`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`APPLE INC.,
`Petitioner,
`v.
`PARUS HOLDINGS, INC.,
`Patent Owner.
`
`Case No. IPR2020-00686
`U.S. Patent No. 7,076,431
`
`PATENT OWNER’S RESPONSE
`PURSUANT TO 37 C.F.R. § 42.120(a)
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`
`2.
`
`3.
`
`I.
`II.
`
`TABLE OF CONTENTS
`Introduction ...................................................................................................... 1
`The ’431 Patent ................................................................................................ 3
`A.
`Prior Art Interactive Voice Systems Suffered From Numerous
`Drawbacks ............................................................................................. 4
`1.
`Typical Prior Art Systems For Accessing Web Sites
`Were Not Sufficiently Portable, Comprehensive, And
`Affordable ................................................................................... 4
`Voice Enabled Options Introduced Additional Problems
`and Drawbacks ............................................................................ 5
`Prior Art “Interactive Voice Response” Systems Suffered
`From A Lack Of Fault Tolerance, Limited Webpage
`Resources, And Generic Search Options And Results ............... 6
`Prior Art Speaker-Dependent and Speaker-Independent
`Systems Suffered from Various Drawbacks ............................... 8
`The ’431 Patent’s Solution .................................................................... 9
`1.
`Overview Of The ʼ431 Patent’s Voice Browser System .......... 11
`2.
`Speaker-Independent Speech Recognition Device ................... 14
`3.
`Sequentially Accessing A Plurality of Pre-selected Web
`Sites ........................................................................................... 15
`The Challenged ’431 Patent Claims .................................................... 18
`C.
`Claim Construction.............................................................................. 21
`D.
`III. Apple’s Petition & Grounds .......................................................................... 24
`A.
`Ladd ..................................................................................................... 26
`B.
`Kurosawa ............................................................................................. 28
`C.
`Goedken ............................................................................................... 29
`
`4.
`
`B.
`
`i
`
`
`
`B.
`
`C.
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`Houser.................................................................................................. 32
`D.
`IV. Argument ....................................................................................................... 33
`A.
`All Grounds Fail Because The Petition Does Not Identify The
`“at least one speaker-independent speech recognition device,
`said speaker-independent speech recognition device operatively
`connected to said computer and to said voice enabled device”
`Limitation Required In Every Challenged Claim ............................... 34
`The Petition Does Not Identify the “sequential[] access” of pre-
`selected web sites Limitation Required In Every Challenged
`Claim ................................................................................................... 38
`1.
`Goedken does not teach sequentially accessing
`preselected web sites until the requested information is
`found or all pre-selected web sites have been accessed ........... 38
`Kurosawa does not teach sequentially accessing pre-
`selected web sites until the requested information is
`found or all pre-selected web sites have been accessed ........... 40
`The Petition Relies On Impermissible Hindsight To Support
`The Alleged Obviousness Combinations ............................................ 41
`1.
`There is no motivation to combine Ladd with Kurosawa
`from either reference ................................................................. 43
`There is no motivation to combine Goedken with Ladd
`and Kurosawa in any of the references ..................................... 46
`The Petition relies on impermissible hindsight for its
`motivation to combine arguments ............................................. 48
`The Petition Does Not Establish It Was Obvious To Combine
`Ladd, Kurosawa, and Goedken As Required In All Asserted
`Grounds ............................................................................................... 48
`1.
`A POSITA Would Not Combine Ladd With Kurosawa .......... 49
`2.
`A POSITA Would Not Combine Ladd With Kurosawa
`And Goedken ............................................................................. 53
`
`D.
`
`2.
`
`2.
`
`3.
`
`ii
`
`
`
`E.
`
`F.
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`The Petition does not identify any reference that teaches
`“wherein said speaker-independent speech recognition device is
`configured to analyze phonemes to recognize said speech
`commands” limitation found in claim 5 .............................................. 56
`The Petition does not identify any reference that teaches
`“wherein said speaker-independent speech recognition device is
`configured to recognize naturally spoken speech commands”
`limitation found in claim 6 .................................................................. 57
`Grounds 2 – 4 fail because the addition of Houser, Madnick, or
`Rutledge does not cure the shortcomings of the proposed
`combination of Ladd, Kurosawa, and Goedken .................................. 58
`The Petition has not provided sufficient evidence that the
`proposed combination of Ladd, Kurosawa, and Goedken, could
`be further combined with Houser, Madnick, or Rutledge would
`have been obvious to a person of ordinary skill .................................. 59
`Conclusion ..................................................................................................... 61
`
`H.
