Google LLC v. Dialect LLC

1. Case Identification

• Case #: IPR2024-00749
• Patent #: 7,693,720
• Filed: April 5, 2024
• Petitioner(s): Google LLC
• Patent Owner(s): Dialect, LLC
• Challenged Claims: 1-16, 18-25, 27, 29-43, 52, 55

2. Patent Overview

• Title: Mobile Systems and Methods For Responding To Natural Language Speech Utterance
• Brief Description: The ’720 patent describes a mobile system, typically implemented in a vehicle, that processes natural language speech from a user. The system uses a modular architecture with a plurality of "domain agents" to interpret the user's utterance, select an appropriate function or application, and generate a response.

3. Grounds for Unpatentability

Ground 1: Obviousness over Core Conversational System Art - Claims 1-16, 22-25, 27, 29-31, 35-43, 52, and 55 are obvious over Coffman, Kanevsky, Kneser, and Zadrozny.

• Prior Art Relied Upon: Coffman (International Publication No. WO 00/20962), Kanevsky (Patent 5,897,616), Kneser (Patent 6,157,912), and Zadrozny (a 1994 publication).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Coffman disclosed the core architecture of the challenged claims, including a mobile "conversational car PC" system that receives natural language speech and uses a natural language processing system with multiple "conversationally-aware applications" (the claimed "domain agents"). The system included a speech recognition engine and a parser. To implement the well-known components of this system, Petitioner asserted a person of ordinary skill in the art (POSITA) would have looked to other references. Kanevsky was cited for its teaching of using trigram language models for speech recognition, which represent the claimed "plurality of dictionary and phrase entries." Kneser was cited for teaching that these language models could be "adaptive"—dynamically updated based on user interaction and dialog history. Finally, Zadrozny was used to show the obviousness of using an application-specific "grammar of constructions" to parse a user's utterance and formulate a command for the selected application to execute.
  • Motivation to Combine: A POSITA would combine these references as they represent a collective body of work from the same field and, in large part, the same assignee (IBM). Coffman expressly incorporated Kanevsky by reference, and Kanevsky cited Zadrozny. The shared inventorship and assignment provided a strong motivation for a POSITA implementing Coffman's system to consult the other references for known techniques to realize its functions, such as using n-gram models for speech recognition (Kanevsky), adapting those models (Kneser), and using grammars for parsing (Zadrozny). The combination was presented as simply applying known techniques to improve a known system.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success because the combination involved applying conventional and well-understood technologies for speech recognition and natural language processing to the system described in Coffman.

Ground 2: Obviousness with In-Vehicle Audio Enhancement - Claims 18, 20-21, and 32-33 are obvious over the Ground 1 Prior Art in further view of DeLine.

• Prior Art Relied Upon: Coffman, Kanevsky, Kneser, Zadrozny, and DeLine (Patent 6,420,975).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground built upon the combination in Ground 1. Petitioner argued that claims 18, 20-21, 32, and 33 add limitations related to specific hardware for improving audio quality in a vehicle: a speech coder, an "array microphone," and a "filter that optimizes a signal to noise ratio." While Coffman mentioned acoustic signal processing, DeLine was cited for explicitly teaching an in-vehicle system that used a microphone array and a digital sound processor (a filter) to perform noise cancellation techniques. This enhanced the vocal content and reduced ambient noise, which is a known problem in vehicle environments.
  • Motivation to Combine: A POSITA would combine DeLine with the Ground 1 art because Coffman's system was designed for a vehicle environment, and DeLine addressed the well-known problem of degraded speech input in such noisy settings. Implementing DeLine’s microphone array and filtering techniques in Coffman's system was argued to be an obvious design choice to improve its performance and reliability for the intended use case.
  • Expectation of Success: A POSITA would have expected success in integrating DeLine's noise-cancellation technology into Coffman's system, as it was a known solution for a known problem.

Ground 3: Obviousness with Adaptive Echo Cancellation - Claims 19 and 34 are obvious over the Ground 2 Prior Art in further view of Gerson.

• Prior Art Relied Upon: Coffman, Kanevsky, Kneser, Zadrozny, DeLine, and Gerson (Patent 6,937,977).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground added Gerson to the combination of Ground 2 to address limitations requiring the filter to employ "adaptive echo cancellation." While DeLine taught general noise cancellation, including echo cancellation, Gerson specifically disclosed an adaptive echo cancellation system for "barge-in" processing. This allows a user to speak over the system's audio output, which is a critical feature for natural-feeling conversational interfaces.
  • Motivation to Combine: A POSITA would combine Gerson's teachings because the goal of Coffman's system was to provide a natural, conversational interaction. Allowing for "barge-in" through adaptive echo cancellation, as taught by Gerson, directly served this goal. It was argued to be a known and advantageous feature for any interactive speech system, making its inclusion in the Coffman/DeLine system an obvious improvement.
  • Expectation of Success: The combination involved applying a conventional technique (adaptive echo cancellation) to improve a speech-based system, for which a POSITA would have had a reasonable expectation of success.

4. Arguments Regarding Discretionary Denial

• Petitioner argued against discretionary denial under 35 U.S.C. §314(a), noting that the co-pending district court litigation was in its early stages with no trial date set, discovery yet to occur, and no claim construction.
• Petitioner also argued against denial under 35 U.S.C. §325(d), acknowledging that a U.S. counterpart to the Coffman reference was before the Examiner on an Information Disclosure Statement (IDS) but was never discussed or applied. Petitioner contended that the Examiner also did not have the benefit of Kanevsky, Kneser, or Zadrozny, which provide essential context and detail for the obviousness arguments.

5. Relief Requested

• Petitioner requests the institution of an inter partes review and the cancellation of claims 1-16, 18-25, 27, 29-43, 52, and 55 of the ’720 patent as unpatentable.