US 11929073

Summary of U.S. Patent 11,929,073

Title: Hybrid arbitration system

Assignee: Cerence Operating Co.

Inventor: Min Tang

Filing Date: October 3, 2022

Issue Date: March 12, 2024

Abstract:
A method for selecting a speech recognition result on a computing device includes receiving a first speech recognition result determined by the computing device, receiving first features, at least some of the features being determined using the first speech recognition result, determining whether to select the first speech recognition result or to wait for a second speech recognition result determined by a cloud computing service based at least in part on the first speech recognition result and the first features.

---

Overview of Independent Claims

U.S. Patent 11,929,073 has three independent claims: Claim 1, Claim 14, and Claim 15.

Claim 1: A method for selecting a speech recognition result

In plain language, this claim describes a process performed on a user's device (like a smartphone or car infotainment system). When a person speaks a command, the device itself tries to understand it (the "first speech recognition result") and at the same time, sends the voice data to a more powerful cloud service for a second opinion. This patented method is about deciding, very quickly, whether the device's own interpretation is good enough to act on immediately, without waiting for the cloud's response. This decision is made by analyzing the device's result and associated data ("first plurality of features"). If the device's result is deemed reliable enough, it is selected and used right away, which saves time.

Claim 14: A system for selecting a speech recognition result

This claim covers the physical hardware that performs the method described in Claim 1. It outlines a system that includes:

• An input to capture the user's speech.
• An output to send the speech data to both the local device's processor and a cloud service.
• A second input to receive the results from the local processor.
• One or more processors that are programmed to analyze the local result and decide whether to use it immediately or wait for the cloud's result, and then select the local result if it meets the criteria.

Essentially, this claim protects the tangible components that make up the "hybrid arbitration system."

Claim 15: Software for selecting a speech recognition result

This claim protects the software that enables the process. It describes a non-transitory, computer-readable medium (like a memory chip) that stores instructions. When these instructions are run by a processor, they cause the device to perform the same steps outlined in Claim 1: get the speech data, send it to the local processor and the cloud, receive the local result, and then decide whether to use the local result right away or wait for the cloud's more detailed analysis. This claim ensures that the computer program itself, which is the core of the invention, is also protected.

Litigation Status

As of the current date, U.S. Patent 11,929,073 is the subject of litigation. Court records indicate that Cerence Operating Company has filed a patent infringement lawsuit against Amazon.com, Inc., Amazon.com Services LLC, and Amazon Web Services, Inc. in the U.S. District Court for the Eastern District of Texas.

A search of the CAFC (Court of Appeals for the Federal Circuit) 2026 dockets did not reveal any appeals related to this case at this time. However, district court proceedings are ongoing.