`
`G.
`
`V.
`
`iii
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`
`TABLE OF AUTHORITIES
`
` Page(s)
`
`Cases
`Intamin Ltd. v. Magnetar Techs., Corp.,
`483 F.3d 1328 (Fed. Cir. 2007) .......................................................................... 22
`Parus Holdings Inc. v. Apple Inc.,
`No. 6:19-cv-00432-ADA, ECF No. 167 (W.D. Tex. Sept. 8, 2020) .................. 24
`Phillips v. AWH Corp.,
`415 F.3d 1303 (Fed. Cir. 2005) .................................................................... 22, 23
`TQ Delta, LLC v. Cisco Sys.,
`942 F.3d 1352 (Fed. Cir. 2019) .............................................................. 42, 45, 48
`
`iv
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`
`TABLE OF EXHIBITS
`
`Exhibit
`2001
`
`2002
`2003
`2004
`2005
`2006
`
`2007
`2008
`2009
`2010
`2011
`
`2012
`2013
`2014
`2015
`2016
`2017
`2018
`2019
`2020
`2021
`2022
`2023
`2024
`2025
`
`Description
`Order Denying Motion to Stay Pending Inter Partes Review, C.A. No.
`6:18-cv-00207-ADA
`Exhibit A3 Ladd Claim Chart 7076431
`Exhibit C Obviousness Claim Chart 7076431 (Corrected)
`Reserved
`Reserved
`Standing Order Regarding Scheduled Hearings in Civil Cases, 6:19-
`cv-00432-ADA
`Claim Construction Order, 1:20-cv-00351-ADA
`Claim Construction Order, 6:19-cv-00532-ADA
`Claim Construction Order, 6:18-cv-00308-ADA
`U.S. Patent No. 6,157,705 (Perrone)
`“instruction set” excerpt from 1997 Novell’s Dictionary of
`Networking
`Defendants’ Opening Claim Construction Brief, 6:19-cv-00432-ADA
`Excerpt of Case Docket Sheet, 6-19-cv-00278-ADA
`Excerpt of Case Docket Sheet, 6-19-cv-00514-ADA
`Excerpt of Case Docket Sheet, 6-19-cv-00515-ADA
`Excerpt of Case Docket Sheet, 7-18-cv-00147-ADA
`Markman Hearing Transcript, 6:19-CV-00432-ADA
`10/2/2020 Email to Court
`Excerpt of Case Docket Sheet, 6:18-cv-00308-ADA
`5/30/2019 Order Denying Stay, C.A. No. 6:18-cv-00207
`6/23/2020 Order Denying Stay, C.A. No. 6:19-cv-00514
`6/23/2020 Order Denying Stay, C.A. No. 6:19-cv-00515
`7/22/2020 Order Denying Stay, C.A. No. 7:18-cv-00147
`December 16, 2020 Deposition Transcript of Loren Terveen, Ph.D.
`Declaration of Benedict Occhiogrosso in Support of Patent Owner’s
`Response to Inter Partes Review of U.S. Patent No. 7,076,431
`
`v
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`
`I.
`
`Introduction
`Parus began developing its voice response systems in the late 1990’s. In
`
`developing those systems, which are still in use today, the inventors at Parus realized
`
`that combining a speaker-independent speech recognition device with flexible and
`
`fast access to multiple websites was a powerful tool. The Patent and Trademark
`
`Office issued Parus several patents on this technology including the ’431 Patent.
`
`Petitioner resorts to combining three or more references, without any motivation to
`
`combine those references, in order to attempt to disclose Parus’s invention. It fails
`
`to do so. Even the ungainly combination of Ladd, Kurosowa, and Goedken, which
`
`is used as the basis for all grounds of the Petition, lacks a teaching of the speaker-
`
`independent voice recognition device or the sequential access of websites required
`
`by the challenged claims of the ’431 Patent.
