Applications of Sub-linear Algorithms on Other Fields
summary
Argues for applying sublinear methods to medical imaging with concrete results and open problems: using coresets and importance sampling to approximate GLCM-based features reduces a pipeline step from 30 to 3 minutes, but currently has no provable guarantees; coreset-based lifelong learning cuts time and communication for edge, cross-silo settings; universal sketching (UnivMon) finds real deployment in networks and inspired load‑balancing at Intel; data growth and cloud cost can be enormous — roughly $1M/year for radiomics in some practices — creating urgent federated, multimodal, sequence-aware sublinear challenges.
AI-written summary · about