Privacy-Preserving Natural Language Processing via Intermediary Server Mediation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Intelligent automated assistants lack native configuration to handle private data domains, such as sensitive user-specific or health data, leading to security and efficiency issues when interacting with user requests related to these domains.

Innovation Solution

The system determines whether a natural-language input corresponds to a private domain and either performs a task associated with the input or forgoes task determination based on domain type, ensuring authorized access and reducing power usage by preventing improper task execution.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the digital assistant processes natural language input for private domains (health data, financial data, sensitive user-specific data), then the assistant can provide more comprehensive services, but security and reliability are compromised due to lack of native configuration for these domains

Engineering Contradiction:

Improvecapability to handle private data domainsVSAvoidsecurity of personal information

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The patent introduces an intermediary server that mediates between the digital assistant and private data domains. The server receives natural language input, determines the domain type, and selectively provides process flows for authorized domains while blocking unauthorized access. This intermediary layer enables the assistant to handle private data domains without directly accessing or storing sensitive information, thus maintaining security while expanding capability.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system segments the processing of natural language input into distinct phases: receiving input, determining domain type, and executing appropriate process flows. By segmenting this way, the system can handle different domain types (public vs. private) through different processing paths, allowing comprehensive service coverage for authorized domains while maintaining security through selective execution.

Inventive Principle:

Principle #1Segmentation

2Productivity

If the digital assistant attempts to determine tasks for all domains including private ones, then more tasks can be performed, but computational resources are wasted on unauthorized or improper task execution

Engineering Contradiction:

Improvenumber of tasks that can be performedVSAvoidpower usage for task determination

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary domain type determination before executing any task processing. By first identifying whether the input corresponds to a public domain or a private domain, and only then selecting appropriate process flows, the system avoids wasting computational resources on unauthorized task determination. This preliminary classification action filters out unnecessary processing early in the workflow.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system applies preliminary anti-action by blocking or preventing task determination for unauthorized private domains before any processing occurs. When the domain type is identified as private and the device lacks appropriate configuration or authorization, the system proactively prevents task execution rather than attempting to process it, thereby avoiding wasteful computational effort.

Inventive Principle:

Principle #9Preliminary anti-action

3Adaptability or versatility

If the digital assistant is configured to handle all domain types, then versatility is improved, but unauthorized task execution may occur in private domains

Engineering Contradiction:

Improvedomain type coverageVSAvoidunauthorized access to private data

Core Design Contradiction:

Adaptability or versatilityVSObject-affected harmful factors

Solution Approach 1:

The system applies local quality by treating different domain types differently based on their sensitivity. Public domains are processed with standard task determination, while private domains receive specialized handling through selective process flow provision. This differentiated approach ensures that comprehensive versatility is maintained for authorized domains while unauthorized access to private data is prevented through domain-specific processing rules.

Inventive Principle:

Principle #3Local quality

Data Source

PatentEP3602540B1Client server processing of a natural language input for maintaining privacy of personal information

Publication Date: 2020.08.05 APPLE INC

Figure 1
Figure 2A
Figure 2B

AI summary

Systems and processes for operating an intelligent automated assistant to perform intelligent list reading are provided. In accordance with one example, a method includes, at an electronic device having one or more processors, receiving a natural-language input corresponding to a domain; providing the natural-language input to an external device; receiving, from the external device, a process flow corresponding to the domain; determining, with the process flow corresponding to the domain, a task associated with the natural-language input; performing the task; and providing an output indicating whether the task has been performed.