find_optimal
The job allows you to perform optimization-based searches across your archived data, providing efficient retrieval and enhanced data discovery capabilities.
You can use the job to analyze experimental scores associated with content in your archive. By passing the id_to_score parameter, the job utilizes gradient-based optimization to estimate the theoretical ideal vector that would achieve the maximum score.
Once the optimal vector is predicted, the job searches all indexed vectors within the archive, or a specified subset, and returns the ID of the most similar indexed vector to the predicted best vector given the experimental data.
Required Account Privileges: "read"
Request JSON ["inputs"]:
"archive":
string (3 <= len <= 30) unique in account
null NOT allowed
A unique string identifier for the archive within your account.
"archive_content_ids_subset":
list of ints
null allowed
Optional. A list of integers representing the IDs of the specific contents to consider for filtering. If not provided, all contents in the archive will be considered.
"id_to_score":
dict of int to float
null NOT allowed
This parameter is a dictionary where each key is an identifier (ID) corresponding to a specific piece of content in the archive, and each value is the associated score from your experiments. The scores should reflect the performance or relevance of the content based on your experimental data.Response JSON ["results"]
"id_closest_to_max":
int or null