Hyperdimensional computing and vector symbolic architectures
Representing structured meaning in very high-dimensional random vectors.
Hyperdimensional computing works with vectors of thousands of dimensions, where nearly all random vectors are close to orthogonal. In that regime, symbols can be combined and taken apart with simple, reversible operations: binding ties two symbols into a new one, superposition adds symbols into a set, and permutation encodes order. A whole data structure becomes a single fixed-width vector.
This is the substance of vector symbolic architectures. Because binding is invertible and superposition is similarity-preserving, a bound structure can be queried by unbinding, and a noisy vector still resolves to the nearest clean symbol. The representation is holographic — information is spread across all dimensions — so it degrades gracefully and maps well onto wide, parallel, error-tolerant hardware.
The neuromorphic studio uses hyperdimensional representations as one of its computing substrates, alongside spiking and stochastic styles, which is why the same repository spans encoders, classifiers, and symbolic binding under one evidence contract.
The vocabulary of the topic.
- Near-orthogonality
- In thousands of dimensions almost all random vectors are dissimilar, giving each symbol its own direction.
- Binding
- A reversible operation that combines two symbols into a new, dissimilar one.
- Superposition
- Adding vectors to represent a set while preserving similarity to its members.
- Holographic representation
- Information spread across all dimensions, so partial corruption degrades gracefully.
Studios working on this.
Each runs standalone in its own repository and federates its evidence through the platform.
Run it, don't take it on faith.
The runnable lessons and sealed measurements this site ships for the topic.
Read it at source, in order.
Canonical references for the topic, ordered from the foundations to current work. Every one was verified at source — a DOI resolves through doi.org, a standard through its issuer — so each link goes to the real record.
Start with the foundations
- 1
The introduction to computing with high-dimensional random vectors.
- 2
Binding and superposition of symbols in a fixed-width vector — the algebra behind VSA.
Current work
- 3
The current survey of models, operations, and hardware for HDC/VSA.
Related topics.
Where this topic connects to the rest of the federation's work.