PTAB

IPR2025-01333

Meta Platforms Inc v. Dialect LLC

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Processing Speech and Non-Speech Communications
  • Brief Description: The ’039 patent discloses methods for processing combined speech and non-speech user inputs to retrieve information or process commands. The system uses personal profile information and context determination to improve the reliability of interpreting a user's intent and presenting expected results.

3. Grounds for Unpatentability

Ground 1: Claims 13-15 and 17-18 are obvious over Maes in view of Ross.

  • Prior Art Relied Upon: Maes (Patent 6,964,023) and Ross (Application # 2002/0133354).
  • Core Argument for this Ground:

    • Prior Art Mapping: Petitioner argued that Maes taught a multi-modal conversational computing system that meets the initial steps of independent claim 13. Specifically, Maes disclosed a system that receives and processes both speech (audio input) and non-speech (video input, e.g., lip movement) communications. Maes further taught transcribing these inputs to create separate data streams (a decoded text script and a visual viseme sequence) and then merging or synchronizing them to generate a generalized query representing the user's intent. Petitioner contended that while Maes taught the basic framework, it did not explicitly disclose comparing the resulting query against specific application grammars or using a relevance score to select a target application.

      Petitioner asserted that Ross supplied these missing elements. Ross taught a system for determining the context of a user's spoken command by comparing it against multiple "context description grammars," where each grammar is associated with a specific application (a "domain agent"). Ross explicitly described testing a recognized utterance against entries in these grammars to find a match. Furthermore, Ross disclosed maintaining a prioritized "context list" of applications, where priority is determined by a relevance score based on recency of access. This score is used to select the correct domain agent when an utterance could be accepted by multiple grammars. The combination, therefore, allegedly taught every limitation of claim 13. Dependent claims 14 (aggregate response), 15 (follow-up query), 17 (context stack), and 18 (generating contexts using fuzzy possibilities) were also argued to be obvious, as Maes and Ross were said to teach these additional features either individually or in combination.

    • Motivation to Combine: Petitioner presented several motivations for a person of ordinary skill in the art (POSITA) to combine Maes and Ross. The primary motivation argued was to improve the context identification capabilities of Maes's multi-modal system by incorporating Ross's more advanced technique of using application-specific context grammars and a prioritized context list. Petitioner argued that both references were analogous art aimed at solving the same problem: reliably interpreting user utterances to invoke the correct computer function. Maes provided a robust multi-modal input system, while Ross provided a sophisticated method for disambiguating intent among multiple speech-enabled applications. A POSITA would combine Ross's grammar-based context determination with Maes's system to enhance its accuracy and allow it to more effectively manage commands across various registered applications.

    • Expectation of Success: Petitioner argued a POSITA would have had a reasonable expectation of success in combining the references. The integration was presented as straightforward, involving the application of Ross's known grammar techniques to Maes's similar system architecture. Both systems were described as software-based, utilizing conventional, commercially-available speech recognition components, and did not require specialized hardware. Implementing Ross's context grammars and priority list would be a predictable modification to the grammar database and context stack already present in Maes's system, leading to the expected result of an improved, more accurate conversational computing system.

4. Arguments Regarding Discretionary Denial

  • Petitioner argued that the Board should not exercise its discretion to deny institution. The petition asserted that review should be instituted based on efficiency and fairness considerations and the "overwhelming strength on the merits" of the asserted unpatentability ground, citing precedential authority in Apple Inc. v. Fintiv, Inc.

5. Relief Requested

  • Petitioner requests the institution of an inter partes review and the cancellation of claims 13-15, 17-18 of the ’039 patent as unpatentable.