group vsa
UnderstandingHyperdimensionalComputingfor ParallelSingle-PassLearning
Weextend our analysis to the more general class of vector symbolic architectures (VSA), which compute withhigh-dimensional vectors(hypervectors) thatarenotnecessarily binary.
Appendix: Understanding Hyperdimensional Computing for Parallel Single-Pass Learning A Proofs of Lemmas, Statements and Theorems
Binary HDC cannot learn the following task. A binary HDC that achieves this matrix is: ( 1, 1, 1), (1, 1, 1), (1, 1, 1) . Suppose the first VSA's group is Note the fact that all irreducible representations of finite Abelian groups are 1-dimensional. Consider the binary icosahedral group expressed as a subset of the quaternions. This means that the more vectors we bundle together, the closer θ is to 90 degrees.
Country:
- North America > United States (0.28)
- Africa > Chad > Salamat (0.05)
- Europe > United Kingdom > England > Cambridgeshire > Cambridge (0.04)
- Asia (0.04)
Technology:
- Information Technology > Artificial Intelligence > Representation & Reasoning (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Statistical Learning (1.00)
- Information Technology > Artificial Intelligence > Machine Learning > Neural Networks (0.69)
- Information Technology > Data Science (0.68)