Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

HDC-X: Efficient Medical Data Classification for Embedded Devices

Jianglan Wei, Zhenyu Zhang, Pengcheng Wang, Mingjie Zeng, Zhigang Zeng

Abstract

Energy-efficient medical data classification is essential for modern disease screening, particularly in home and field healthcare where embedded devices are prevalent. While deep learning models achieve state-of-the-art accuracy, their substantial energy consumption and reliance on GPUs limit deployment on such platforms. We present HDC-X, a lightweight classification framework designed for low-power devices. HDC-X encodes data into high-dimensional hypervectors, aggregates them into multiple cluster-specific prototypes, and performs classification through similarity search in hyperspace. We evaluate HDC-X across three medical classification tasks; on heart sound classification, HDC-X is $350\times$ more energy-efficient than Bayesian ResNet with less than 1% accuracy difference. Moreover, HDC-X demonstrates exceptional robustness to noise, limited training data, and hardware error, supported by both theoretical analysis and empirical results, highlighting its potential for reliable deployment in real-world settings. Code is available at https://github.com/jianglanwei/HDC-X.

HDC-X: Efficient Medical Data Classification for Embedded Devices

Abstract

Energy-efficient medical data classification is essential for modern disease screening, particularly in home and field healthcare where embedded devices are prevalent. While deep learning models achieve state-of-the-art accuracy, their substantial energy consumption and reliance on GPUs limit deployment on such platforms. We present HDC-X, a lightweight classification framework designed for low-power devices. HDC-X encodes data into high-dimensional hypervectors, aggregates them into multiple cluster-specific prototypes, and performs classification through similarity search in hyperspace. We evaluate HDC-X across three medical classification tasks; on heart sound classification, HDC-X is more energy-efficient than Bayesian ResNet with less than 1% accuracy difference. Moreover, HDC-X demonstrates exceptional robustness to noise, limited training data, and hardware error, supported by both theoretical analysis and empirical results, highlighting its potential for reliable deployment in real-world settings. Code is available at https://github.com/jianglanwei/HDC-X.

Paper Structure

This paper contains 17 sections, 3 theorems, 11 equations, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Hyperdimensional Computing (HDC)
  3. Medical Data Classification through HDC-X
  4. Encode Sample into Hypervector
  5. Class-Wise Hyperspace Clustering
  6. Retrain Cluster Prototypes
  7. Results and Discussions
  8. Datasets
  9. Experiment Setup
  10. Comparison with Baseline Models
  11. Impact of HDC-X Hyperparameters
  12. Robustness for Real-World Deployment
  13. Resilience to Input Noise
  14. Resilience to Limited Training Data
  15. Resilience to Hardware Errors
  16. ...and 2 more sections

Key Result

Theorem 1

Let $s^{(1)}, s^{(2)} \in \mathcal{S} \subseteq \mathbb{R}^d$ be two feature vectors. $S = f_{\mathcal{ID}, \mathcal{L}, \Theta}(s)$ denotes the hypervector encoding defined in Equation eqn: mapping. Suppose that for all indices of the features $n \in \{1, \dots, d\}$, where $\Delta_i$ is the difference between the upper and lower bounds of the $i^{th}$ feature value, and $\delta \in [0, 1]$ deno

Figures (5)

  • Figure 1: Automated disease screening through medical data classification (left) and HDC-X performance on heart sound classification (right). HDC-X is 350$\times$ more energy-efficient than Bayesian ResNet and supports GPU-free inference, highlighting its potential for embedded deployment.
  • Figure 2: Divide feature $n$'s value range into $M$ intervals.
  • Figure 3: Medical data classification through HDC-X: Training samples are encoded into sample hypervectors (Sample-HVs) and aggregated into a compact set of cluster prototypes (Cluster-HVs); new samples are classified by selecting the Cluster-HV with highest similarity. The figure illustrates a binary classification example, though HDC-X is not limited to binary tasks.
  • Figure 4: HDC-X sensitivity to hyperparameters, input noise, limited training data, and hardware errors on PhysioNet 2016.
  • Figure 5: Conceptual hardware framework for HDC-X.

Theorems & Definitions (3)

  • Theorem 1: Robustness to Input Noise
  • Theorem 2: Distance Between Cluster Prototype and Constituents
  • Theorem 3: Robustness to Hardware Error