Complex-Valued Sparse Distributed Memory for Lower Hardware Overhead

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional SDM implementations require a large hardware overhead to store and retrieve vectors of integer values due to the need for multiple index locations for each element, leading to inefficient storage and retrieval processes.

Innovation Solution

Implementing a complex valued sparse distributed memory (CV-SDM) technique that represents integer elements as complex numbers, reducing storage requirements from r*n to 2*n by using real and imaginary components for each element, thereby optimizing hardware usage.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If integer SDM techniques are used to represent data, then the ability to represent real-world data is improved, but the hardware overhead increases significantly

Engineering Contradiction:

Improveability to represent real-world dataVSAvoidhardware overhead

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent changes the parameter representation from integer values to complex number values. Instead of using r index locations for each integer value, the system uses complex numbers with real and imaginary parts to represent the same information, reducing the storage requirement from r*n to 2*n hardware components.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent transitions from a one-dimensional integer representation to a two-dimensional complex number representation (real and imaginary parts). This dimensional change allows the system to represent the same information using fewer storage components, as each complex number occupies 2 storage locations instead of requiring r separate index locations for each possible integer value.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Quantity of substance

If multiple index locations are used for each element in conventional SDM, then the storage capacity is improved, but the retrieval process becomes less efficient

Engineering Contradiction:

Improvestorage capacityVSAvoidretrieval efficiency

Core Design Contradiction:

Quantity of substanceVSProductivity

Solution Approach 1:

The patent merges the storage and retrieval operations into a more efficient process. By using complex numbers, the system combines the representation of multiple integer values into a single mathematical operation, allowing simultaneous storage and retrieval without requiring separate index location management, thus improving retrieval efficiency while maintaining storage capacity.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20250355794A1Complex valued sparse distributed memories

Publication Date: 2025.11.20 HEWLETT PACKARD ENTERPRISE DEV LP

AI summary

In certain examples, a method includes obtaining a SDM word to be written to a contents matrix; obtaining an address corresponding to the SDM word; determining a set of activated address rows of an address matrix using the address; selecting a set of activated contents matrix rows of the contents matrix based on the set of activated address rows; calculating a complex value based on an element of the SDM word, wherein an element position of the element corresponds to a contents matrix column pair of the contents matrix; and for each row of the set of activated contents matrix rows: modifying a real component value of a first column of the contents matrix column pair based on a real part of the complex value; and modifying an imaginary component value of a second column of the contents matrix column pair based on an imaginary part of the complex value.