thinQmaster Virtual testing partner Adaptive practice exams that find your weak spots and drill them. Every answer explained. From $69
Enroll

thinQmaster

Find your exam

Five questions free, before you decide.

Step 1 · choose a vendor

In one of our camps? Your practice exams come with the tuition — they are on your study page already.

AI & Machine Learning · Intermediate

Reward Function

The signal that tells a reinforcement learning agent how good an action or outcome was.

Definition

The reward function defines the objective the agent optimizes. Poorly specified rewards lead to reward hacking, where the agent exploits loopholes rather than achieving the intended goal. In RLHF a learned reward model stands in for a hand-written reward function.

Example

Awarding +1 for reaching a goal and -0.01 per step to encourage efficiency.

Exam tip

Reward hacking is a common alignment risk to recognize.

Related terms