PTAB

IPR2018-00475

Google LLC v. IPA Technologies Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Speech-Based Navigation of Electronic Data
  • Brief Description: The ’021 patent relates to systems and methods for navigating electronic data using spoken natural language requests. The technology leverages an Open Agent Architecture (OAA) with multiple autonomous software "agents" and a "facilitator" agent to process requests, resolve ambiguities, and retrieve information.

3. Grounds for Unpatentability

Ground 1: Obviousness over Cheyer and Shwartz - Claims 1, 8, 72, and 79 are obvious over Cheyer in view of Shwartz.

  • Prior Art Relied Upon: Cheyer (“Multimodal Maps: An Agent-based Approach,” a 1995 conference paper) and Shwartz (Patent 5,197,005).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Cheyer, a paper co-authored by inventors of the ’021 patent, discloses the core framework of the invention. Cheyer describes a multimodal system using the same OAA as the ’021 patent for speech-based navigation of remote data sources like the World Wide Web. It discloses receiving a spoken request, using speech recognition (SR) and natural language (NL) agents to interpret it, and using a facilitator to delegate tasks. Petitioner contended that while Cheyer discloses using database requests, it lacks detail on constructing them. Shwartz, which addresses database retrieval via a natural language interface, was argued to supply this missing detail by teaching the generation of a structured database query (e.g., SQL) from a natural language input to retrieve data. Claim 8, which adds that soliciting additional input occurs in response to a deficiency, was argued to be taught by Cheyer’s disclosure of prompting a user when a request is ambiguous.
    • Motivation to Combine: A POSITA would combine Cheyer with Shwartz to implement the general database request functionality of Cheyer with the specific query construction details from Shwartz. Both references are in the same technical field of natural language interfaces for data retrieval, making the combination a predictable solution to a known problem.
    • Expectation of Success: A POSITA would have a high expectation of success because combining Shwartz’s well-understood query generation methods with Cheyer’s agent-based architecture was a straightforward implementation involving known technologies to achieve the predictable result of improved data retrieval.

Ground 2: Obviousness over Cheyer, Shwartz, and Johnson - Claims 1 and 72 are obvious over Cheyer and Shwartz in view of Johnson.

  • Prior Art Relied Upon: Cheyer, Shwartz, and Johnson (Patent 5,748,974).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground builds on the Cheyer and Shwartz combination to address the limitation of soliciting additional non-spoken input without requiring the user to request that modality. Petitioner argued that to the extent Cheyer does not explicitly teach this, Johnson does. Johnson, which is directed to a multimodal natural language interface, explicitly discloses resolving ambiguity by presenting a user with choices in a pop-up window (a non-spoken modality) and asking the user to "select one." Johnson's system does this automatically in response to ambiguity without requiring the user to first request the pop-up window. This was argued to teach the claimed method of soliciting additional input in a different, non-spoken modality to refine a query.
    • Motivation to Combine: A POSITA would be motivated to incorporate Johnson’s user-friendly ambiguity resolution technique into the Cheyer/Shwartz system. Providing a pop-up with choices is a simple and convenient way to clarify user intent, consistent with Cheyer's goal of creating a more natural user interface.
    • Expectation of Success: The integration would be predictable, as implementing a pop-up window for clarification was a well-known user interface design choice for handling ambiguous query results.

Ground 3: Obviousness over Cheyer, Shwartz, and Kistler - Claims 3, 4, 74, and 75 are obvious over Cheyer and Shwartz in view of Kistler.

  • Prior Art Relied Upon: Cheyer, Shwartz, and Kistler (“WebL—a programming language for the Web,” a 1998 journal article).

  • Core Argument for this Ground:

    • Prior Art Mapping: This ground addresses dependent claims requiring the step of constructing a navigation query to include "extracting an input template for an online scripted interface" and using it to build the query. Petitioner asserted that Kistler discloses WebL, a scripting language for automating web-related tasks, including extracting data from HTML pages and filling in web-based input forms. This process of extracting the structure of a web form to automate interaction was argued to be equivalent to extracting an input template. Claim 4’s limitation of “dynamically scraping” the interface was argued to be taught by Kistler’s disclosure of using an automated scripting language to extract information in response to a user query.
    • Motivation to Combine: A POSITA implementing the Cheyer/Shwartz system to access web data sources would be motivated to use a tool like Kistler's WebL. Since many websites in the relevant time period used scripted interfaces (e.g., CGI scripts), automating the process of extracting form data would be necessary for seamless agent-based interaction, improving efficiency and reliability.
    • Expectation of Success: Success would be expected, as Kistler provides an explicit tool (WebL) for the express purpose of automating interaction with web pages and forms, a straightforward programming task for a POSITA.

  • Additional Grounds: Petitioner asserted additional obviousness challenges based on combinations including Thrift (for performing interpretation locally), Wyard (for handling unresolved words and mixed-initiative dialogue), and Burns (for iterative query refinement), but relied on similar design modification theories.

4. Key Claim Construction Positions

  • Petitioner addressed the term "navigation query," which is central to the independent claims. For the purposes of the petition, Petitioner adopted the Patent Owner’s proposed construction from related district court litigation: “an electronic query, form, series of menu selections, or the like; being structured appropriately so as to navigate a particular data source of interest in search of desired information.” This strategic choice aimed to show that the claims were obvious even under the Patent Owner’s own interpretation of the key term.

5. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 3, 4, 6, 9, 14, 21, 74, 75, 77, and 80 of the ’021 patent as unpatentable.