Search Unity

  1. We are migrating the Unity Forums to Unity Discussions. On July 12, the Unity Forums will become read-only.

    Please, do not make any changes to your username or email addresses at id.unity.com during this transition time.

    It's still possible to reply to existing private message conversations during the migration, but any new replies you post will be missing after the main migration is complete. We'll do our best to migrate these messages in a follow-up step.

    On July 15, Unity Discussions will become read-only until July 18, when the new design and the migrated forum contents will go live.


    Read our full announcement for more information and let us know if you have any questions.

Question Metrics on which observations are significant?

Discussion in 'ML-Agents' started by sarachan, May 29, 2022.

  1. sarachan

    sarachan

    Joined:
    Jul 21, 2013
    Posts:
    43
    Hi,
    I have a model that trains fairly well and that uses a few geometric observations. I have been trying to determine whether all the observations are significant for the model or whether some can be left out by retraining with different combinations of observations. Unfortunately, this is tedious and time consuming. Does the training process report any metrics on how significant each observation type is for the trained model? that would be extremely helpful in knowing whether some observations can be left out, and the impact of leaving them out if they do have some small contribution to the model.

    A related question: Is there a downside to having some "unnecessary" observations or will the training process just learn to ignore some observations if they don't matter. I would guess that training will take longer if there are more observations.

    I hope these questions are clear!

    Thanks!
     
    iffalseelsetrue likes this.
  2. iffalseelsetrue

    iffalseelsetrue

    Joined:
    May 3, 2018
    Posts:
    11
    like your question, also would like to know the answer!
     
    sarachan likes this.