randsel

torchhd.randsel(input: VSATensor, other: VSATensor, *, p: float = 0.5, generator: Generator | None = None) → VSATensor[source]

Bundles two hypervectors by selecting random elements.

A bundling operation is used to aggregate information into a single hypervector. The resulting hypervector has elements selected at random from input or other.

\[\oplus: \mathcal{H} \times \mathcal{H} \to \mathcal{H}\]
Parameters:

  • input (VSATensor) – input hypervector

  • other (VSATensor) – other input hypervector

  • p (float, optional) – probability of selecting elements from the input hypervector. Default: 0.5.

  • generator (torch.Generator, optional) – a pseudorandom number generator for sampling.

Shapes:

  • Input: \((*)\)

  • Other: \((*)\)

  • Output: \((*)\)

Examples:

>>> a, b = torchhd.random(2, 6, "FHRR")
>>> a
tensor([-0.7404-0.6721j,  0.8280-0.5608j, -0.5059+0.8626j, -0.9965-0.0841j, -0.7337+0.6795j, -0.9925-0.1223j])
>>> b
tensor([-0.5593+0.8290j,  0.8097-0.5869j,  0.8306+0.5569j, -0.4970+0.8678j,  0.9962+0.0875j, -0.6631+0.7485j])
>>> torchhd.randsel(a, b)
tensor([-0.7404-0.6721j,  0.8280-0.5608j, -0.5059+0.8626j, -0.9965-0.0841j, -0.7337+0.6795j, -0.9925-0.1223j])