[Verse 1] In nineteen eighty-eight, Kanerva showed the way Sparse distributed memory, where patterns come to play High dimensions hold the key, addresses spread so wide Binary vectors dance in space, where memories reside Ten thousand bits or more, scattered through the void Each pattern finds its neighbors, similarity deployed [Chorus] High-dimensional spaces, where the cortex learns to see Sparse and distributed, that's the way it's meant to be Hyperdimensional computing, vectors in the air Holographic representations, information everywhere SDM and HDC, the brain's computational spree Random vectors holding all we need to be memory [Verse 2] Two thousand nine brought clarity, hyperdimensional dreams Kanerva mapped the territory where distributed gleams Random vectors ten thousand strong, each dimension plays its part Bundling operations bind the concepts from the start Circular convolution, binding symbols into one Permutation operators make the sequences run [Chorus] High-dimensional spaces, where the cortex learns to see Sparse and distributed, that's the way it's meant to be Hyperdimensional computing, vectors in the air Holographic representations, information everywhere SDM and HDC, the brain's computational spree Random vectors holding all we need to be memory [Verse 3] Plate in ninety-five revealed holographic reduced forms HRRs compress the knowledge, breaking through the norms Convolution binds the pairs, correlation pulls apart Cleanup memory restores what noise tried to thwart Compositional structures built from vector multiplication Hierarchies emerge from simple transformation [Verse 4] Language processing flows through high-dimensional streams Semantic vectors capture meaning beyond our wildest dreams Nouns and verbs as vectors, syntax becomes geometry Understanding emerges from computational symmetry Neural networks learn the patterns, synaptic weights align High-dimensional cognition makes the human mind divine [Bridge] Cortical columns compute like this, each one a processing node Sparse activation patterns carry the episodic load From Kanerva's scattered memory to Plate's holographic view The mathematics shows us how the brain makes something new Distributed representations, no single point of failure Vector symbolic architecture, nature's computational savior [Chorus] High-dimensional spaces, where the cortex learns to see Sparse and distributed, that's the way it's meant to be Hyperdimensional computing, vectors in the air Holographic representations, information everywhere SDM and HDC, the brain's computational spree Random vectors holding all we need to be memory [Outro] Ten thousand dimensions dancing in the neural night Sparse distributed memory makes everything feel right
← When the Space Grows Wide | Memory Patterns Fold at Dawn →