`
`IN THE UNITED STATES DISTRICT COURT
`
`FOR THE DISTRICT OF DELAWARE
`
` )
`
`
`
`
`)
`)
`)
`) C.A. No. 15-cv-262-SLR-SRF
`)
`)
`)
`)
`)
`
`
`IMPROVED SEARCH LLC,
`
`
`Plaintiff,
`
`
`
`v.
`
`
`AOL INC.,
`
`
`
`
`
`
`Defendant.
`
`DECLARATION OF JAIME CARBONELL, PH.D IN SUPPORT OF
`IMPROVED SEARCH OPENING CLAIM CONSTRUCTION BRIEF
`
`
`I, Jaime Carbonell, Ph.D., do hereby declare as follows:
`
`I.
`
`INTRODUCTION
`
`1.
`
`I, Jaime Carbonell, Professor at the Language Technologies Institute, in the
`
`School of Computer Science at Carnegie Mellon University, located at 5000 Forbes Avenue,
`
`Pittsburgh PA 15213, am over eighteen years of age, and I am competent to testify as to the
`
`matters set forth herein if I am called upon to do so.
`
`2.
`
`I have been retained by Improved Search LLC (“Improved Search”) to provide
`
`expert testimony in the above captioned matters. In particular, I have been asked to provide my
`
`expert opinions on the proper construction of claim terms in U.S. Patent Nos. 6,604,101 (the
`
`“’101 Patent”) and 7,516,154 (the “’154 Patent). I have also been asked to opine on the
`
`understanding that a person of ordinary skill in the art at the time of the inventions claimed in the
`
`‘101 Patent and ‘154 Patent would have had with respect to those terms. I am being
`
`compensated at the rate of $550 per hour.
`
`
`
` 1
`
`AOL Ex. 1013
`Page 1 of 15
`
`

`

`
`
`II.
`
`BACKGROUND AND EXPERIENCE
`
`3.
`
`I received Bachelor of Science degrees in both Physics and Mathematics in 1975
`
`from the Massachusetts Institute of Technology. I received M.S., M.Phil. and Ph.D. degrees in
`
`Computer Science from Yale University in 1976, 1977, and 1979, respectively.
`
`4.
`
`I have held the position of Allen Newell Professor of Computer Science at
`
`Carnegie Mellon University from 1995 to the present. I currently also hold the title of Director
`
`of the Language Technologies Institute at Carnegie Mellon University. I first joined Carnegie
`
`Mellon as an Assistant Professor of Computer Science in 1979. In 1987, I was appointed as a
`
`Professor of Computer Science at Carnegie Mellon.
`
`5.
`
`Since 1979, I have taught a wide variety of graduate and undergraduate courses at
`
`Carnegie Mellon that fall within the general field of Computer Science, including courses in
`
`software engineering, data mining, natural language processing, search engines, machine
`
`translation, electronic commerce, and artificial intelligence. I have been involved in a number of
`
`different professional organizations and activities, including memberships in the Association of
`
`Computing Machinery (“ACM”), the Association for the Advancement of Artificial Intelligence
`
`(“AAAI”), and the Cognitive Science Society. I have also held leadership positions within
`
`professional organizations. From 1983 to 1985, I served as Chair of the ACM’s Special Interest
`
`Group on Artificial Intelligence (“SIGART”). From 1988 to the present, I have been a Fellow of
`
`the AAAI. From 1990 to 1992, I served on the AAAI executive committee. I have also served
`
`on a number of different government committees, including the Computer, Information Science
`
`& Engineering Advisory Committee of the National Science Foundation (2010 to 2014); the
`
`Human Genome Scientific Advisory Committee to the National Institute of Health, also known,
`
`colloquially, as the “Watson Committee” (from 1988 through 1992); and the Scientific Advisory
`
` 2
`
`
`
`AOL Ex. 1013
`Page 2 of 15
`
`

