objectivestatistic and by varying the actions we gain insight into the type of learningrate adjustment strategies that are successful for neural network optimization.

Using Deep Q-Learning to Control Optimization Hyperparameters