Super learner
This user is a Super Learner. To become a Super Learner, you need to reach Level 8.
Last answered:

04 Jul 2024

Posted on:

02 Jul 2024

0

REWARD FUNCTIONS

Are reward functions not functions we are trying to maximize?
If I got the lesson correctly, the higher the reward function, the better the output. This means higher reward function is a joy to the Machine Learning scientist.

Please educate me.

2 answers ( 0 marked as helpful)
Instructor
Posted on:

04 Jul 2024

1

Hey Jonathan,

Good to hear from you!

Yes, you are correct! Reward functions in the context of machine learning, particularly in reinforcement learning, are designed to provide feedback to the model about how well it is performing. Higher rewards indicate better performance, and improving these rewards is a key objective for machine learning practitioners.

Best,

Ned

Super learner
This user is a Super Learner. To become a Super Learner, you need to reach Level 8.
Posted on:

04 Jul 2024

0

Thank you Ned for the clarification.

I asked the question because of what the correct answer was stated as in the number 2 question:

Reward functions are NOT:

I chose the option: "functions we are trying to maximize"

However, I got the feedback that the correct answer should be: "functions we are trying to minimize"


Did I misunderstand the question?

Submit an answer