`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`______________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`______________
`
`AMERICAN EXPRESS COMPANY, AMERICAN EXPRESS TRAVEL
`COMPANY, INC., EXPEDIA, INC., HOTELS.COM LP, HOTELS.COM
`GP, LLC, HOTWIRE, INC., ORBITZ WORLDWIDE, INC.,
`PRICELINE.COM, INC., TRAVELOCITY.COM LP, and YAHOO! INC.
`Petitioner
`v.
`
`METASEARCH SYSTEMS, LLC.
`Patent Owner
`
`______________
`
`Case No. CBM2014‐00001
`Patent Number 8,326,924 B1
`
`______________
`
`Before the Honorable HOWARD B. BLANKENSHIP, KARL D. EASTHOM, and
`BARBARA A. BENOIT, Administrative Patent Judges.
`
`
`DECLARATION OF DR. JAIME CARBONELL
`IN SUPPORT OF PATENT OWNER METASEARCH SYSTEMS, LLC’s MOTION TO
`AMEND
`
`
`
`
`84941757.1
`
`1
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-1
`
`
`
`Declaration of Dr. Jaime Carbonell
`Qualifications
`
`1.
`
`I am presently a chaired professor at Carnegie Mellon University’s
`
`School of Computer Science and Director of the Language Technologies Institute.
`2.
`
`I am also presently an adjunct faculty member at the University of
`
`Pittsburgh Medical School’s Department of Computational Biology.
`3.
`
`I received my B.S. in Physics and Mathematics from the
`
`Massachusetts Institute of Technology in 1975. I received my MS in Computer
`
`Science and my Ph.D. in Computer Science from Yale University in 1976 and 1979,
`
`respectively.
`4.
`
`I have been a faculty member of Carnegie Mellon University since
`
`1979. While at Carnegie Mellon University, I served as Director of the Center for
`
`Machine Translation from 1986 to 1996, which evolved into the Language
`
`Technologies Institute in 1996. I am also a co‐Founder and Board Chairman of
`
`Carnegie Speech Incorporated, and was Co‐Founder and Board Chairman of
`
`Wisdom Technologies Corporation.
`5. My research interests and areas of expertise include data mining,
`
`natural language processing, search engines, electronic commerce, databases,
`
`algorithms, and machine learning.
`6.
`
`I have authored or co‐authored over 330 publications and have given
`
`over 500 invited or refereed paper presentations. During the late 1990s and early
`
`2000s, I published or presented on a number of topics related to database
`
`
`84941757.1
`
`2
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-2
`
`
`
`searching and data mining. This includes Ralf D. Brown, Thomas Pierce, Yiming
`
`Yang, and Jaime G. Carbonell, ‘‘Link Detection ‐ Results and Analysis,’’ Topic
`
`Detection and Tracking Workshop, 1999; Carbonell, J., and Goldstein, J., ‘‘MMR
`
`and Diversity‐Based Reranking for Reordering Documents and Producing
`
`Summaries,’’ Proceedings of the 21st meeting of International ACM SIGIR
`
`Conference, Melbourne, Australia, August 1998, pp. 335‐336; Goldstein, J. and
`
`Carbonell, J., Carbonell. J. G., Yang, Y,. Frederking, R. E., Brown, R., Geng, Y., and
`
`Lee, D., ‘‘Translingual Information Retrieval: A Comparative Evaluation,’’
`
`Proceedings of The International Conference on Artificial Intelligence, Nagoya,
`
`Japan, 1997, (Distinguished paper award). and Carbonell. J. G., Goldstein, J,.
`
`Yibing, G., ‘‘Automated Query‐Relevant Summarization and diversity‐based
`
`reranking,’’ Proceedings of the IJCAI‐97 workshop on AI in Digital Libraries, Nagoya,
`
`Japan, 1997.
`7.
`
`During my tenure at Carnegie Mellon University, I was a Ph.D. advisor
`
`to Michael Mauldin, the founder of Lycos Inc. (an early web search engine), and
`
`co‐advisor to Oren Etzioni, founder of MetaCrawler.
`8.
`
`I regularly teach courses in search engines, data mining and
`
`electronic commerce at CMU.
`9.
