Fix MICRON handling of unequal code lists - #1234
Open
fbonc wants to merge 2 commits into
Open
Conversation
_pool_embedding padded a length-1 temporal dim up to exactly 2, then features were concatenated along the embedding dim (requiring equal temporal lengths), so any real sequence length != 2 crashed. Pad each feature to the longest feature sequence (>= 2 for the residual step) before concatenating, and combine per-feature masks with OR.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Issue
MICRON treats individual codes as visits. Unequal diagnosis and procedure counts can crash concatenation, while padding alone can cause predictions to ignore the shorter list.
Fix
Pool flat code lists into one visit, preserve nested visit histories, and return zero reconstruction loss for single-visit inputs.
Notes
Regression tests verify that both features receive prediction gradients and nested inputs retain their visits.