XL-HD uses fixed Sobol projections and real-to-binary prototype optimization to deliver competitive accuracy on MNIST, UCIHAR, and ISOLET with a 0.395 mm² IMC engine at 0.40 μJ per inference.
Hassan Najafi, and Mohsen Imani
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
XL-HD: Extended Learning in Hyperdimensional Computing via Deterministic Projections for In-Memory Accelerators
XL-HD uses fixed Sobol projections and real-to-binary prototype optimization to deliver competitive accuracy on MNIST, UCIHAR, and ISOLET with a 0.395 mm² IMC engine at 0.40 μJ per inference.