`
`Each of the Grounds fail to invalidate any of the claims of the ’431 Patent
`
`because Ladd does not disclose a speaker-independent speech recognition device as
`
`required by the challenged claims. Petitioner relies entirely on Ladd to provide this
`
`teaching. However, a proper construction of the term “speaker-independent speech
`
`recognition device” disclaims recognition based on a predefined voice pattern. Ladd
`
`specifically discloses its automatic speech recognition unit as recognizing voice
`
`patterns.
`
`1
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`None of the prior art references identified in the Grounds of the petition
`
`disclose the element of “sequentially access[ing] said plurality of web sites until said
`
`information to be retrieved is found or until said plurality of web sites has been
`
`accessed.” (Ex. 1001, 20:26-33). Petitioner relies on a combination of Ladd,
`
`Kurosowa, and Goedken to provide this teaching. None of the references teach
`
`sequentially accessing websites until information is found. The closest the
`
`references come is to sequentially accessing database records until the information
`
`is found, which is fundamentally different than the claimed limitation.
`
`The proposed combination of Ladd, Kurosowa, and Goedken fails as a
`
`combination because the combination is cobbled together using impermissible
`
`hindsight and none of the references disclose or hint at the motivation identified by
`
`Petitioner to combine the references. Petitioner conjures a motivation to combine
`
`Kurosowa with Ladd, then finds a motivation to combine Goedken with the
`
`combination of Ladd and Kurosowa based entirely on expert testimony and a
`
`document not cited or otherwise related to any of the references in the combination,
`
`which is a violation of Federal Circuit law. What makes this combination worse is
`
`that the motivation to combine Ladd and Kurosowa (flexibility) conflicts with the
`
`motivation to combine Goedken with the combination of Ladd and Kurosowa
`
`(speed). Combining the teachings of Kurosowa and Goedken with the teachings of
`
`Ladd necessarily slows the system down. Finally, the motivations to combine the
`
`2
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`references are entirely made up by Petitioner’s expert. They are not found in the
`
`references themselves. Petitioner doesn’t even try to allege that Goedken contains
`
`the motivation and the allegation that Ladd contains the motivation to make the
`
`system more flexible is extraordinarily thin.
`
`Finally, Ground 3 fails because Ladd does not disclose the additional
`
`limitations of Claims 5 and 6. Ladd does not disclose a “speaker-independent speech
`
`recognition device” based on phonemes or one that uses natural language as
`
`discussed in more detail below.
`
`Patent Owner asks the Board to find that none of the challenged claims are
`
`unpatentable based on the asserted grounds.
`
`II.
`
`The ’431 Patent
`The ’431 Patent is directed to an “interactive voice system” system that allows
`
`a user to request information from web sites using voice commands and, in turn,
`
`provides retrieved information to the user in audio form. More specifically, the
`
`claims of the ’431 Patent combine the use of a flexible speaker-independent speech
`
`recognition device with efficient information retrieval from web sites to provide an
`
`adaptable and effective way to get information from the Internet using your voice.
`
`3
`
`
`
`A.
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`Prior Art Interactive Voice Systems Suffered From Numerous
`Drawbacks
`1.
`Typical Prior Art Systems For Accessing Web Sites Were
`Not Sufficiently Portable, Comprehensive, And Affordable
`At the time of the ʼ431 Patent’s inventions, there was a need for a system that
`
`allowed a user to efficiently and quickly access web sites from various locations,
`
`rather than being tied down to the permanent site of a non-portable desktop
`
`computer. As the ʼ431 Patent describes, portable options existed for users wishing
`
`to quickly gather information from a web site accessible over the Internet were
`
`limited, each having distinct drawbacks. (Ex. 1001, 1:30-43). These limited options
`
`included (1) heavy and bulky laptop computers with limited access to power and
`
`communication lines; (2) very expensive Personal Digital Assistants (PDAs), which
`
`required expensive service plans for Internet access and could only access web sites
`
`specially designed to be compatible with PDAs (which many web sites were not);
`
`and (3) web-phones or web-pagers that suffered from similar drawbacks. (Ex. 1001,
`
`1:42-2:24).
