When does fusing hand-crafted knowledge with learned representations pay? A cost-normalized benchmark of stacking, substitution, and interference
Fusing prior knowledge with data-driven learning is attractive where data is scarce, yet no controlled account says when it helps, is redundant, or harms. We benchmark one fixed hand-crafted knowledge source, a pinned bank of Gabor targets injected only during training at $\sim$2\% overhead, against data-driven alternatives (SimCLR, SimSiam, DINO, ImageNet transfer, augmentation, learned teachers) under one frozen recipe with fixed subsets: 13 datasets, 9 backbones, 150 to 1.28M images, 32--224\,px, 2.5M--86M parameters ($\computeCells$ classification configurations over $\computeRuns$ runs, p
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 52%Fractale-350M-base: memory as trained behaviour instead of long context, a fully open research release →
- PossiblePossibly related (embedding) · 47%Fine-tuning →
- PossiblePossibly related (embedding) · 45%When Unlearning Is Free: Leveraging Low Influence Points to Reduce Computational Costs - Apple Machine Learning Research →
- LinkedLinked via arxiv author · 85%Ahmad AlMughrabi →
“When does fusing hand-crafted knowledge with learned representations pay? A cost-normalized benchmark of stacking, subst”
- LinkedLinked via arxiv author · 85%Albert Clop →
“When does fusing hand-crafted knowledge with learned representations pay? A cost-normalized benchmark of stacking, subst”
- LinkedLinked via arxiv author · 85%Benjamin Busam →
“When does fusing hand-crafted knowledge with learned representations pay? A cost-normalized benchmark of stacking, subst”
- LinkedLinked via arxiv author · 85%Ricardo Marques →
“When does fusing hand-crafted knowledge with learned representations pay? A cost-normalized benchmark of stacking, subst”