`
`I have been a member of numerous professional organizations
`
`including the Association for Computing Machinery (ACM), the Association for
`
`Computational Linguistics (ACL), and the Association for the Advancement of
`
`Artificial Intelligence (AAAI). I was elected chair of ACM’s Special Interest Group
`
`
`84941757.1
`
`3
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-3
`
`
`
`on Artificial Intelligence (SIGART) from 1983 to 1985. I have been an AAAI Fellow
`
`since 1988.
`10.
`
`I have participated in a number of governmental groups. To name a
`
`few, I have been a member of the National Science Foundation’s Directorate for
`
`Computer and Information Science and Engineering Advisory Committee since
`
`2010. From 1988 to 1992 I was a member of the National Institute of Health’s
`
`Human Genome Scientific Advisory Committee. I was also a member of the
`
`Scientific Advisory Committee to the National Institute of Standards and
`
`Technology (NIST) that focused on speech, language and information retrieval
`
`from 1997 to 2001.
`11. My industry experience included consulting in support of the launch
`
`of the Lycos Inc. internet search engine in 1998‐1999, and consulting for the
`
`launch of the search engine, Searchline, from 2002 to 2004. I also served on
`
`Citibank’s technology advisory board for 10 years, focusing on text and data
`
`mining, fraud detection, and optimization. In total, I have designed or co‐designed
`
`three data mining engines, and evaluated and helped improve numerous other
`
`data mining engines, including the machine learning algorithms that they are
`
`based on.
`12.
`
`I am a named inventor on several issued patents including U.S.
`
`Patent No. 5,995,920, “Computer‐based method and system for monolingual
`
`document development,” and U.S. Patent No. 6,163,785, “Integrated authoring
`
`and translation system.”
`
`
`84941757.1
`
`4
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-4
`
`
`
`13.
`
`I am being compensated by counsel at my rate of $550/hour plus
`
`reimbursement for normal expenses incurred. I have no financial interest in the
`
`outcome of the related litigations or this proceeding.
`
`Scope of this Declaration
`
`14.
`
`I offer this declaration in support of Patent Owner’s Motion to Amend
`
`claims of U.S. Patent No. 8,326,924 (the “’924 patent”). I first address
`
`construction of the term “database,” which is used in the proposed substitute
`
`claims, claims 13 and 14. Then I address how claims 13 and 14 do not claim an
`
`abstract idea.
`
`Claim Construction of “Database”
`
`15.
`
`Substitute claims 13 and 14 use the phrase “ database.” I understand data
`
`base to have its common computer‐science meaning, namely an organized
`
`electronic repository of structured, semi‐structured or unstructured information.
`
`The patent’s specification and drawings show a database that can store previously
`
`searched information. See Fig 53A and B, and Ex. 2001 at p. 243, Col. 36, l. 54‐Col.
`
`37, l. 32.; Ex. 2117 at p. 177, l. 11‐ p. 179, l. 23. for support). Generally speaking,
`
`databases store data for subsequent rapid use. The use of the database, claimed
`
`in substitute claim 13 (Exhibit 2018) and substitute claim 14 (Exhibit 2019)
`
`conforms to this general rule.
`16.
`
` If the entries in the database are recent (up to date), then such a
`
`database saves time, rendering the metasearch more efficient. If it is not recent,
`
`its results can still be combined with those from external searches, though that
`
`
`84941757.1
`
`5
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-5
`
`
`
`combination process requires more inventiveness as to how and when to
`
`combine the results with those from on‐the‐fly external engines.
`17.
`
`The construction of database specified above conforms to the plain
`
`meaning given to that term by one of ordinary skill in the art in 2000. Databases
`
`have been used for over 30 years to act as a repository that stores previously
`
`processed information, including accessed data, sometimes recently accessed
`
`data and other times all previously accessed data. One example of such
`
`databases circa 2000 are “materialized views” of database computations such as
`
`multi‐source joins, which are popular in the structured database community (e.g.,
`
`see Exhibits 2020‐2022).
`
`
`The Substitute Claims’ Steps using the Database Further Establish that the Claims
`do not Claim an Abstract Idea
`The invention does not claim an abstract idea, for at least two reasons.