`
`There was thus a need for an alternative and improved system that could
`
`provide more portable, ubiquitous, and comprehensive access to web sites to any
`
`and every one without adding additional expenses.
`
`4
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`Voice Enabled Options Introduced Additional Problems
`and Drawbacks
`People used phone lines to access the Internet through dial-up connections,
`
`2.
`
`which tethered any access to a static environment. There existed no way for a person
`
`on the move to access the internet from any location. One potential alternative to
`
`expand internet use to a dynamic environment in which a person moving locations
`
`could access the internet from anywhere was to use voice enabled devices to connect
`
`to the Internet and access web sites. However, voice enabled devices at the time
`
`introduced a number of additional considerations and problems that had to be
`
`addressed that did not exist with typical computers. Mobile users expected results
`
`quickly and would not tolerate latency. Moreover, systems would have to be built
`
`to provide a user with Internet search results quickly and mobile users could not
`
`quickly browse through search results to select from them what they desired.
`
`For example, when a user accesses a web site with a desktop or laptop
`
`computer, delays in fetching and rendering the web site are tolerated and even
`
`expected. (Ex. 1001, 2:42-45). In contrast, when a user communicates using a
`
`telephone, users expect the communications to occur immediately with a minimal
`
`amount of delay. (Ex. 1001, 2:45-48).
`
`In addition, using a typical desktop or laptop computer, a user is able to search
`
`the Internet, quickly read the resulting list of possible web sites, and then choose and
`
`5
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`open the most pertinent web site to access the desired information. But using a list
`
`of possible web sites in this manner is not feasible with a voice enabled device that
`
`provides the requested information as audio to the user—the user would have neither
`
`the time nor the patience to listen to a long list of potential web sites before choosing
`
`the appropriate one to receive the desired information. Therefore, voice enabled
`
`devices needed to present the requested information from a website without
`
`providing the traditional list of relevant web sites.
`
`3.
`
`Prior Art “Interactive Voice Response” Systems Suffered
`From A Lack Of Fault Tolerance, Limited Webpage
`Resources, And Generic Search Options And Results
`Some prior art attempts to develop a voice enabled system to meet and solve
`
`the foregoing problems used an “Interactive Voice Response” (IVR) system. An
`
`IVR system allows a user to place a phone call into a system and navigate through a
`
`number of options in response to voice prompts and retrieve information stored in a
`
`computer database. Well-known examples of IVR systems include a voicemail
`
`system and automated customer service call centers.
`
`During the prosecution of the ʼ431 Patent, the Examiner cited—but allowed
`
`the challenged claims over—U.S. Patent No. 6,157,705 (“Perrone”), which
`
`discloses a typical prior art attempt to use an IVR system to retrieve information
`
`from a web site. In Perrone, a user places a telephone call to the IVR system, issues
`
`6
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`a voice command requesting information from the Internet, and (ideally) receives an
`
`audio message containing the requested information. (Ex. 2010 Perrone, Abstract).
`
`In the Perrone system, the IVR system utilizes a table that maps specific voice
`
`commands to a specific web site resource. When the Perrone IVR system receives
`
`a voice command (such as “stocks”), it determines whether it can be mapped to a
`
`web site resource in the table. If so, the Perrone IVR system accesses the web site
`
`resource to retrieve its information, converts that information into speech, and
`
`delivers it to the user in audio form. Id.
`
`But Perrone’s and similar attempts to address the problems associated with a
`
`voice enabled web browser (the aforementioned increased need for speed and the
`
`user’s inability to quickly browse through Internet search results) suffered from a
`
`variety of additional drawbacks.
`
`First, these systems were not fault tolerant. Speech commands mapped to
`
`specific resources on specific web sites. If the web site mapped to a given speech
`
`command was inaccessible, there would be no way for the system to retrieve the
`
`requested information. For example, in Perrone’s IVR system, if the resource
`
`identifier “news” corresponded to “www.news.com,” and the news.com site was
`
`currently not accessible, there would be no way to get the news.