`

`
`
`Committee of the Information Access Division of the National Institute of Standards and
`
`Technology (from 1997 through 2001).
`
`6.
`
`I am an author or co-author on more than 330 technical papers published as
`
`invited contributions and/or in peer-reviewed journals or conferences. These papers present the
`
`results of my research, which is generally directed at computer implemented algorithms and
`
`methods that relate to machine learning, natural language processing and information retrieval,
`
`including such applications as cross-language information retrieval (best paper award),
`
`optimizing database access, machine, parsing natural language (a.k.a., “content analysis”), search
`
`engine optimization, and text mining. I have served as an editor and peer-reviewer for a number
`
`of different technical journals in my field, including the Machine Learning Journal (from 1984
`
`through 2000), the Machine Translation Journal (the 1980’s), and the Artificial Intelligence
`
`Journal (1984 through 2008). I was also a Co-editor of the book series Lecture Notes in
`
`Artificial Intelligence, which was published by Springer from 1996 through 2008.
`
`7.
`
`I received a “recognition of service” award from the Association for Computing
`
`Machinery for my role as chair of the ACM’s special interest group in Artificial Intelligence
`
`(SIGART) between 1983 and 1985. In 1986, I received the Sperry Fellowship for excellence in
`
`artificial intelligence research. In 1987, I received the Carnegie Mellon University Herb Simon
`
`Computer Science Department’s teaching award.
`
`8.
`
`I have also worked as a technical consultant on Computer Science applications for
`
`a variety of industrial clients. This includes consulting on data mining applications for Industrial
`
`Scientific Corporation (data mining to improve workplace safety); Carnegie Group Inc. (artificial
`
`intelligence and natural language processing); Citicorp (financial data mining, natural language);
`
`Wisdom Technologies (financial optimization); Dynamix Technologies (large-scale algorithms,
`
` 3
`
`
`
`AOL Ex. 1013
`Page 3 of 15
`
`

`

`
`
`databases and information retrieval, with applications to Homeland Security), and Meaningful
`
`Machines in natural language processing and machine translation. I have experience in many
`
`aspects of computing technology, including electronic commerce, where I regularly teach two
`
`classes every year, including in data mining and business processes, and in search engine
`
`optimization.
`
`9.
`
`I am a named inventor on a number of issued U.S. Patents, including: U.S. Patent
`
`No. 5,677,835 (“Integrated authoring and translation system”); U.S. Patent No. 5,995,920
`
`(“Computer-based method and system for monolingual document development”); U.S. Patent
`
`No. 6,139,201 (“Integrated authoring and translation system”); U.S. Patent No. 6,163,785
`
`(“Integrated authoring and translation system”); and U.S. Patent No. 7,406,443 (“Method and
`
`system for multi-dimensional trading”).
`
`10.
`
`A current copy of my curriculum vitae setting forth details of my background and
`
`relevant experience, including a full list of my publications and a listing of cases for which I
`
`have provided expert testimony over the last seven years as Exhibit A.
`
`III. MATERIALS REVIEWED
`
`11.
`
`In performing the analysis that is the subject of this Declaration, I have reviewed:
`
`•
`
`•
`
`•
`
`•
`
`•
`
`•
`
`US Patent No. 6,604,101
`
`US Patent No. 7,516,154
`
`File History for the US Patent No. 6,604,101
`
`File History for the US Patent No. 7,516,154
`
`Joint Claim Construction Statement
`
`Exhibits B-J attached hereto
`
`
`
` 4
`
`
`
`AOL Ex. 1013
`Page 4 of 15
`
`

`

`
`
`IV. UNDERSTANDING THE LAW TO BE APPLIED TO INTERPRET CLAIMS
`
`12.
`
`In formulating my opinions and conclusions in this declaration, I have been
`
`provided with an understanding of some of the prevailing principles of United States patent law
`
`that govern the issues of patent claim interpretation, including claim construction, applicable to
`
`my declaration.
`
`V.
`
`LEVEL OF ORDINARY SKILL IN THE ART
`
`13.
`
`It is my understanding that my analysis of the interpretation of a claim term must
`
`be undertaken from the perspective of what would have been known or understood by a person
`
`having ordinary skill in the art (POSITA) at the time of the invention. In my opinion a POSITA
`
`would have had a Bachelor’s degree in computer science or computer engineering, plus two or
`
`more years of either work experience or graduate study involving web-based search engines, as
`
`well as familiarity with multi-lingual text processing and databases.
`
`VI. CLAIM TERMS
`
`A.
`
`Translating
`
`14.
`
` The common and ordinary meaning of “translating”, consistent with the
`
`specifications of the ‘101 and ‘154 patents is “changing the text in one language into an
`
`equivalent text in a different language.” No further narrowing is warranted, certainly not
`
`constraining the different language to have been preselected by the user. I have worked in
`
`translation since college in the 1970’s, and later in machine translation, and nowhere is there a
`
`pre-selection restriction placed on the second language. Nor do the specifications place such a
`
`restriction; a single example embodiment where (‘101: 6:5-9) the user pre-selects the text does
`
`not override the multitudes of other referents to translation in the specification without this
`
`
`
` 5
`
`restriction.
`
`
`
`AOL Ex. 1013
`Page 5 of 15
`
`