`
`First, adding hardware—a database‐‐to the previously‐claimed process,
`
`18.
`19.
`
`creates a technological invention. Handling structured, unstructured, and
`
`semistructured information is a significant technological and analytical challenge.
`
`Much of the data available on the web is unstructured data, and by storing the
`
`data in a database, this data would inevitably become structured or semi‐
`
`structured. Modern databases all use structured data (e.g. relational tables), or
`
`semi‐structured data (e.g., indexed and/or parsed text, XML, etc.). Storing such
`
`fields as time of the initial query, the content of the initial query, the external
`
`source (provenance) of the search results, and likely other fields in a database
`6
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`
`84941757.1
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-6
`
`
`
`would force structure on the (previously unstructured) data.
`20. By giving the database structure, a person of ordinary skill in the art would
`
`appreciate that many benefits would flow. For example, data that comes from a
`
`slow or unreliable web site could be presented reliably to a user.
`21.
`
`Second, the process of creating, maintaining and using a database is rather
`
`complex but is very well specified. Handling structured, semistructured, and
`
`unstructured information is a significant technological and analytical challenge. It
`
`requires multiple sequential well defined steps, including: 1) Retaining and
`
`indexing the previous query; 2) determining what information to include in the
`
`database, such as the results of previous searches that were ranked sufficiently
`
`high by the metasearch engine to be included in the combination shown to the
`
`user; 3) indexing these results with the terms of the original query as the key‐
`
`fields for rapid retrieval if and when the same query or another query with the
`
`same terms is issued by a user; 4) Storing the data and time of the initial query,
`
`linked to the query results; 5) Storing the external source (provenance) of the
`
`search results, linked to the query term and the results themselves; 6) upon an
`
`new query with the same terms, determine whether to use the d database based
`
`on: metasearch engine load (s are used to reduce load), recency of information in
`
`the (comparing the current time and data with the stored time and date), and
`
`availability of external information sourced (e.g. if an external search engines
`
`times out – meaning it is not available or is too heavily loaded – then its database
`
`results can be used instead); 7) combining the database results with the “fresh”
`
`
`84941757.1
`
`7
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-7
`
`
`
`results provided by the external data sources and search engines.
`22. As I have explained above, the substitute claims require a specific
`
`computer—a metasearch engine—capable of receiving specific types of
`
`requests—“Hypertext Transfer Protocol” requests—from a user device for the
`
`purposes of querying a plurality of heterogeneous resources. Thus, the originally‐
`
`claimed invention is a fairly general but well defined method, rather than an
`
`abstract concept. But with the addition of the d database limitation to claims 13
`
`and 14 it is considerably more detailed, concrete, and specific
`23.
`
` Under the substitute claims, the use of advertisements and online ordering
`
`would not be foreclosed for all metasearch sites. One would be free to provide
`
`advertisements and online ordering for metasearch sites which did not use a
`
`database as claimed in the two proposed substitute claims. Advertisements could
`
`be contained in the database or could be passed through from the search results
`
`of the individual search engines accessed by the metasearch engine, or could
`
`come from a separate advertisement server. These marketing concepts could be
`
`applied to distributed search engines that only handle one type of information
`
`(structured, unstructured, semistructured).
`
`
`
`
`
`
`
`84941757.1
`
`8
`Declaration of Dr. Jaime Carbonell – Motion to Amend
`
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-8
`
`
`
`I hereby declare under penalty of perjury under the laws of the United
`
`States of America that the forgoing is true and correct, and that all statements
`
`made of my own knowledge are true and that all statements made on
`
`information and belief are believed to be true. I understand that willful false
`
`statements are punishable by fine or imprisonment or both (18 U.S.C. § 1001).
`
`Executed on the 15th day ofJuly, 2014, in Pittsburgh, Pennsylvania.
`
`
` '
`
`" .me Carbonell, Ph.D.
`
`9
`
`Declaration of Dr. Jaime Carbonell — Motion to Amend
`
`84941757.]
`
`AMERICAN EXPRESS V. METASEARCH
`
`CBM2014-00001 EXHIBIT 2026-9
`
`AMERICAN EXPRESS v. METASEARCH
`CBM2014-00001 EXHIBIT 2026-9
`
`