`
`Further, these systems were limited to get a specific resource from a specific
`
`web site. For example, if a user wanted to get a traffic update, they were limited to
`
`7
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`the traffic resource that is mapped to the traffic update command. This traffic
`
`resource may not be the user’s preferred resource for traffic information, may not be
`
`the best traffic source with current traffic conditions, or perhaps the user desires the
`
`traffic updates from a plurality of different resources. IVR systems like that in
`
`Perrone did not offer an avenue to solve this problem.
`
`Finally, these systems provided only generic information in response to a
`
`mapped command. If a user was searching for specific information, such as an
`
`answer to a question, unless the question was mapped to a resource identifier, there
`
`would be no way to get the specific answer to the specific question.
`
`4.
`
`Prior Art Speaker-Dependent and Speaker-Independent
`Systems Suffered from Various Drawbacks
`Prior art voice response systems generally fell into two broad categories,
`
`speaker-dependent and speaker-independent. Speaker-dependent systems required
`
`training by individual users prior to being able to be used. (Ex. 2025, at ¶¶ 50-52).
`
`Without advance training of the system by each individual user, such systems could
`
`not recognize what was spoken. Speaker-independent systems did not require
`
`training by each individual in order to operate and came in various flavors. Id., (Ex.
`
`1001, 4:38-42).
`
`Many prior art speaker-independent systems used voice patterns to recognize
`
`spoken voice commands. The speaker-independent system of Ladd is an example
`
`8
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`of such a speaker-independent system. Specifically, the speaker-independent system
`
`of Ladd was described as “[w]hen the ASR unit 254 identifies a selected speech
`
`pattern of the speech input, the ASR unit 254 sends an output signal to implement
`
`the specific function associated with the recognized voice pattern.” (Ex. 1004, 9:35-
`
`40). Such a speaker-independent system is an improvement over speaker-dependent
`
`systems because no training by individual users is necessary, however it still has
`
`significant drawbacks. Recognizing a speech pattern is resource intensive, in that
`
`such systems also required advanced training for each spoken word and could only
`
`recognize a word that the system was familiar with in advance. For this, many
`
`different sounds have to be analyzed together in order to recognize the spoken word
`
`in a voice command. (Ex. 2025, at ¶¶ 52-55). Another, more significant drawback
`
`to this type of speaker-independent system is that it sharply restricts the library of
`
`phrases that can be used with the system, requiring highly structured and specific
`
`commands to be used. (Ex. 1004, 17:1-33, 18:15-32, and 38:4-11; Ex. 2025, at ¶
`
`82).
`
`The ’431 Patent’s Solution
`B.
`The ʼ431 inventors were able to overcome the drawbacks of prior art
`
`traditional systems (lack of portability, comprehensive coverage, expensive), voice
`
`enabled systems (need for increased speed to provide results quickly from reliable
`
`sources and inability to provide search results as a list of web pages that a user must
`
`9
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`remember and then select a desired resource) and IVR systems (lack of fault
`
`tolerance, limited web site resources, and generic search options and results and
`
`more).
`
`To this end, the ’431 Patent discloses “a robust and highly reliable system that
`
`allows users to browse web sites and retrieve information by using conversational
`
`voice commands,” where the retrieved information is “converted into an audio
`
`message [and] transmitted to the user’s voice enabled device.” (Ex. 1001, 1:20-24,
`
`3:41-56). For purposes of this Response, material improvements over prior art IVR
`
`systems include the following:
`
` An ability to access a first web site of a plurality of web sites and, if the
`
`information to be retrieved is not found at the first web site, the computer
`
`configured to sequentially access the plurality of web sites until the
`
`information to be retrieved is found or until the plurality of web sites has
`
`been accessed. This ability allowed the system to adjust for unavailable
`
`web sites and provide more responsive and accurate information to more
`
`specific requests, while allowing for a rapid response, thereby solving prior
`
`art IVR system’s lack of fault tolerance, limited web site resources, and
`
`generic search options and results.
`
` A flexible and highly adaptable speaker-independent speech recognition
`
`device that “use[s] phonemes to recognize spoken words and not
`
`10
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`predefined voice patterns,” which substantially increases the adaptability
`
`of the system.