`

`B.
`
`Second language
`
`
`
`15.
`
`In my opinion, in the context of the ‘101 and ‘154 patents, “second language” is
`
`meant to be inclusive of a language different from the first language, and other dialects of the
`
`first language. I was informed that the terms should be construed as their ordinary and
`
`customary meaning to a person of ordinary skill in the art at the time of the invention, consistent
`
`with the claim language, written description, and prosecution history. For instance, Croatian and
`
`Serbian are different dialects of Serbo-Croatian, yet there are translators from one to the other
`
`(including Google translate (e.g. typing “translation from Croatian to Serbian” in the Google
`
`search box brings up exactly such a translator with the “second language” with Serbian, Serbian
`
`and Croatian being dialects of the same language, Serbo-Croatian)). For example, Google
`
`translate will translate Croatian “Dobro jutro. Drago mi je” into Serbian “Добро јутро. Драго
`
`ми је,” which roughly means, “Good morning. Pleased to meet you.” Attached hereto as
`
`Exhibit J. The ‘101 specification explicitly cites an example of Chinese dialects (‘101, 5:44-47),
`
`and hence there is both intrinsic and extrinsic evidence to support the broader definition of
`
`second language. Other extrinsic evidence shows that translation includes dialects, such as
`
`translating between Mandarin and Cantonese which are both dialects of the same language,
`
`Chinese, e.g. the second language may be a dialect of the first. See Zhang “Dialect MT: A Case
`
`Study between Cantonese and Mandarin” (1998), Attached hereto as Exhibit B; Petras
`
`“Translating Dialects in Search: Mapping between Specialized Languages of Discourse and
`
`Documentary Languages” (2006) at 111-153, Attached hereto as Exhibit C.
`
`C.
`
`Content word
`
`16.
`
` “Content word” is a basic concept in language and linguistics. A “content word”
`
`is any word that carries semantic content. Merriam Webster defines “content word” as “a word
`
`
`
` 6
`
`AOL Ex. 1013
`Page 6 of 15
`
`

`

`
`
`that primarily expresses lexical meaning” (http://www.merriam-
`
`webster.com/dictionary/content%20word). Attached hereto as Exhibit D. Words that carry
`
`meaning (i.e. semantic content) include nouns, verbs, adjectives and adverbs, but also include
`
`proper nouns (names), quantifiers (e.g. cardinal and ordinal numbers), qualifiers (common in
`
`Chinese, for instance), neologisms, and artificial words such as part numbers or product codes,
`
`such as “EOS-ID X”. For a more complete list and explanation of parts of speech, see Santorini,
`
`“Part-of-Speech Tagging Guidelines for the Penn Treebank Project (3rd Revision)” (1990).
`
`Attached as Exhibit E. Search engines also cope with model numbers, acronyms, etc. which go
`
`beyond the traditional parts of speech. Under a strict interpretation of “nouns, verbs, adjectives
`
`and adverbs” search queries such as “EOS-1D X under $4000” would have no content words at
`
`all – it is a model number, a preposition and a cardinal (numerical quantity, i.e.(a price), and yet
`
`if typed to Google the above query retrieves appropriate pages, and hence contains content
`
`words. Attached hereto as Exhibit F Search engines typically operate on what they consider
`
`content words.
`
`17.
`
`The ‘154 at 4:41-45 recites: “in a typical deployment, as soon as a content word
`
`or a keyword is extracted from the user’s query input, the server conducts a search in the
`
`database and returns to the user one or more advertisements relevant to the content word or
`
`keyword” which implies that data-base searchable key words are used, and these include names,
`
`model numbers, quantities, etc., -- whatever is defined as a key field in the data base, and not
`
`restricted to just nouns, verbs, adjectives and adverbs. For example, the PC Magazine
`
`encyclopedia defines “key field” as: “A field in a record that holds unique data which identifies
`
`that record from all the other records in the file or database. Account number, product code and
`
`
`
` 7
`
`AOL Ex. 1013
`Page 7 of 15
`
`

