PTAB

IPR2018-00794

Microsoft Corp v. IPA Technologies Inc

Key Events

Petition

petition

Case #: IPR2018-00794
Patent #: 6,742,021
Filed: March 21, 2018
Petitioner(s): Microsoft Corporation
Patent Owner(s): IPA Technologies, Inc.
Challenged Claims: 2-4, 14, 16-17, 19, 21, 24, 26, 28-30, 43-45, 47-49, 61, 63-64, 66-67, 69, 71, 73-75, 87, 89, 91-93, 106-108, 110-112, 124, and 126

Title: Navigating Network-Based Electronic Information Using Spoken Input with Multimodal Error Feedback
Brief Description: The ’021 patent discloses a system and method for processing a user's spoken natural language request to retrieve information from remote electronic data sources. The system uses a platform of distributed software agents and employs multimodal feedback mechanisms to resolve errors or ambiguities in the interpretation of the spoken request.

Prior Art Relied Upon: Moran (a 1997 publication titled "Multimodal User Interfaces in the Open Agent Architecture").
Core Argument for this Ground:
- Prior Art Mapping: Petitioner asserted that Moran, a publication co-authored by two inventors of the ’021 patent, describes every element of the challenged claims. Moran discloses a multimodal, agent-based system using the Open Agent Architecture (OAA) that accepts spoken natural language commands to access and retrieve information from distributed, remote data sources such as online databases. Petitioner argued Moran's system receives a spoken request, renders an interpretation using speech recognition and natural language understanding agents, and constructs a composite query. To resolve ambiguities, Moran's system solicits additional input from the user and can accept this clarifying input through a different modality (e.g., a pen gesture), thereby refining the query to retrieve the desired information.

Prior Art Relied Upon: Moran (1997 publication) and Kistler (a 1998 publication titled "WebL – a programming language for the Web").
Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that to the extent Moran does not teach interaction with online scripted interfaces, Kistler provides the missing element. Kistler discloses a programming language (WebL) specifically for processing web documents, which it treats as a distributed database. It explicitly teaches extracting structured information by "scraping" webpages, including analyzing online forms to create an input template and using that template to construct a query to retrieve data.
- Motivation to Combine: A POSITA would combine Kistler's specific web data extraction techniques with Moran's broader multimodal agent framework to improve the system's ability to retrieve information from web-based sources. Moran's OAA framework explicitly encourages a "mix-and-match" approach with pre-existing components to facilitate rapid prototyping. Both references address the analogous problem of retrieving information from distributed online sources.
- Expectation of Success: A POSITA would have reasonably expected success in combining these references. The combination would involve supplementing Moran's general data retrieval capabilities with Kistler's well-defined, specific methods for automated web scraping, a predictable integration to enhance a known function.

Prior Art Relied Upon: Moran (1997 publication) and Burns (Patent 5,454,106).
Core Argument for this Ground:
- Prior Art Mapping: Petitioner argued that to the extent Moran is deficient in teaching the iterative refinement of a query, Burns discloses this feature. Burns describes a database retrieval system that handles ambiguous natural language queries by repeating the steps of requesting alternate or additional user input until all parts of the query are understood and can be processed. Its system flowchart expressly depicts loops for re-soliciting user input upon failure to parse a query.
- Motivation to Combine: A POSITA would be motivated to incorporate Burns' detailed error-correction and query refinement loops into Moran's multimodal system to improve its robustness and achieve the shared goal of successful information retrieval. Moran's flexible agent architecture is designed to incorporate such specialized, pre-existing functionalities to improve performance.
- Expectation of Success: Success would be predictable, as the combination involves applying Burns' established techniques for resolving ambiguity in database queries to Moran's agent-based system. This represents a straightforward application of a known error-correction technique to a similar system to achieve an improved, predictable result.

Petitioner noted that if the Board requires claim construction, it should adopt the interpretations proposed by the Patent Owner in a related district court litigation to streamline the proceeding and focus on the prior art. Key proposed constructions include:
- "navigation query": Interpreted as "an electronic query, form, series of menu selections, or the like; being structured appropriately so as to navigate a particular data source of interest in search of desired information."
- "rendering an interpretation of the spoken request": Interpreted as "determining a meaning of the spoken request using a computing device, such as that provided by extracting speech data...and linguistically parsing the speech data."

Petitioner requested the institution of an inter partes review of claims 2-4, 14, 16-17, 19, 21, 24, 26, 28-30, 43-45, 47-49, 61, 63-64, 66-67, 69, 71, 73-75, 87, 89, 91-93, 106-108, 110-112, 124, and 126, and that those claims be found unpatentable and canceled.