`
`The ability to sequentially access web sites in their ranked order until the
`
`requested information is retrieved resolves many issue in the prior art. (Ex. 1001,
`
`20:4-33). This sequential accessing of pre-selected web sites allows the ʼ431 Patent’s
`
`system to provide the speed and audio response required by a voice enabled system,
`
`while still utilizing the plurality of web sites that improve upon prior art IVR
`
`systems.
`
`The speaker-independent speech recognition system allows a user to leverage
`
`the flexibility of the voice recognition system. The ’431 Patents do not require a
`
`user “to learn a special language or command set in order to communicate with the
`
`voice browsing system.” (Ex. 1001, 4:34-36). Instead, the phoneme based speaker-
`
`independent speech recognition device of the ’431 Patent allows users to use natural
`
`language in a flexible way that is not possible with other speech recognition systems
`
`that rely on recognition of predefined voice patterns. (Ex. 1001, 4:38-43).
`
`Overview Of The ʼ431 Patent’s Voice Browser System
`1.
`In one embodiment, the ʼ431 Patent describes “a browsing system and method
`
`that allows users to browse web sites using conversational voice commands spoken
`
`into any type of voice enabled device … These spoken commands are then converted
`
`into data messages by a speech recognition software engine.” (Ex. 1001, 3:41-46).
`
`11
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`Figure 1 (reproduced below) depicts media servers 106, which include (among
`
`other things) the speech recognition software engine 300. (Ex. 1001, 3:41-46, 5:60-
`
`6:24).
`
`After converting the spoken commands to data messages, the media server
`
`106 then processes the resulting data message to recognize keywords by using sub-
`
`words and/or phonemes. (Ex. 1030, 4:34-43, 5:54-59). For example, if the request
`
`was “what is the weather in Chicago?”, the converted data messages may be used to
`
`identify the keywords “weather” and “Chicago.” (Ex. 1001, 6:44-54).
`
`12
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`The media server 106 then uses those keywords to search website records
`
`stored in database 100 (also shown in Figure 1 above). (Ex. 1001, 6:44-56). This
`
`process and database are explained in more detail below, but as a result of this search,
`
`the media server 106 identifies a first web site likely to have the requested
`
`information.
`
`A web browsing server 106 (Figure 1 above) receives a website record 200
`
`concerning both the web site found in the search (including its URL) as well as
`
`information concerning the user’s request. (Ex. 1001, 5:5-11, 6:52-56, 7:14-36). For
`
`example, the website record (illustrated in Figure 2 below) includes “extraction
`
`agent” commands 206 that identify which information should be retrieved from the
`
`web site to fulfill the user’s request, and how to accomplish that retrieval. (Ex. 1001,
`
`5:5-11, 6:52-56, 7:14-36).
`
`13
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`The web browsing server 106 uses this web site record to access the identified
`
`web site and extract the information identified for retrieval, through a process known
`
`as “content extraction.” (Ex. 1001, 6:65-7:36). The web browsing server invokes
`
`the “extraction agent command” 206 contained in the record to identify which
`
`information (i.e., “content”) to retrieve from the web site, where to find that content
`
`at the web site, and how to request and extract that content. (Ex. 1001, 7:14-37).
`
`Once the web browsing server 102 accesses the web site specified in the URL
`
`to obtain the identified information to be retrieved, it forwards that retrieved
`
`information to media sever 106. (Ex. 1001, 15:32-34). Media server 106 further
`
`comprises a speech synthesis engine 302, which it uses the retrieved information “to
`
`create an audio message that is transmitted to the user’s voice enabled device 112.”
`
`(Ex. 1001, 15:32-337).
`
`The foregoing is merely a general high-level description of an embodiment of
`
`the ʼ431 Patent’s system. Additional details and innovations concerning the
`
`database 100, media server 106, and web browsing server 102 are relevant to the
`
`claimed invention and the substantive failings of the Petition, and are discussed in
`
`the next section.
`
`Speaker-Independent Speech Recognition Device
`2.