`

`
`
`customer name are typical key fields.” (http://www.pcmag.com/encyclopedia/term/45766/key-
`
`field) Attached as Exhibit G.
`
`18.
`
`Therefore the more general definition of content word is warranted, correct, and
`
`commonly used in the practice.
`
`D.
`
`Contextual search / contextually searching
`
`19.
`
`In my opinion “contextual search” applies to web search both in normal ordinary
`
`usage and in terms of intrinsic evidence (e.g. ‘101, 1:16-20), and also explicit mentions of URLs
`
`in the patent, which are identifiers of web-based document, providing access thereto, rather than
`
`being typical database designators. Hence the construction “Identification of/identifying relevant
`
`domain-unlimited set of documents available on the World Wide Web, based on the words
`
`contained in the documents” is appropriate. I note that the construction proposed by AOL
`
`referring to “external set of unidentified documents” could be consistent with Improved Search’s
`
`construction, but could also be problematic. External to what? If external means external to the
`
`user’s computer, including the web, then it could make sense, but “external” could mean
`
`“outside the company” or “outside the country” and thus exclude important parts of the web.
`
`Also the word “unidentified” is confusing in the context of these patents, unidentified by whom?
`
`If it means documents not previously identified by the user, it is acceptable, but it could mean
`
`“documents not previously identified by the search engine” which would rule out web search – a
`
`search engine must crawl and index the web, before searches are conducted, and in the process
`
`identifies the documents and corresponding URLs. Hence in my opinion AOL’s proposed
`
`construction is fraught with indeterminacy, and the much better stated Improved Search
`
`construction should be adopted.
`
`
`
`
`
` 8
`
`AOL Ex. 1013
`Page 8 of 15
`
`

`

`
`
`E.
`
`Dialectal standardization / dialectally standardizing / dialectally standardized
`
`20.
`
`In my opinion dialectical standardization includes both standardization within a
`
`dialect and across dialects of a language, therefore “to map keywords from different styles and
`
`dialects into standard and less ambiguous keywords” is the correct construction. The intrinsic
`
`evidence is clear to one of ordinary skill in the art reviewing the patents. For example, ‘101 5:63-
`
`37 gives variants for standardization including “auto”, “automobile” and “transportation vehicle”
`
`all of which are in Standard American English (e.g. within-dialect standardization, not requiring
`
`cross-dialect mapping). Other within dialect variants would include “car” and “motorcar.” Other
`
`examples of within-dialect standardization are “delimit”, “demarcate” and “differentiate”, or
`
`“airplane”, “plane”, and “aircraft.” It makes sense to standardize to “airplane”, which is less
`
`ambiguous (“plane”, for instance, can also refer to a carpenter’s tool, to a geometric figure, or
`
`other meaning). Typing “planes and level” to Google, yields a mix bag of results: geometric
`
`figures, carpenter’s tools, level flight, etc. (Attached as Exhibit H), whereas typing “airplanes
`
`and level” yields only flight-related results (Attached as Exhibit I).
`
`21.
`
`The specification also gives examples of cross-dialect standardization (e.g. “lorry
`
`vs. truck”), hence both within-dialect and cross-dialect are including in dialectical
`
`standardization, not just the latter, as would be the case if the unjustifiably narrow AOL
`
`construction were adopted.
`
`F.
`
`Search in the second language / searching in the second language
`
`22.
`
`The meaning of “search in the second language” should be self-evident as
`
`searching using words that are used in that second language. However there is an important
`
`subtlety that ruins the normal search engine process if the construction adds the word “only.”
`
`“Only words in the second language” could and probably would be interpreted to exclude words
`
`that may be used in that second language but are not properly part of said language. For instance
`
`
`
` 9
`
`AOL Ex. 1013
`Page 9 of 15
`
`

