Speech Processing Device Feedback Loop for Task Accuracy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Electronic devices often misinterpret user speech, leading to unintended tasks being executed or multiple tasks being performed incorrectly due to limitations in natural language understanding (NLU) processing.

Innovation Solution

An electronic device with a processor, memory, touchscreen display, microphone, and wireless communication circuit that sends user inputs to an external server for processing, allowing for accurate task execution and cancellation based on received responses, enabling improved NLU and task management.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If natural language understanding (NLU) processing is used to recognize and execute user speech commands, then the electronic device can perform tasks based on user intent, but the device may misinterpret speech and execute unintended tasks or multiple tasks incorrectly

Engineering Contradiction:
Improveautomatic task execution based on speech recognitionVSAvoidaccuracy of task execution
Core Design Contradiction:
Extent of automationVSReliability

Solution Approach 1:

The patent implements a feedback mechanism where the electronic device provides speech recognition results and task execution status back to the user through the user interface. The user can review recognized speech, select intended tasks from multiple options, and confirm or correct task execution. This feedback loop allows the system to adjust and verify task execution accuracy, resolving the contradiction between automated speech execution and execution reliability.

Inventive Principle:
Principle #23Feedback

2Adaptability or versatility

If multiple tasks are performed based on user speech, then the electronic device can handle complex user requests, but unintended tasks may be performed due to misinterpretation

Engineering Contradiction:
Improveability to perform multiple tasksVSAvoidaccuracy of task execution
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent applies partial action by executing only the tasks that the user confirms as intended, rather than automatically executing all recognized tasks. When multiple tasks are identified from user speech, the system presents them to the user for selection and confirmation. This allows the device to handle complex multi-task requests while maintaining reliability by executing only the verified intended tasks, filtering out unintended ones.

Inventive Principle:
Principle #16Partial or excessive action

3Ease of operation

If speech recognition and NLU processing are implemented, then the electronic device can receive natural language inputs, but the device may not operate as the user desires when identified commands differ from intended commands

Engineering Contradiction:
Improvenatural language input capabilityVSAvoidalignment between user intent and device action
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The system provides feedback by displaying recognized speech and identified tasks to the user through the user interface, allowing the user to verify whether the device correctly understood their intent. The user can then confirm or correct the identified tasks before execution. This feedback mechanism bridges the gap between natural language input ease and reliable intent execution.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The user interface acts as an intermediary between speech recognition and task execution. Instead of directly executing recognized commands, the system uses the user interface to present recognized speech and identified tasks, allowing the user to mediate and verify the interpretation before final execution. This intermediary layer ensures that ease of natural language input does not compromise execution accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3612925B1Electronic device and method for processing user speech
Publication Date: 2023.12.13 SAMSUNG ELECTRONICS CO LTD
  • EP3612925B1 patent drawingFigure 1
  • EP3612925B1 patent drawingFigure 2
  • EP3612925B1 patent drawingFigure 3

AI summary

An electronic device and method are provided for processing user speech. The electronic device includes a housing, a touchscreen display disposed in the housing and exposed through a first portion of the housing, a microphone disposed in the housing and exposed through a second portion of the housing, at least one speaker disposed in the housing and exposed through a third portion of the housing, a wireless communication circuit disposed in the housing, at least one processor disposed in the housing and electrically connected with the touchscreen display, the microphone, the speaker, and the wireless communication circuit, and a memory disposed in the housing and electrically connected with the processor, wherein the memory stores instructions executed to enable the at least one processor to receive a first user input through the touchscreen display or the microphone, the first user input including a request to perform tasks using the electronic device, send data associated with the first user input through the wireless communication circuit to an external server, receive a response from the external server through the wireless communication circuit, the response including information about a plurality of states of the electronic device to perform the tasks in an order, perform the tasks by causing the electronic device to have the plurality of states having the order after receiving the response, receive a second user input through the touchscreen display or the microphone, the second user input including a request to cancel at least one of the performed tasks, and cause the electronic device to return to one of the plurality of states, based on the second user input.