`The ’431 Patent requires the use of a specific type of speaker-independent
`
`speech recognition device. The speaker-independent speech recognition device
`
`14
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`recognizes natural speech commands without needing to be “trained to recognize the
`
`voice patterns of each individual user.” (Ex. 1001, 4:34-43). The intrinsic record of
`
`the ’431 patent teaches that the “speaker-independent speech recognition device” is,
`
`at a minimum, a “speech recognition device that recognizes spoken words without
`
`using predefined voice patterns.” (Ex. 1001, 4:42-43).
`
`The specification disclaims the use of predefined voice patterns and states that
`
`the speaker-independent system “use[s] phonemes to recognize spoken words and
`
`not predefined voice patterns.” (Ex. 1001, 4:42-43 (emphasis added)). Therefore,
`
`in the context of the ’431 Patent’s specification, “speaker-independent” refers to a
`
`system that uses a segmental modeling approach based on recognizing discrete
`
`phonetic units or “phonemes” that the system uses to construct words. (Ex. 2025, at
`
`¶ 53). This approach is able to use large vocabularies and allows for natural
`
`language commands. Id. The specification further explains that the system “is
`
`speaker-independent; it does not have to be trained to recognize the Voice patterns
`
`of each individual users.” (Ex. 1001, 4:38-42).
`
`Sequentially Accessing A Plurality of Pre-selected Web Sites
`3.
`Unlike Perrone and other prior art IVR systems—including Ladd, the
`
`Petition’s primary reference—the ʼ431 Patent system did not rely on a single web
`
`site to retrieve information in response to a particular request. Rather, the ʼ431
`
`Patent discloses an “instruction set” for identifying a plurality of web sites
`
`15
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`containing the information to be retrieved. Identifying and using multiple potential
`
`web sites to satisfy the user’s verbal command or question allows the system to
`
`adjust for unavailable web sites and provide more responsive and accurate
`
`information in a manner that the prior art IVR systems could not.
`
`But where typical desktop and laptop web browsers could return a plurality of
`
`web sites without issue, a voice enabled browser could not because of the
`
`aforementioned need for speed and an audio response. To achieve these ends and
`
`still effectively use a plurality of web sites in a voice enabled system, the ʼ431 Patent
`
`further disclosed that the instruction set includes a database of pre-selected web
`
`sites that are ranked both for speed and the usefulness of the information stored
`
`thereon, so that the web sites may be sequentially accessed in their ranked order
`
`until the identified information is retrieved.
`
`For example, the ʼ431 Patent describes that database 100 contains web site
`
`information for not just a single web site for answering a particular command or
`
`request, but a plurality of pre-selected possible sites. (Ex. 1001, 5:3-53 (including
`
`Table 1, depicting stored data for two distinct web sites, each for potential use in
`
`answering a query concerning “weather”), 16:31-43). In addition, each of these pre-
`
`selected potential web sites is ranked, on a continual basis, based on aspects such as
`
`speed of response speed and the accuracy and completeness of the provided data.
`
`(Ex. 1001, 16:56-17:28).
`
`16
`
`
`
`Case No. IPR2020-00686
`Patent No. 7,076,431
`As described, “the database 100 contains a separate set of records for each
`
`web site accessible by the system”:
`
`An example of a web site record is shown in FIG. 2. Each web site
`record 200 contains the rank number of the web site 202, the
`associated Uniform Resource Locator (URL) 204, and a command that
`enables the appropriate “extraction agent” 206 that is required in order
`to generate proper requests sent to[,] and to format data received from[,]
`the web site.
`(Ex. 1001, 16:56-17:28 (emphasis added)).
`
`Thus, when the media server uses keywords generated from the user’s vocal
`
`request to search the database 100’s web site records to identify the different pre-
`
`selected web sites that may be used to retrieve information for the user, it will only
`
`select and forward to the web browsing server 102 the web site record for the
`
`relevant web site with the highest rank:
`
`For instance, if the user’s request was “What is the weather in
`Chicago?”, the keywords “weather” and “Chicago” would be
`recognized [by media server 106]. A web site record 200 with the
`highest rank number from the “weather” category within the database
`100 would then be selected and transmitted to the web browser server
`102 along with an identifier indicating that Chicago weather is being
`requested.
`(Ex. 1001, 6:44-56).
`
`The web browsing server 102 will then access the first-ranked pre-selected
`
`web site for the information to be retrieved.