`

`
`
`in English we may search for a “Sushi restaurant” or for a “Tapas restaurant” or for “Toyota
`
`automobiles”, or for “haute couture fashion”, but all of those searches could be excluded because
`
`they contain words that are not “only” in the second language (in English, in this case). A more
`
`comprehensive example: Suppose a Japanese person visiting the United States types “すし屋”
`
`(sushi-ya) as a query in Japanese, the query translator turns this into “sushi restaurant” and this
`
`query in the second language is passed to the contextual search engine, which finds sushi
`
`restaurant web pages returning same (or translations thereof into Japanese). This would be a
`
`typical use of the invention, but would be excluded if “sushi” (written in roman letters, but still a
`
`Japanese word) were excluded from the search in the second language because said search in the
`
`second language may use only words that in that language. The same would happen if the initial
`
`query would contain “トヨタ” (Toyota), because it is a Japanese word and therefore forbidden
`
`in a search (e.g. for “Toyota dealers”) that can only use English words (Toyota is a Japanese
`
`word, even if written in roman characters).
`
`23.
`
`Therefore, adding the word “only” to the construction is an arbitrary and
`
`unwarranted restriction whose sole function is to significantly reduce the value of the invention.
`
`The correct construction is simply “searching using words in the second language”.
`
`G.
`
`Advertising cues
`
`24.
`
`The ‘154 patent is clear in its use of the term “advertising cues” to mean
`
`advertisements, links to advertisements (e.g. URLs/hyperlinks) or other references to
`
`advertisements. Hence AOLs proposed construction of that term to just “advertisements” is in
`
`my opinion unwarranted. For instance ‘154 claims 2 and 8 mention examples of signals of
`
`references to advertisements. A hyperlink is an example of a reference. The specification of
`
`the‘154 at 6:29-34 mentions “a hyperlink to an advertisement page”, “a pop-up window” “a
`
`
`10
`
`AOL Ex. 1013
`Page 10 of 15
`
`

`

`
`
`flag”, “audio advertisement” and “non-textual visual advertisement”. All of these are
`
`“advertising cues”; essentially the same information is repeated in ‘154 7:25-30. Claim 2 of the
`
`‘154 recites: “The method of claim 1, wherein the advertising cues comprise any of: a hyperlink
`
`to an advertisement page in the first language; a pop-up window containing content in the first
`
`language; a flag containing content in the first language; an audio advertisement; and a non-
`
`textual visual advertisement”, making it even clearer that advertising cues contain signals and
`
`references (flags and hyperlinks).
`
`H. Means for receiving from the user through an input device a query in the
`first language
`
`25.
`
`It is my understanding that a “means plus function” claim or claim element entail
`
`first specifying the function and then identifying the structure that supports that function, either
`
`in the claim itself or in the specification. The function is clearly stated in a manner a POSITA
`
`would understand: “receiving from the user through an input device a query in the first
`
`language.” The structure is provided by the specification: that “the user inputs a query in her
`
`native language (i.e. the source language) through an input device such as a keyboard.”
`
`(emphasis added) in ‘154 1:60-61; 4:53-55; 6:39-41. Moreover a keyboard is the standard means
`
`for inputting information (queries) to a search engine. Hence the structure is “a keyboard or its
`
`equivalents”.
`
`I.
`
`Dialectal controller for dialectally standardizing a content word extracted
`from the query
`
`
`26.
`
` “Dialectal controller” is definite in light of the claim language, specification and
`
`prosecution history. The ‘101 patent states at 5:27-33: “The query is received by a dialectal
`
`controller which processes the query and identifies a keyword from the query input 120. The
`
`dialectal controller extracts content word out of the query. The next step involves dialectal
`
`
`11
`
`AOL Ex. 1013
`Page 11 of 15
`
`

`

`
`
`standardization 122, wherein the dialectal controller at server backend picks up the keyword and
`
`standardizes it to a commonly known word and/or term.” The specification further defines
`
`“dialectal controller” at 7:9-16: “The dialectal controller uses processing logic to identify the
`
`keyword 152. Statistical data in conjunction with syntactic analysis provides the foundation for
`
`the processing logic so as to include and exclude certain kind of verbal entries. Thereafter, the
`
`dialectal controller applies dialectal standardization logic to standardize keyword 154. Such a
`
`logic is used so as to standardize the keyword to a commonly known word/term.” A person of
`
`ordinary skill in the art would understand that “dialectal controller” is a server implementing
`
`processing logic using statistical data in conjunction with syntactic analysis to dialectally
`
`standardize the keyword to a commonly known word, and is thus sufficiently definite.
`
`27.
`
`If this term were to be governed as a means-plus-function element, then the
`
`function is clearly stated in the phrase “dialectally standardizing a content word extracted from
`
`the query,” and the corresponding structure evident to a POSITA reading the specification is a
`
`“server.”
`
`28.
`
`Support for this structure is evident in the ‘154 at 4:37-39, which recites: “The
`
`server, which is connected to a search engine through the Internet, hosts a dialectal controller”;
`
`and at 4:55-63, which recites: “Step 102: The input is received by a dialectal controller in the
`
`server which processes the query input, identifies the user’s input language, and extracts a
`
`content word or keyword out of the query input. The dialectal controller at the server backend
`
`picks up the keyword and standardizes it to a commonly known word or term. This is done to
`
`bring about a consistency in the meaning of a word notwithstanding dialectal variations.”
`
`29.
`
`The specification of the ‘101 patent further adds to this structure, e.g. ‘101 patent
`
`at 7:10-16 states that” statistical data in conjunction with syntactic analysis” at the server is the
`
`
`12
`
`AOL Ex. 1013
`Page 12 of 15
`
`

