Dice at N = 10 labels โ skill vs the strongest net
Each domain: the best LLM-skill path (teal) against the strongest matched-N neural net found across seeds and architectures (clay). Ordered by margin. Bars start at 0.50 Dice.
Where the skills win is set by the target's visual cue
The GPU-free classical decoder wins exactly where the target has an exploitable appearance cue, and needs a different decode skill where it doesn't (isodense CT โ shape-prior; filaments โ vesselness).
| Domain | Modality | Target contrast cue | best net @10 | LLM skill @10 | ฮ | skill path |
|---|
Skill flat, net climbing โ the crossover budget is the whole story
The skill line is near-constant in N (the drawer is GT-free; only a light genome uses labels, saturating by Nโ10) while the net climbs with more labels. Where the skill leads at N=10 it holds through N=25/50 until the net catches up โ the crossover is the honest measure of the low-label edge. Shaded = skill ahead.
Four transferable decode skills โ and where each deploys
Every skill is data / numpy, LLM-agnostic, and SkillOpt-developed (rollout โ reflect โ strict-improve validation gate โ converge). They deploy exactly where a valid GT-free signal exists.
Draw โ GrabCut โ genome โ judge
Claude draws (GT-free), classical GrabCut + numpy genome decodes, a visual judge sends verbal feedback back to the skills. Wins the appearance-cue domains at low N.
Template placed by a spatial prior
On isodense CT the boundary is invisible โ GrabCut carves noise (0.002). A mean-mask template placed by a GT-free localizer recovers it (~250ร). The wall is localizing the unseen edge, not the prior (oracle @GT-box 0.84 > net).
Frangi โ shallow RF decode
Thin retinal vessels can't be boxed. A ridge-filter stack feeds a shallow RandomForest (โค10 labels, CPU). Beats a matched-N net decisively โ the extreme low-data case where nets fail hardest.
Route by a GT-free cue
A two-scalar cue (saturation + within-box Otsu separation) routes appearance targets โ GrabCut and isodense CT โ the shape-prior. Integrates the isodense skill automatically (Spleen 0.002 โ 0.438 from routing alone).
Caveats kept in view, and what's still running
- Corrected margins. The dramatic single-seed leads were an artifact; every number here is vs the arch-max best matched-N net. FetalHead: +0.338 โ +0.105.
- Nets are seed-noisy at low N (Spleen N=10: 0.547 ยฑ 0.206), and no architecture dominates โ mit_b3 wins FetalHead, UNet++ wins Kvasir/Spleen, unet-resnet34 wins Prostate. The skill paths are near-deterministic.
- Subset caveat. Feedback-loop & isodense numbers are on fixed 15โ28-image test subsets (medians reported); net & deployed-pipeline numbers are full test.
- Running now: full-default nnU-Net (1000-epoch, @224, all 6 domains, ~25 h) to make the gold-standard comparison airtight โ its reduced run already shows nnU-Net@10 weak on Kvasir (0.559) and only a tie on CXR; SegFormer N=100 fill; and the Synapse multi-organ CT run applying the shape-prior + bilateral-kidney multi-ROI to a fresh abdominal dataset.
GPU-free feedback loop โ per-stage SkillOpt & the walls
The classical loop (Claude box โ GrabCut / shape-prior / vesselness + numpy genome + judge โ no SAM, no GPU) was pushed stage-by-stage on all six domains. It beats the net where a geometric/appearance cue exists; elsewhere every stage moved (+0.05โ0.09) but hit a GT-box-oracle-proven wall that only the deployed GPU frozen-model lane clears.
| Domain | before [cls] | after [cls] | ฮ | stage pushed | verdict ยท wall |
|---|---|---|---|---|---|
| ISIC | .923 | .940 | +.017 | drawer redraw | WIN all N โ beats net (NN .835/.847/.825) |
| FetalHead | .921 | .936 | +.015 | ellipse-fit refine | WIN all N โ beat old oracle (NN .583/.800/.883) |
| Kvasir | .678 | .771 | +.093 | multi-polyp boxes | WIN N=10; loss Nโฅ25 โ localization wall (GT-box oracle .89) โ [gpu] |
| Prostate | .703 | .751 | +.048 | blur + CLAHE | WIN N=10, tie N=50 โ box-instability wall (oracle .82) โ [gpu] |
| CXR | .842 | .880 | +.038 | CLAHE | LOSS all โ soft-border DECODE wall, GT-oracle .88 โช NN .92โ.95 โ [gpu] |
| Spleen | .50 | .73 | +.23 | laterality localizer | LOSS (mean) โ isodense saturation (oracle .85) โ [gpu] |
Failure cases โ where the feedback loop breaks
The three lowest-Dice test images per domain from the deployed GPU-free feedback loop (GrabCut + numpy genome; isodense shape-prior for Spleen). Green = ground truth, red = prediction. Each domain's tail is a single, nameable perception failure โ not random noise โ which is exactly what the GT-free gate cannot always catch.
faint diffuse periphery โ the pigment fades into skin with no crisp rim for GrabCut to snap to, so the mask stops short of the true lesion extent.
acoustically-dropped skull walls โ where the bright skull ring fades in the shadowed lateral/inferior arc, the boundary is guessed and the ellipse leaks or clips.
polyp โ mucosa mislocalization โ a low-contrast sessile polyp is the same colour as the surrounding mucosa, so the box+GrabCut latches the wrong structure or the dark lumen.
isodense gland under-seg โ apex/base slices where the gland is small and iso-intense with adjacent muscle/bladder; the colour model under-covers or grabs the neighbour.
soft diaphragm over-extension โ with no clean intensity edge at the diaphragm/cardiac border, the per-lung box bleeds below the costophrenic angle into the abdomen.
whole-volume localizer miss on isodense CT โ the GT-free membership prior mislocalizes atypical caudal/apical slices entirely (Dice 0.00); no intensity edge exists to recover the crescent.