US 8015006

US patent 8015006, titled "Systems and methods for processing natural language speech utterances with context-specific domain agents," was issued on September 6, 2011, from an application filed on May 30, 2008. The original assignee was VoiceBox Technologies Corp, with the current assignee listed as Dialect LLC. The inventors are Robert A. Kennewick, David Locke, Michael R. Kennewick, SR., Michael R. Kennewick, JR., Richard Kennewick, and Tom Freeman.

Abstract:
The patent describes systems and methods for receiving and executing natural language queries and/or commands. It addresses limitations of prior speech query and response systems by offering a comprehensive speech-based environment for information querying, retrieval, presentation, and command. This environment maximizes the use of context, prior information, domain knowledge, and user-specific profile data to create a natural interaction across multiple domains. The system also generates, stores, and utilizes extensive personal profile information for each user to enhance the accuracy of context determination and the relevance of presented results for specific questions or commands.

Plain-Language Overview of Independent Claims:

The patent includes multiple independent claims, each outlining a distinct aspect of the invention.

• Claim 1: This claim describes a computer-implemented method for processing a natural language speech utterance. The method involves several steps:

  1. Receiving a natural language speech utterance from a user.
  2. Accurately recognizing and parsing this utterance, then interpreting it to determine its meaning, which could be a query or a command.
  3. Identifying the appropriate domain of expertise and context required by the interpreted utterance. This involves using a real-time scoring system and invoking relevant software "agents" (which contain domain-specific knowledge and behaviors).
  4. Formulating one or more machine-processable queries or commands based on the interpreted utterance, context, and domain. These may be directed to local or network data sources, or to local or remote devices or the system itself.
  5. Executing these queries or commands asynchronously (meaning they can run at the same time and independently) while managing potential failures gracefully.
  6. Extracting the necessary information from the results returned by the queries, even if the results are in different formats.
  7. Evaluating and combining these results into a single "best" answer, even if the initial results are ambiguous, incomplete, or contradictory. This step uses probabilistic or fuzzy set decision and matching methods.
  8. Formatting the combined result with variable substitutions and transformations to make it easily understandable and natural for the user.
  9. Presenting this formatted result to the user through a text-to-speech engine.
     The entire process accounts for the required domain expertise, the context of the utterance, domain-specific information, the user's interaction history, user preferences, available information sources, and responses received.

• Claim 13: This claim describes a system designed to process natural language speech utterances. The system comprises a computer and software components that work together:

  1. A speech recognition module: This component captures the user's spoken input.
  2. A parser: This module takes the recognized input and transforms it into understandable commands and questions.
  3. A text-to-speech engine module: This converts text into spoken responses.
  4. A network interface module: This allows the computer to connect with one or more networks.
  5. A graphical user interface module: This provides a visual way for users to interact with the system.
  6. An event manager: This component coordinates interactions between other parts of the system and is connected to various other modules.
  7. A dictionary and phrases module: This contains linguistic data for recognition and parsing.
  8. A user profile module: This stores and manages user-specific information.
  9. A personality module: This enables the system to generate responses with simulated human-like characteristics.
  10. An agent module: This contains multiple autonomous "agents" that process and respond to user requests, where each agent provides functionality for a specific domain or application.
  11. An update manager: This module adds new agents or updates existing ones.
  12. One or more databases: These store various types of information used by the system.
      This claim emphasizes that domain-specific behavior and information are organized into agents, which are essentially self-contained packages of executable code, scripts, or links to information. These agents can be updated remotely and can interact with other agents and system resources. The system also includes license management capabilities for these agents, allowing third parties to sell them.

• Claim 16: This claim describes a computer-implemented method similar to Claim 1, but specifically focuses on processing natural language speech commands to control devices. The steps are:

  1. Capturing a user's natural language speech command using accurate speech recognition in real-world environments.
  2. Parsing and interpreting the captured command to understand its meaning.
  3. Determining the domain and context of the command, then invoking appropriate resources, including specific agents.
  4. Formulating one or more device-specific commands for either the system itself or for external devices.
  5. Routing these generated commands to the appropriate systems and/or external devices.
  6. Receiving and processing the results of these executed commands, including any error messages.
  7. Optionally, providing a response to the user, indicating whether the command was successful or failed, and potentially including status information.

Litigation Information:

US Patent 8015006 has been involved in various litigation. Google Patents indicates that the family has litigation, with cases filed in the Texas Western District Court (case 7:25-cv-00061 and 7:25-cv-00060), Virginia Eastern District Court (case 1:23-cv-00581), Washington Western District Court (case 2:26-cv-00810), Texas Eastern District Court (case 2:24-cv-01067), California Northern District Court (case 3:24-cv-04388), and Delaware District Court (case 1:23-cv-00378). Additionally, there have been PTAB cases (IPR2025-01335, which was not instituted due to procedural reasons, and IPR2024-00748, which was not instituted on merits). The Google Patents record also notes "First worldwide family litigation filed."

A search of the CAFC 2026 dockets did not return any specific cases directly involving US patent 8015006 in 2026. The search results provided general CAFC news and summaries of other patent cases from 2026, but not for this specific patent. Therefore, while there is ongoing litigation related to the patent family in various district courts and PTAB, no specific CAFC dockets for 2026 for patent 8015006 were identified in the provided search results.