Google LLC v. IPA Technologies Inc

1. Case Identification

• Case #: IPR2019-00728
• Patent #: 6,851,115
• Filed: February 26, 2019
• Petitioner(s): Google LLC
• Patent Owner(s): IPA Technologies Inc.
• Challenged Claims: 61-70

2. Patent Overview

• Title: Software Architecture for Cooperative Agents
• Brief Description: The ’115 patent describes a software-based architecture for constructing distributed systems. The architecture uses autonomous electronic agents that cooperate to complete tasks, with communication and coordination brokered by one or more "facilitator" agents.

3. Grounds for Unpatentability

Ground 1: Obviousness over Martin and Kohn - Claims 61 and 63-70 are obvious over Martin in view of Kohn.

• Prior Art Relied Upon: Martin ("Building Distributed Software Systems with the Open Agent Architecture," a 1998 conference proceeding) and Kohn (Patent 6,088,689).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that Martin, which describes an Open Agent Architecture (OAA), disclosed nearly all limitations of independent claim 61. Martin’s OAA used a central "facilitator" to coordinate "cooperative task completion" among distributed agents. It taught an agent registry (a "knowledge base"), a facilitating engine to parse requests and construct a "goal satisfaction plan," and an Interagent Communication Language (ICL) with conversational and content layers. For the limitation in claim 61 requiring interpretation of a compound goal with "both local and global constraints," Petitioner asserted that while Martin taught constraints generally, Kohn explicitly taught a multi-agent architecture that used both local and global constraints to govern agent behavior and system goals.
  • Motivation to Combine: Petitioner contended a person of ordinary skill in the art (POSITA) would combine Kohn with Martin because both relate to distributed, multi-agent architectures. A POSITA would have been motivated to modify Martin's system to incorporate Kohn’s use of both local and global constraints to provide greater flexibility and control in fulfilling complex service requests. This would allow the system to manage goals with varying scopes and dependencies, which was a known objective in the field.
  • Expectation of Success: The proposed combination involved applying a known type of constraint (local and global constraints from Kohn) to a known system architecture (Martin's OAA). Petitioner argued this was a straightforward integration of known elements for their intended purpose, leading to a predictable improvement in system flexibility.

Ground 2: Obviousness over Martin, Kohn, and Pollock - Claim 62 is obvious over Martin in view of Kohn and Pollock.

• Prior Art Relied Upon: Martin (a 1998 conference proceeding), Kohn (Patent 6,088,689), and Pollock (Patent 5,706,406).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground challenged claim 62, which depends on claim 61 and adds the limitation that the facilitating engine is capable of modifying the goal satisfaction plan during execution. Petitioner relied on the Martin and Kohn combination from Ground 1 to teach the base facilitator agent and asserted that Pollock supplied the missing element. Pollock described an architecture for an artificial agent capable of "planning and executing plans" and expressly taught that a plan could be modified during execution in response to new information. Pollock’s "TAIL-MONITOR" component was responsible for replacing the tail end of a plan with a modified version based on events that occur after execution has begun.
  • Motivation to Combine: Petitioner argued that Pollock was in the same technical field of agent-based planning and execution. A POSITA implementing the Martin/Kohn system would have been motivated to incorporate Pollock’s dynamic plan modification to improve the system's robustness and adaptability. For example, Martin acknowledged that agents could go offline; Pollock’s teachings would provide a known method to modify an existing plan to re-route tasks from a departed agent to an active one, a clear and desirable improvement.
  • Expectation of Success: A POSITA would have had a reasonable expectation of success in integrating Pollock's dynamic modification capability into the Martin/Kohn agent framework. The concept of monitoring execution and adapting plans based on new events was a known technique to make distributed systems more resilient, and its implementation was considered predictable.

4. Key Technical Contentions (Beyond Claim Construction)

• Prior Art Status of Martin: A central contention of the petition was that the Martin reference was improperly disqualified as prior art during the original prosecution of the ’115 patent. The applicants had submitted inventor declarations averring that the subject matter of Martin was their own invention, and that a third co-author, Dr. Douglas Moran, was not a co-inventor of the claimed subject matter. Petitioner argued these were conclusory, uncorroborated declarations. To counter them, Petitioner submitted a new declaration from Dr. Moran himself, stating that he contributed to the conception of core networking and distributed agent concepts described in Martin and led the project. Petitioner argued that this new evidence established that Martin described the work of a different inventive entity than that of the ’115 patent, making it valid prior art under pre-AIA 35 U.S.C. §102(a).

5. Relief Requested

• Petitioner requested institution of an inter partes review and cancellation of claims 61-70 of the ’115 patent as unpatentable.