primitive_frequency_mix: a second identification carrier for 002#6
Merged
Conversation
- classifier rebuilt on per-ink footprints from the overprint arithmetic; fully diagonal on the five primitives (recall and precision 1.0); edge- touching components excluded from the mix (unbiased small-fragment estimates) - scoring: L1 vs committed mix with n-dependent tolerance (committed x max(1, sqrt(n_ref/n)), n_ref 16) so multinomial sampling noise is not punished as disagreement - iso-002.yaml v1.1.0: mix committed (reconstruction's composition model per SI-018), weights rebalanced ink 0.60 / mix 0.40 - same-ink/different-composition impostors: 0/624 fragments identified across all four single-primitive compositions (closes SI-022's dangerous half); genuine held-out surfaces 0.89 identified, frac-0.2 mean 0.639 - SI-008 resolved-for-reconstruction, SI-022 updated, SI-026 added - 21 mix tests; 146 total green
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Milestone 3, Phase 9 — closes the dangerous half of SI-022 ("two grid dialects sharing an ink set are indistinguishable") by measuring, committing and scoring the primitive-frequency mix.
committed × max(1, √(16/n))) so multinomial sampling noise on small fragments is priced as uncertainty, not punished as disagreement; the working reports n, effective tolerance and L1.expected [0.30, 0.10, 0.24, 0.20, 0.16], tolerance 0.95), weights rebalanced ink 0.60 / mix 0.40. Explicit caveat per SI-018: this commits the reconstruction's composition model, not Studio.Build's original.Results
Review note
The first draft was rejected in director review for gutting genuine small-fragment strength (0.30 at frac 0.2) and letting all-staircase cross 0.70; both were fixed by the edge-exclusion + n-dependent-tolerance redesign and three classifier recall fixes.
Tests
21 mix tests; merged with main (incl. #5): 154 passed.