Superposed Inference for Hyperdimensional Computing

2026-09-26

AI Summary

This study addresses the high-dimensional projection overhead in traditional hyperdimensional computing (HDC) caused by independently encoding queries. We propose SupHDC, a paradigm that repurposes high-dimensional redundancy as shared inference capacity for the first time. By superimposing multiple queries through random slot keys with shared encoding and employing a lightweight classifier to directly recover predictions, SupHDC demonstrates that categorical evidence can be preserved without exact reconstruction. Bridging HDC with the perspective of random feature kernels, this method achieves a measured 2.01× speedup on a Raspberry Pi 5 with only a 2.26 percentage point drop in average accuracy. Consequently, it provides an efficient superposition-encoded inference solution tailored for resource-constrained scenarios.

Abstract

Hyperdimensional computing (HDC) is attractive for efficient and robust learning, but conventional inference still encodes every query independently, repeatedly paying the cost of high-dimensional projection. We introduce SupHDC, a new inference paradigm that processes multiple queries through a shared encoding computation. SupHDC assigns lightweight random slot keys, superposes the keyed queries before encoding, and uses slot-specific classifiers to recover their individual predictions. A random-feature kernel view explains why exact recovery of each hypervector is unnecessary: inference only needs to preserve the class evidence that determines the prediction. Across ten datasets, SupHDC achieves 1.39x analytical speedup with no average accuracy loss, and up to 2.08x speedup with only a 2.67 percentage-point mean accuracy loss. On a Raspberry Pi~5, it delivers 2.01x measured wall-clock speedup with a 2.26 percentage-point loss in mean prediction accuracy. SupHDC shows that high-dimensional redundancy can be used not only for robustness, but also as capacity for shared inference.

Problem

• Hyperdimensional Computing
• Inference Efficiency
• High-dimensional Projection
• Superposed Inference

Innovation

• Hyperdimensional Computing
• Superposed Inference
• Shared Encoding
• Random Feature Kernel
• Slot Keys