PTAB

IPR2015-00347

Google Inc v. Network 1 Technologies Inc

Log In
Key Events
Petition
petition

1. Case Identification

2. Patent Overview

  • Title: Associating an Electronic Work with an Action
  • Brief Description: The ’988 patent relates to a method for identifying an unknown media work and performing a corresponding action. The method involves electronically extracting features from the work, determining its identity using a non-exhaustive search to find a "neighbor" in a database, determining an action based on that identity, and electronically performing the action.

3. Grounds for Unpatentability

Ground 1: Anticipation over Ghias - Claims 1-3, 7-9, 13-17, 21-23, 28, 30-31, 40, 49, and 51 are anticipated by Ghias under 35 U.S.C. § 102(b).

  • Prior Art Relied Upon: Ghias (Patent 5,874,686)
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Ghias, which discloses a system for identifying a hummed melody, teaches every element of the independent claims. Ghias discloses (1) extracting a "contour representation" of a melody, which meets the "extracting features" limitation; (2) searching a melody database using an "efficient approximate pattern matching algorithm" to locate a close match, which meets the "non-exhaustive search identifying a neighbor" limitation; (3) determining an action based on search results, such as the number of matches to display; and (4) performing the action by outputting a ranked list of matching melodies. Petitioner contended that Ghias also discloses the "sublinear" search required by dependent claims 3 and 17, as its disclosed algorithms are substantially faster than brute-force search.

Ground 2: Anticipation over Wood - Claims 1-2, 7-9, 13-16, 21-23, 27-33, 38, 40-42, 47, 49, and 51 are anticipated by Wood under 35 U.S.C. § 102(e).

  • Prior Art Relied Upon: Wood (Patent 7,743,092)
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner asserted that Wood, which describes a system to identify a song hummed or sung by a user over the internet, anticipates the challenged claims. Wood allegedly discloses (1) a "feature extraction unit" that extracts musical features characterizing a song; (2) comparing the extracted pattern with patterns in a library to find a candidate if the "distan[ce] score" is smaller than a predetermined value, which constitutes a neighbor search. Wood further discloses this search can be non-exhaustive by comparing only memorable portions of a song. Finally, Wood teaches (3) determining an action, such as transmitting a "corrupted version" of the song for confirmation or displaying information, and (4) performing that action.

Ground 3: Obviousness over Levy and Arya - Claims 1-3, 7-17, 21-34, 37-38, 40-43, 46-47, and 49-52 are obvious over Levy in view of Arya under 35 U.S.C. § 103.

  • Prior Art Relied Upon: Levy (Patent 6,505,160) and Arya (a July 1998 publication titled "An Optimal Algorithm for Approximate Nearest Neighbor Searching in Fixed Dimensions").

  • Core Argument for this Ground:

    • Prior Art Mapping: Petitioner argued Levy discloses a system that meets most claim limitations by extracting a "fingerprint" from an audio signal, sending it to a server, mapping the identifier to an action (like displaying information or performing a transaction), and executing it. However, Levy does not explicitly detail a non-exhaustive, neighbor-finding search. Arya remedies this by disclosing an efficient algorithm for "approximate nearest neighbor" searching that provides significant improvements over brute-force methods, explicitly for applications like multimedia databases.
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine Levy with Arya to improve Levy's system. A POSITA would have recognized that fingerprints of media works from different sources would inevitably have slight discrepancies, necessitating a search for close or "neighbor" matches, not just exact ones. Arya provided a well-known and more efficient solution to this exact problem, making its integration into a system like Levy's a logical step to improve real-world performance.
    • Expectation of Success: A POSITA would have had a reasonable expectation of success in combining the references because it involved applying a known, efficient search algorithm (Arya) to a known type of media identification system (Levy) to solve the predictable and known problem of finding near matches for content fingerprints.

  • Additional Grounds: Petitioner asserted additional challenges, including that claims 15-17, 21-22, 24-25, 27-28, 31-32, 37, 40-41, and 46 are obvious over Iggulden in view of Böhm; claims 1-3, 7, 9, 13-17, 21, 23, 27-33, 38, 40-42, 47, 49, and 51 are anticipated by Iwamura; and other claims are obvious over Ghias and Wood as single references.

4. Key Claim Construction Positions

Petitioner argued for the following "broadest reasonable constructions" for key terms, which are central to its arguments that the prior art meets the claim limitations:

  • "extracting features": Proposed construction is "generating a quantitative representation of the features of a media work, or a subset of the features of a media work." This broad construction encompasses various fingerprinting and feature representation methods disclosed in the prior art.
  • "non-exhaustive search": Proposed construction is "a search that locates a match without conducting a brute force comparison of all possible matches, and all data within all possible matches." This covers any search algorithm more efficient than a brute-force comparison.
  • "identifying a neighbor": Proposed construction is "identifying a close, but not necessarily exact, match." This construction is critical for applying prior art that discloses approximate or similarity searching rather than identical matching.

5. Relief Requested

  • Petitioner requests institution of an inter partes review and cancellation of claims 1-3, 7-17, 21-34, 37-38, 40-43, 46-47, and 49-52 of the ’988 patent as unpatentable.