Google LLC v. IPA Technologies Inc

1. Case Identification

• Case #: IPR2019-00731
• Patent #: 7,069,560
• Filed: February 26, 2019
• Petitioner(s): Google LLC
• Patent Owner(s): IPA Technologies Inc.
• Challenged Claims: 26-35, 47, 48, and 52

2. Patent Overview

• Title: Software-Based Agent Architecture
• Brief Description: The ’560 patent discloses a software-based architecture supporting cooperative task completion by autonomous electronic agents. The system relies on one or more "facilitator" agents to broker communications and cooperation between other agents to satisfy complex user goals.

3. Grounds for Unpatentability

Ground 1: Claims 47 and 52 are obvious over Martin.

• Prior Art Relied Upon: Martin ("Building Distribute Software Systems with the Open Agent Architecture," a 1998 conference proceeding).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Martin, which describes the Open Agent Architecture (OAA), discloses a nearly identical agent-based system to that claimed in the ’560 patent. Martin’s “facilitator” agent coordinates autonomous agents in a distributed environment. It maintains a “knowledge base” of agent capabilities (the claimed “registry”) and decomposes compound goals into subgoals that are allocated to capable agents (the claimed “facilitating engine”). Martin’s Interagent Communication Language (ICL) is described as a conversational protocol with event types and parameter lists that refine event semantics, meeting the final limitations of claim 47. Petitioner asserted that the process steps of claim 52 are the necessary result of operating the system disclosed in Martin and taught by the facilitator agent of claim 47.
  • Motivation to Combine (for §103 grounds): Not applicable as this ground relies on a single reference. Petitioner contended that Martin alone renders the claims obvious, as any minor differences would have been obvious modifications to a POSITA.

Ground 2: Claims 26 and 28-35 are obvious over Martin in view of Farley.

• Prior Art Relied Upon: Martin (1998 conference proceeding) and Farley (Java Distributed Computing, 1998).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner asserted that Martin teaches the core facilitator agent architecture, including the registry and facilitating engine, as argued in Ground 1. Independent claim 26 adds the limitation that the facilitator agent is "functionally distributed across at least two computer processes." Farley, a textbook on distributed computing, explicitly teaches distributing agents across multiple processes to achieve benefits like parallel computation, fault tolerance, and efficient data handling. Petitioner also argued Martin teaches or suggests the elements of dependent claims 28-35, such as using symbolic names, unique addresses, trigger mechanisms (communication, data, and task triggers), and a global database, all of which were well-known concepts in agent-based systems.
  • Motivation to Combine (for §103 grounds): A POSITA would combine Martin’s agent architecture with Farley’s distributed computing techniques to improve system performance, scalability, and reliability. Applying Farley’s teachings to distribute Martin’s facilitator agent was argued to be a predictable combination of known technologies to achieve expected benefits.
  • Expectation of Success (for §103 grounds): A POSITA would have had a high expectation of success, as distributing software components was a standard and well-understood practice for enhancing system architecture at the time.

Ground 3: Claim 27 is obvious over Martin in view of Farley and Pollock.

• Prior Art Relied Upon: Martin (1998 conference proceeding), Farley (Java Distributed Computing, 1998), and Pollock (Patent 5,706,406).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground builds on the Martin-Farley combination from Ground 2. Dependent claim 27 adds the capability for the facilitating engine to modify the goal satisfaction plan during execution based on events like new agent declarations or information from remote agents. While Martin describes a dynamic system where agents can go offline, Pollock explicitly teaches an artificial agent architecture capable of modifying its plan during execution in response to new information.
  • Motivation to Combine (for §103 grounds): A POSITA seeking to make the Martin-Farley distributed agent system more robust and adaptive would have been motivated to incorporate Pollock’s dynamic planning capabilities. This would allow the system to intelligently re-route tasks if, for example, a remote agent became unavailable, which was a known challenge in distributed systems.

Ground 4: Claim 48 is obvious over Martin in view of Farley and White.

• Prior Art Relied Upon: Martin (1998 conference proceeding), Farley (Java Distributed Computing, 1998), and White (Patent 5,603,031).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground addresses dependent claim 48, which specifies that the facilitator agent is "functionally distributed across at least two electronic agents." The combination of Martin and Farley teaches distributing the facilitator across multiple processes. White, which relates to distributed computing, reinforces this by defining an "agent" as a mobile "process."
  • Motivation to Combine (for §103 grounds): A POSITA would combine these references to implement a distributed agent-based system. White’s disclosure that an agent is a type of process, combined with Farley’s teaching of distributing agents across multiple processes, would lead a POSITA to understand that Martin’s facilitator could be functionally distributed across multiple electronic agents (i.e., processes) to gain the known benefits of parallel computation and fault tolerance.

4. Key Technical Contentions (Beyond Claim Construction)

• Prior Art Status of Martin: A central contention of the petition was that the Martin reference is proper prior art under pre-AIA 35 U.S.C. §102(a), contrary to a finding during the patent’s original prosecution. Petitioner argued that Martin was co-authored by D. Martin, A. Cheyer, and D. Moran, while the ’560 patent only names D. Martin and A. Cheyer as inventors. Therefore, Martin represents the work of a different inventive entity. Petitioner asserted that inventor declarations submitted during prosecution to disqualify Martin were legally insufficient as "naked assertions" under Federal Circuit precedent. To support this, Petitioner submitted a new declaration from Dr. Moran, the third co-author of Martin, detailing his inventive contributions to the work described in the paper.

5. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 26-35, 47, 48, and 52 of the ’560 patent as unpatentable.