PTAB

IPR2024-00747

Google LLC v. Dialect LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Systems and Methods for Responding to Natural Language Speech Utterance
  • Brief Description: The ’738 patent describes a system for processing natural language speech commands and queries. The system uses an agent-based architecture where a parser selects a domain-specific agent to handle a request based on the context of a user's utterance, and an update manager allows for the purchase of new agents.

3. Grounds for Unpatentability

Ground 1: Claims 1-11, 18, 20-23, 26, 31-37, and 39 are obvious over Coffman in view of Kanevsky and Ronning.

  • Prior Art Relied Upon: Coffman (International Publication No. WO 00/20962), Kanevsky (Patent 5,897,616), and Ronning (Patent 9,817,650).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Coffman discloses the core framework of the claimed system, including a conversational computing system with an "agent architecture" (comprising "conversationally aware applications"), a "parser" (its Natural Language Understanding or NLU system), and an "event manager" (its Conversational Virtual Machine or CVM that mediates interactions). Petitioner argued Coffman’s NLU system determines utterance context to select an appropriate application. To meet the claim limitation of using keywords to determine context, Petitioner combined Coffman with Kanevsky, which teaches implementing NLU using "a combination of classical parsing capabilities with key word spotting." To satisfy the "update manager" limitation for purchasing agents from a third party, Petitioner combined the system with Ronning, which describes a software agent for purchasing, downloading, and managing software applications from an e-commerce system.
    • Motivation to Combine: A POSITA would combine Coffman and Kanevsky because Coffman expressly incorporates Kanevsky by reference, they share a common inventor, and both were assigned to IBM, making them part of a collective body of work. A POSITA would combine Ronning to add a well-known e-commerce and software management functionality to Coffman's system, which already disclosed installing and upgrading applications. Petitioner argued this combination would improve Coffman’s system and provide a known commercial method to incentivize the third-party developers for whom Coffman already provided development tools.
    • Expectation of Success: A POSITA would have had a high expectation of success, as the combination involved applying known, conventional techniques (keyword spotting from Kanevsky, software purchasing agents from Ronning) to improve the functionality of a base system (Coffman) in a predictable manner.

Ground 2: Claims 16, 17, 25, 30, and 38 are obvious over the Ground 1 Prior Art in further view of Zadrozny.

  • Prior Art Relied Upon: Coffman (WO 00/20962), Kanevsky (Patent 5,897,616), Ronning (Patent 9,817,650), and Zadrozny (a 1994 publication titled “Natural Language Understanding with a Grammar of Constructions”).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground builds on the combination in Ground 1 and adds Zadrozny to address the limitations of claim 16, which requires the parser to formulate a request "in accordance with a grammar" used by the selected domain agent. Petitioner argued that while Coffman discloses that its applications provide their own grammars, Zadrozny provides the explicit teaching of using a "grammar of constructions" to analyze a recognized utterance. Zadrozny’s system formulates the utterance as an intended action with specific parameters (e.g., schedule a meeting at 5 PM) and translates it into a form that is executable by the application, directly teaching the claimed grammar-based request formulation.
    • Motivation to Combine: The motivation to combine the Ground 1 references applies here as well. A POSITA would be further motivated to incorporate Zadrozny because Kanevsky expressly cites Zadrozny as a primary example of a "semantic analyzer." Petitioner contended that a POSITA seeking to implement the semantic analysis and parsing capabilities described at a high level in Coffman and Kanevsky would have been directly motivated to consult Zadrozny for its detailed, working implementation of grammar-based NLU. The shared IBM affiliation between the authors of Zadrozny, Coffman, and Kanevsky reinforces this motivation.
    • Expectation of Success: A POSITA would expect the combination to succeed because Zadrozny describes a functional, implemented technique for grammar-based parsing that directly addresses and provides a detailed solution for the high-level concepts disclosed in the Coffman/Kanevsky system.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued against discretionary denial under 35 U.S.C. §314(a) by noting that the parallel district court litigation is in a very early stage. The case was recently transferred, and as of the petition filing, no discovery, claim construction, or trial date had occurred, weighing against denial under the Fintiv factors.
  • Petitioner also argued against denial under 35 U.S.C. §325(d), asserting that while a U.S. counterpart to the primary reference (Coffman) was listed on an Information Disclosure Statement during prosecution, it was one of over 100 references and was not substantively considered by the Examiner. Critically, the Examiner did not have the benefit of Kanevsky, Ronning, or Zadrozny, which Petitioner contended are necessary to provide context and demonstrate the compelling obviousness of the challenged claims.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-11, 16-18, 20-23, 25, 26, and 30-39 of the ’738 patent as unpatentable.