What is a target network in DQN?

Updated May 17, 2026

Short answer

A fixed copy of the Q-network used for stable target computation.

Deep explanation

The target network is updated less frequently to reduce oscillations in Q-value updates.

Real-world example

Used in autonomous driving simulation training.

Common mistakes

Updating target network too frequently.

Follow-up questions

How often should target network update?
What happens without it?

More Q-Learning interview questions

How does Q-Learning handle exploration-exploitation under uncertainty in large state spaces?senior
What is the relationship between Q-Learning and fixed-point convergence?senior
How does Q-Learning behave when reward signals are delayed and noisy simultaneously?senior
What is the impact of state representation quality on Q-Learning convergence?senior
How does Q-Learning handle catastrophic bootstrapping errors?senior
What is the role of reward normalization in stabilizing deep Q-networks?senior
How does Q-Learning behave under function approximation + off-policy mismatch?senior
How does Q-Learning interact with non-convex function approximation landscapes?senior