This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
2
Calculating KL-Divergence Between Two Q-Learning Policies?
Post Flair (click to view more posts with a particular flair)
Post Body
Hi everyone,
Iām looking to calculate the KL-Divergence between two policies trained using Q-learning. Since Q-learning selects actions based on the highest Q-value rather than generating a probability distribution, should these policies be represented as one-hot vectors? If so, how can we calculate KL-Divergence given the issues with zero probabilities in one-hot vectors?
Author
Account Strength
80%
Account Age
4 years
Verified Email
Yes
Verified Flair
No
Total Karma
159
Link Karma
112
Comment Karma
47
Profile updated: 1 week ago
Posts updated: 1 day ago
Subreddit
Post Details
We try to extract some basic information from the post title. This is not
always successful or accurate, please use your best judgement and compare
these values to the post title and body for confirmation.
- Posted
- 7 months ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/reinforceme...