LogHD: Robust Compression of Hyperdimensional Classifiers via Logarithmic Class-Axis Reduction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yun, Sanggeon, Oh, Hyunwoo, Masukawa, Ryozo, Mercati, Pietro, Bastian, Nathaniel D., Imani, Mohsen
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912869971394560
author Yun, Sanggeon
Oh, Hyunwoo
Masukawa, Ryozo
Mercati, Pietro
Bastian, Nathaniel D.
Imani, Mohsen
author_facet Yun, Sanggeon
Oh, Hyunwoo
Masukawa, Ryozo
Mercati, Pietro
Bastian, Nathaniel D.
Imani, Mohsen
contents Hyperdimensional computing (HDC) suits memory, energy, and reliability-constrained systems, yet the standard "one prototype per class" design requires $O(CD)$ memory (with $C$ classes and dimensionality $D$). Prior compaction reduces $D$ (feature axis), improving storage/compute but weakening robustness. We introduce LogHD, a logarithmic class-axis reduction that replaces the $C$ per-class prototypes with $n\!\approx\!\lceil\log_k C\rceil$ bundle hypervectors (alphabet size $k$) and decodes in an $n$-dimensional activation space, cutting memory to $O(D\log_k C)$ while preserving $D$. LogHD uses a capacity-aware codebook and profile-based decoding, and composes with feature-axis sparsification. Across datasets and injected bit flips, LogHD attains competitive accuracy with smaller models and higher resilience at matched memory. Under equal memory, it sustains target accuracy at roughly $2.5$-$3.0\times$ higher bit-flip rates than feature-axis compression; an ASIC instantiation delivers $498\times$ energy efficiency and $62.6\times$ speedup over an AMD Ryzen 9 9950X and $24.3\times$/$6.58\times$ over an NVIDIA RTX 4090, and is $4.06\times$ more energy-efficient and $2.19\times$ faster than a feature-axis HDC ASIC baseline.
format Preprint
id arxiv_https___arxiv_org_abs_2511_03938
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LogHD: Robust Compression of Hyperdimensional Classifiers via Logarithmic Class-Axis Reduction
Yun, Sanggeon
Oh, Hyunwoo
Masukawa, Ryozo
Mercati, Pietro
Bastian, Nathaniel D.
Imani, Mohsen
Machine Learning
Hyperdimensional computing (HDC) suits memory, energy, and reliability-constrained systems, yet the standard "one prototype per class" design requires $O(CD)$ memory (with $C$ classes and dimensionality $D$). Prior compaction reduces $D$ (feature axis), improving storage/compute but weakening robustness. We introduce LogHD, a logarithmic class-axis reduction that replaces the $C$ per-class prototypes with $n\!\approx\!\lceil\log_k C\rceil$ bundle hypervectors (alphabet size $k$) and decodes in an $n$-dimensional activation space, cutting memory to $O(D\log_k C)$ while preserving $D$. LogHD uses a capacity-aware codebook and profile-based decoding, and composes with feature-axis sparsification. Across datasets and injected bit flips, LogHD attains competitive accuracy with smaller models and higher resilience at matched memory. Under equal memory, it sustains target accuracy at roughly $2.5$-$3.0\times$ higher bit-flip rates than feature-axis compression; an ASIC instantiation delivers $498\times$ energy efficiency and $62.6\times$ speedup over an AMD Ryzen 9 9950X and $24.3\times$/$6.58\times$ over an NVIDIA RTX 4090, and is $4.06\times$ more energy-efficient and $2.19\times$ faster than a feature-axis HDC ASIC baseline.
title LogHD: Robust Compression of Hyperdimensional Classifiers via Logarithmic Class-Axis Reduction
topic Machine Learning
url https://arxiv.org/abs/2511.03938