SPARCL: Spectral Partitioned Analytic Continual Learning
Analytic continual learning has emerged as a strong exemplar-free alternative to gradient-based class-incremental learning because it replaces iterative optimization with closed-form ridge updates. Yet the usual forgetting narrative, centered on stochastic gradient overwriting, does not explain why analytic methods still drift on old classes despite exact recursive solvers. We identify the culprit as spectral interference: the joint ridge classifier for all tasks shares the inverse autocorrelation operator $(R+λI)^{-1}$, so incoming task samples that load onto old dominant eigendirections dilu
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) · 57%Exploring continual learning without replay buffers: Our findings using dynamic task-similarity routing [P] →
- PossiblePossibly related (embedding) · 56%What's your take on continual learning? [D] →
- PossiblePossibly related (embedding) · 54%Live Continual Learning in Machine Learning [D] →
- PossiblePossibly related (embedding) · 48%Benchmarks for live continual learning from ModelBrew [N] →
- FuzzySimilar title/name (fuzzy) · 59%aymericdamien/TopDeepLearning →
“Fuzzy title match (0.73): “SPARCL: Spectral Partitioned Analytic Continual Learning” ≈ “aymericdamien/TopDeepLearning””
- LinkedLinked via arxiv author · 85%James Hartley →
“SPARCL: Spectral Partitioned Analytic Continual Learning”
- LinkedLinked via arxiv author · 85%Zeropy Surio →
“SPARCL: Spectral Partitioned Analytic Continual Learning”
- LinkedLinked via arxiv author · 85%Daniel Whitmore →
“SPARCL: Spectral Partitioned Analytic Continual Learning”