UiPath Inc v. Rule 14 LLC

1. Case Identification

• Case #: IPR2024-01304
• Patent #: 9,229,977
• Filed: September 16, 2024
• Petitioner(s): UiPath, Inc.
• Patent Owner(s): Rule 14 LLC.
• Challenged Claims: 1-21

2. Patent Overview

• Title: Data Mining System
• Brief Description: The ’977 patent discloses a data mining method where a user generates and expands a query, executes it across multiple data sources, and selects at least one source for scheduled monitoring based on the results meeting an accuracy threshold. The system then extracts updated data from the monitored source and establishes a communication channel to provide it to a second user.

3. Grounds for Unpatentability

Ground 1: Obviousness over Kirsch, Eichstaedt, and Lucovsky (Claims 1, 2, 4-6, 8, 10, 11, 13-15, 17, 19-20)

• Prior Art Relied Upon: Kirsch (Patent 6,018,733), Eichstaedt (Patent 7,765,228), and Lucovsky (Patent 7,844,603).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that the combination of these references taught all limitations of the challenged claims. Kirsch, addressing the "collection selection" problem, disclosed generating a user query, expanding search terms, executing the query against multiple "document collections" (data sources), and selecting a "qualified set" of collections that meet a "minimum level of relevance," which Petitioner mapped to the claimed "accuracy thresholds." Eichstaedt taught the monitoring element by disclosing a "poller server" that "wakes up after a predefined period" to apply queries to new content and deliver email alerts. Lucovsky supplied the final limitation by disclosing mechanisms for a first user to share search results with a second user via email, message board, or blog.
  • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine these references because they all addressed search query technology. It would have been obvious to integrate Eichstaedt's periodic monitoring and alert functionality into Kirsch's system to provide users with updates from the selected relevant data sources. Furthermore, a POSITA would have been motivated to add Lucovsky's well-known sharing features to enable collaboration and dissemination of the monitored search results.
  • Expectation of Success: The combination involved applying conventional monitoring and sharing functions to a standard search and selection system, which would have resulted in a predictable and successful system.

Ground 2: Obviousness over Learning Lexis, Kirsch, and Lucovsky (Claims 1, 2, 5, 6, 8, 10, 11, 14, 15, 17, 19-21)

• Prior Art Relied Upon: Learning Lexis (a 2001 lexis.com guide), Kirsch (Patent 6,018,733), and Lucovsky (Patent 7,844,603).
• Core Argument for this Ground:
  • Prior Art Mapping: This ground substituted the primary reference Kirsch with Learning Lexis, a guide Petitioner established as prior art. Learning Lexis taught a system where a user generates a query, which can be expanded with suggested terms or equivalents. The system executes the query across a hierarchy of data sources (e.g., "Combined Federal & State Case Law"). Critically, its "ECLIPSE Feature" allowed users to automatically save and re-run searches at periodic intervals and receive updates via email, teaching the monitoring and communication elements. Petitioner argued that Kirsch was necessary to teach the selection of a subset of data sources from the broader hierarchy based on relevance before monitoring. Lucovsky was again relied upon for sharing the monitored results with a second user.
  • Motivation to Combine: A POSITA using the Learning Lexis system, which could search very broad collections, would have been motivated to incorporate Kirsch's established "collection selection" techniques. This would allow a user to narrow the scope of the ongoing "ECLIPSE" monitoring to only the most relevant jurisdictions or sources identified in an initial broad search, thereby improving efficiency and the relevance of updates. Adding Lucovsky's sharing features was a simple, obvious improvement for collaboration.
  • Expectation of Success: Applying known collection-narrowing techniques from Kirsch to the established monitoring feature in Learning Lexis would have been a straightforward implementation with a high expectation of success.

Ground 3: Obviousness over Liang and Lucovsky (Claims 1, 3, 4, 10, 12, 13, 19)

• Prior Art Relied Upon: Liang (Application # 2006/0106793) and Lucovsky (Patent 7,844,603).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner asserted that Liang alone taught most limitations of claim 1. Liang disclosed an "Intelligent Search Engine" that accepted a natural language query from a user, expanded keywords with synonyms, and executed the query on multiple sources. Liang provided statistics ("Stats") on the number of results from each source, allowing the system to automatically select the top "S" sources based on this data, which Petitioner mapped to selecting sources based on an "accuracy threshold." Liang explicitly taught monitoring these sources for changes on a set frequency and alerting the user via email. Lucovsky was then combined to teach the limitation of sharing these email alerts with a second user.
  • Motivation to Combine: Both references related to search query technology. A POSITA would have found it obvious and desirable to incorporate Lucovsky's user-to-user sharing functionality into Liang's automated search and alert system to allow for easy dissemination of important updates among a team or colleagues.
  • Expectation of Success: Integrating a conventional email sharing feature into a system that already generates email alerts was a simple and predictable modification.
• Additional Grounds: Petitioner asserted numerous additional obviousness challenges, primarily substituting or adding references to the core combinations above. These included using Mellin (WO 01/65336) for group email functionality, Schultz (Patent 5,721,902) for query expansion with synonyms or recursive searching, and Vinsonneau (Patent 7,043,482) for dynamically scaling data extractors based on traffic volume.

4. Key Claim Construction Positions

• "accuracy thresholds": Petitioner argued that this term should be construed to encompass a "relevancy threshold." This construction was based on the patent's specification equating accuracy with relevancy ("the user may judge the results for accuracy based on the intent of the query") and the Patent Owner's own infringement contentions. This interpretation was critical for mapping prior art like Kirsch, which selected data sources based on a "minimum level of relevance."

5. Arguments Regarding Discretionary Denial

• Petitioner argued that discretionary denial under §314(a) is not warranted. Citing the USPTO's Interim Procedure guidance, Petitioner stipulated that, if review is instituted, it will not pursue in the parallel district court litigation the same grounds raised in the petition or any grounds that could have reasonably been raised. Petitioner asserted that this stipulation ensures no overlap between the IPR and the district court case. Further, it was argued that there has been minimal investment in the parallel litigation and claims have not yet been construed.

6. Relief Requested

• Petitioner requests institution of an inter partes review and cancellation of claims 1-21 of the ’977 patent as unpatentable.