PTAB

IPR2019-00556

Comcast Cable Communications, LLC v. Rovi Guides, Inc.

Log In

1. Case Identification

2. Patent Overview

  • Title: Systems and Methods for Identifying and Storing a Portion of a Media Asset
  • Brief Description: The ’014 patent relates to systems and methods for resolving voice commands to retrieve, store, and play back media assets. The technology involves receiving a voice command that may not expressly name a media asset, comparing an identifier from the command to a database of known identifiers using a similarity metric, and suggesting a matching asset to the user for confirmation.

3. Grounds for Unpatentability

Ground A: Obviousness over Chang in view of Cho - Claims 1, 4, 7-8, 11, 14, 17-18

  • Prior Art Relied Upon: Chang (Application # 2011/0119715) and Cho (Patent 7,526,506).
  • Core Argument for this Ground:
    • Prior Art Mapping: Petitioner argued that Chang taught all elements of independent claims 1 and 11 except for the specific method of "calculating a degree of similarity." Chang disclosed a media guidance system that receives inexact voice commands (e.g., "Minor Report" for "Minority Report"), searches an electronic program guide (EPG) database, and suggests the most likely media asset. Cho was argued to supply the missing element by teaching a method for determining similarity between text strings (e.g., a search query and a database entry) using a "fuzzy string comparison" algorithm based on the well-known Levenshtein distance, which calculates a similarity score on a scale of 0-100.
    • Motivation to Combine: A Person of Ordinary Skill in the Art (POSITA) would combine the references to improve Chang's system. Chang’s method for determining similarity was described as basic (e.g., based on number of syllables), and a POSITA would have been motivated to substitute it with Cho's more robust and well-known Levenshtein distance algorithm to improve the accuracy of search results, especially for distinguishing between closely-matching titles.
    • Expectation of Success: A POSITA would have a reasonable expectation of success because both systems operate in the same field (EPG searching), use similar inputs (text strings) and hardware (processors, memory), and produce a similar output (a ranked list of relevant media assets). Implementing the known Levenshtein algorithm from Cho into Chang's system was presented as a predictable modification.

Ground C: Obviousness over Chang, Cho, Jurafsky, and Wang - Claims 3 and 13

  • Prior Art Relied Upon: Chang (Application # 2011/0119715), Cho (Patent 7,526,506), Jurafsky (a 2000 textbook on speech and language processing), and Wang (a 2007 programming book).
  • Core Argument for this Ground:
    • Prior Art Mapping: This ground builds on the Chang and Cho combination to specifically address the limitations of claims 3 and 13, which require calculating similarity based on the number of matching words between strings. Petitioner argued that while the Chang-Cho combination teaches the overall system, Jurafsky explicitly taught this word-matching method. Jurafsky described a vector-space model ("bag of words") for information retrieval where similarity is calculated by summing the number of common words between two strings, irrespective of order. This method was argued to be mathematically equivalent to the embodiment described in the ’014 patent. Wang was cited for the basic programming principle of initializing variables to a default value (e.g., zero) before calculation, a step implicit in any counting algorithm.
    • Motivation to Combine: A POSITA would have been motivated to use Jurafsky's vector-space model as an alternative or improvement to Cho's character-based Levenshtein distance. Jurafsky’s method was known to be simpler, faster, and require less memory, providing a clear design choice for improving the efficiency of the similarity calculation in the Chang-Cho system.
    • Expectation of Success: The expectation of success was high because the vector-space model was a well-documented, fundamental algorithm in information retrieval. A POSITA would have found it straightforward to implement this known algorithm in place of or alongside other similarity metrics within the Chang-Cho framework.

Ground D: Obviousness over Chang, Cho, and Hamano - Claims 5, 9, 15, 19

  • Prior Art Relied Upon: Chang (Application # 2011/0119715), Cho (Patent 7,526,506), and Hamano (Application # 2010/0333137).

  • Core Argument for this Ground:

    • Prior Art Mapping: This ground asserted that adding Hamano to the Chang-Cho combination rendered claims related to user profiles and device-specific actions obvious. Hamano taught an interactive media guidance system that improved upon basic search by using stored user profiles. These profiles contained user preferences (e.g., preferred genres) and information about available playback/recording devices (e.g., "living room TV," "portable media player"). Hamano disclosed tailoring recommendations based on these preferences and selecting a specific device for storage or playback based on a device identifier in a voice command.
    • Motivation to Combine: A POSITA would combine Hamano with the Chang-Cho system to add personalization and device-specific control, thereby improving the user experience. The combination would allow the system not only to find a media asset via a fuzzy voice search but also to make a more intelligent recommendation based on the user’s known preferences or to store the asset on a user-specified device.
    • Expectation of Success: Success would be expected because Chang, Cho, and Hamano all describe media guidance applications implemented on similar hardware (mobile devices, set-top boxes). Integrating Hamano's user profile and device-selection features into the Chang-Cho EPG was presented as the application of a known technique to improve a similar system in a predictable way.

  • Additional Grounds: Petitioner asserted additional obviousness challenges, including combining Chang and Cho with Hoffert (to add media previews to search results) and with Wood (to add functionality for managing insufficient storage space by prompting deletion of old content).

4. Key Claim Construction Positions

  • Petitioner asserted that several terms should be given their explicit lexicographic definitions provided in the ’014 patent specification, including "media asset," "command," and "media asset identifier."
  • Petitioner contended that the terms "communications circuitry" and "control circuitry configured to" are means-plus-function terms under 35 U.S.C. §112, para. 6. Petitioner identified the corresponding structure for "control circuitry" as the control circuitry and software disclosed in the specification that performs the recited functions (e.g., comparing, determining similarity, selecting). Petitioner argued that even if not construed as means-plus-function terms, the prior art disclosed the claimed circuitry.

5. Arguments Regarding Discretionary Denial

  • Petitioner argued that the grounds presented in this petition were not redundant to grounds asserted in three other concurrently filed IPR petitions against the same patent. It was explained that each petition relies on different prior art combinations and different invalidity theories (e.g., anticipation vs. obviousness) and addresses different possible constructions of the key claim term "degree of similarity" (e.g., Levenshtein distance, word count, relevance scores). This distinction was argued to weigh against any discretionary denial based on redundancy or cumulativeness.

6. Relief Requested

  • Petitioner requested institution of an inter partes review and cancellation of claims 1-20 of Patent 9,668,014 as unpatentable.