`

`
`
`structure that implements the “processing logic” supporting, and implementing, the function of
`
`dialectally standardizing a content word extracted from the query. Hence the structure can be
`
`summarized as “server.”
`
`J.
`
`Means to search the database of the advertising cues based on the relevancy
`to the translated content word
`
`
`30.
`
`In my opinion the function would be evident to a POSITA from the description,
`
`namely: “to search the database of the advertisement cues based on the relevance to the
`
`translated content word.” A POSITA reading the specification would conclude that the structure
`
`is a “server.” The ‘154 at 4:39-45 recites: “The server is also associated with a database of
`
`advertisements. In a typical deployment, as soon as a content word or a keyword is extracted
`
`from the user’s query input, the server conducts a search in the database and returns to the user
`
`one or more advertisements relevant to the content word or keyword.” It is typical for a remote
`
`server-based database search to be performed by the server running a database manager (the
`
`DBMS) that searches the database. Server-based databases are not useful without a search
`
`capability, and by definition the DBMS performs that search. Hence the server runs a database
`
`search, implementing this function.
`
`31.
`
`Further structure is provided via the LACE embodiment, ‘154 7:32-44: “In
`
`another equally preferred embodiment, the cross language advertising is incorporated with the
`
`LACE, a system for dynamically returning a remote online user a bilingual annotation or
`
`translation, displayed in a mouse pointer associated callout, on the textual information contained
`
`in the website. When the user initiates a real-time annotation or translation using her mouse
`
`pointer, he is returned one or more advertisements in the user’s language. The system includes a
`
`web server which supports a website on the Internet,” such as English, and “[t]he LACE
`
`application can be activated from the web site but runs on the web site server.” The user
`
`
`13
`
`AOL Ex. 1013
`Page 13 of 15
`
`

`

`
`
`communicates with LACE and with any search engine and advertisement database via a browser.
`
`A POSITA would assemble these elements to see that the keywords come from the user’s query
`
`via the browser, and go to the server for database search to find the relevant advertisements or
`
`advertisement cues, which are then served back to the user’s browser for viewing.
`
`K. Means to send the search results and the matching advertising cues
`to the user’s computer screen
`
`
`32.
`
`A POSITA would understand that the function is contained in the statement,
`
`namely “to send the search results and the matching advertising cues to the user’s computer
`
`screen.” In the context of the client-server framework provided in the specification of the ‘154
`
`the POSITA would further understand that the structure inherently would be the server sending
`
`the search results to the browser, which then displays it on the computer screen. This is the way
`
`browsers and servers operate. The description of the LACE server referenced above, reinforces
`
`this structure more explicitly:
`
`“The remote end user logs on the Internet by using a browser in her
`computer and visits a website. The website is in a target language, such
`as English. The LACE application can be activated from the web site but
`runs on the web site server. Upon activation of the LACE application,
`the user can obtain translation of or bilingual annotation on, a segment
`of textual information in the website by moving her mouse pointer over,
`or pointing the pointer at, the text that she wants to understand. For
`example, when the user moves the pointer over “tax preparer”, a bubble
`or a pop-up callout comes to the screen. The callout is associated with
`the pointer such that a visual reference between the callout and the target
`text is established. At the same time, the server sends the user one or
`more advertisements which are relevant to the text she targeted.” (‘154
`at 7:41-55.)
`
`
`The LACE description explicitly mentions a “bubble or pop-up callout comes to the [user’s]
`
`screen” which as the POSITA knows is done via the browser explicitly mentioned above “the
`
`user … using a browser in her computer.” And the “web site server” sends this information to
`
`
`14
`
`AOL Ex. 1013
`Page 14 of 15
`
`

`

`
`
`the browser. Hence it should be completely clear to a POSITA that the structure is “the server
`
`and/or browser or equivalents.”
`
`I declare under penalty of perjury that the foregoing is true and correct.
`
`
`
`Dated: Nov 3, 2016
`
`
`Jaime Carbonell, Ph.D
`
`
`
`
`15
`
`AOL Ex. 1013
`Page 15 of 15
`
`