Hyperdimensional computing in biomedical sciences: a brief review

Area of Science:

• Biomedical Informatics
• Bioinformatics
• Cheminformatics
• Medical Informatics

Background:

• Hyperdimensional computing (HDC), also known as vector-symbolic architectures (VSA), is an emerging computational paradigm.
• HDC utilizes high-dimensional vectors for information representation and combination.
• It has broad applications across scientific disciplines, including machine learning and artificial intelligence.

Purpose of the Study:

• To provide a critical and comprehensive review of the current HDC landscape in biomedical sciences.
• To highlight the advantages and disadvantages of applying HDC in bioinformatics, cheminformatics, and medical informatics.
• To identify potential future directions and challenges for HDC adoption in these domains.

Main Methods:

• Systematic review of approximately forty scientific articles on HDC applied to biomedical data.
• Analysis of key aspects including vector construction, data encoding, and programming languages used.
• Assessment of open access status, software code availability, and prominent research groups, journals, and conferences.

Main Results:

• Identified key features and methodologies in current HDC applications within biomedical sciences.
• Quantified the prevalence of open access publications and publicly available software code.
• Summarized the distribution of research contributions across authors, journals, and conferences.

Conclusions:

• HDC presents a promising computational paradigm for biomedical data analysis.
• The review outlines the current state, pros, and cons of HDC in specific biomedical domains.
• This work serves as the first open survey, offering valuable insights for researchers in biomedical sciences.