AI Memory Graph for Human-Like Recall

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current language understanding systems and personal digital assistants lack the ability to seamlessly link user memory elements across time, space, and cognitive dimensions, resulting in limited recall and non-user-centric memory management, which hampers task accomplishment and user interaction.

Innovation Solution

An AI memory system that models human memory by creating a user-centric knowledge graph, linking memory elements based on relationships in space, time, and cognitive dimensions, and utilizing machine learning and statistical modeling to enrich and rank memory elements for contextually relevant responses.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If current language understanding systems and personal digital assistants are used, then basic task completion is achieved, but the ability to link user memory elements across time, space, and cognitive dimensions is limited

Engineering Contradiction:

Improveability to link user memory elementsVSAvoidmemory recall capability

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent introduces a multi-dimensional memory graph structure that organizes memory elements across three dimensions: temporal dimension (time-based relationships), spatial dimension (location-based relationships), and cognitive dimension (conceptual relationships). This dimensional expansion allows the system to link memory elements in ways that traditional single-dimension systems cannot, directly resolving the contradiction between adaptability and information retention.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The memory graph employs a nested hierarchical structure where memory elements are organized in layers from general to specific. The graph contains nested sub-graphs that represent different levels of abstraction, allowing the system to both maintain detailed information and provide high-level summaries. This nesting enables efficient retrieval at multiple levels of detail, addressing both the need for comprehensive memory linking and effective information recall.

Inventive Principle:

Principle #7Nested doll (Nesting)

2Ease of operation

If traditional AI systems are used, then basic interactions are maintained, but user-centric memory management is absent

Engineering Contradiction:

Improveuser interaction qualityVSAvoidtask accomplishment time

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The system continuously builds and updates the memory graph in the background as users interact with the application, performing preliminary organization of memory elements before they are needed. This pre-processing of memory data allows for rapid retrieval and contextual understanding when users interact with the system, improving both interaction quality and reducing task completion time.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional keyword-based search mechanisms with a semantic memory graph system that uses machine learning models to understand contextual relationships. This substitution enables the system to comprehend user intent and retrieve relevant information based on semantic meaning rather than exact keyword matching, significantly improving ease of operation and reducing the time needed to accomplish tasks.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If a comprehensive memory graph is created, then memory recall is improved, but system complexity increases

Engineering Contradiction:

Improvememory recall accuracyVSAvoidmemory graph structure

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The memory graph is segmented into multiple independent but interconnected sub-graphs, each representing a specific domain or aspect of user memory (e.g., personal information, task history, preferences, contextual data). This segmentation allows the system to manage complexity by treating each sub-graph as a manageable unit while maintaining their relationships through the overall graph structure, thereby improving recall accuracy without overwhelming system complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces intermediary layers and abstraction mechanisms that mediate between the complex memory graph structure and the retrieval operations. These intermediaries include indexing structures, caching mechanisms, and query optimization layers that simplify access to the underlying complex graph, allowing high recall accuracy to be achieved without exposing the full complexity of the memory structure to users or applications.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS11288574B2Systems and methods for building and utilizing artificial intelligence that models human memory

Publication Date: 2022.03.29 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Systems and methods for creating and/or using an artificial intelligence memory system that models human memory are provided. The AI memory system creates and/or uses a user centric memory graph. The user centric memory graph implicitly links memory elements of a user utilizing relationships created in space, time, and cognitive dimensions similar to how the human brain stores and recalls different memory elements. The user centric memory graph is used by searching and/or constraining the user centric memory graph based on a determined user context and/or a user query.