Responsive Declaration of H. V. Jagadish
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`BEFORE THE PATENT TRIAL AND APPEALS BOARD
`
`INTERNATIONAL BUSINESS MACHINES CORPORATION
`Petitioner,
`
`v.
`
`INTELLECTUAL VENTURES I LLC,
`Patent Owner
`
`Case No.: IPR2015-01481
`Patent No.: 6,510,434
`
`
`RESPONSIVE DECLARATION OF H. V. JAGADISH
`
`
`
`
`
`
`IBM Ex. 1022
`IBM v. IV
`IPR2015-01481
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`BEFORE THE PATENT TRIAL AND APPEALS BOARD
`
`INTERNATIONAL BUSINESS MACHINES CORPORATION
`Petitioner,
`
`v.
`
`INTELLECTUAL VENTURES I LLC,
`Patent Owner
`
`Case No.: IPR2015-01481
`Patent No.: 6,510,434
`
`
`RESPONSIVE DECLARATION OF H. V. JAGADISH
`
`
`
`
`
`
`IBM Ex. 1022
`IBM v. IV
`IPR2015-01481
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`I, H. V. Jagadish, do hereby declare as follows:
`
`1.
`
`I previously prepared and executed a declaration (Ex. 1001) in IPR2015-
`
`01481. This declaration responds to arguments raised in IV’s Patent Owner
`
`Response (Paper 17) and Dr. Papakonstantinou’s declaration (Ex. 2001).
`
`2.
`
`In preparing this Responsive Declaration, I reviewed and considered the
`
`following:
`
` The Board’s Decisions Instituting Inter Partes Review (Paper 12)
`
` IV’s Patent Owner Response (Paper 17)
`
` Declaration of Dr. Papakonstantinou (Ex. 2001)
`
` Deposition testimony of Dr. Papakonstantinou (Ex. 2012)
`
`This material is in addition to the material I reviewed and considered when
`
`preparing my original Declaration.
`
`I. Wical’s Knowledge Base and Directed Graph Are Not One and the
`Same
`
`3.
`
`I understand IV and Dr. Papakonstantinou argue that Wical’s knowledge
`
`base and directed graph are one and the same data structure. This is
`
`incorrect and unsupported by Wical. Wical explains that the knowledge
`
`base comprises categories and terminology based on the arrangement of
`
`categories in a knowledge catalog. Ex. 1006 at 6:7-11. The knowledge
`
`catalog is a baseline arrangement of categories reflecting “a world view of
`
`knowledge.” Ex. 1006 at 5:59-63. The knowledge base is then “augmented
`
`1
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`to include linking and cross referencing among categories for which a
`
`linguistic, semantic, or usage association has been identified.” Ex. 1006 at
`
`11:36-38. These cross-references and links are stored using a directed graph
`
`data structure. Ex. 1006 at 11:56-65.
`
`4. Wical explains that the directed graph data structure describing relationships
`
`among categories is only a part of the entire knowledge base. Ex. 1006 at
`
`11:56-57. This is consistent with Wical’s teaching that the foundation of the
`
`knowledge base is an arrangement of categories based on a knowledge
`
`catalog. Ex. 1006 at 6:7-11, 11:14-17. Wical explains that relationship
`
`information is added to this existing structure using a directed graph. Ex.
`
`1006 at 11:37-38, 11:56-65. Thus a portion of the knowledge base is created
`
`before the directed graph portion describing relationships among categories
`
`is created. A person of ordinary skill at the time of the alleged invention
`
`claimed in the ’434 patent would have understood that Wical’s knowledge
`
`base and its directed graph portion are not one and the same.
`
`II. Wical Discloses Related High Level Categories
`
`5.
`
`I understand IV and Dr. Papakonstantinou argue that high-level categories in
`
`Wical such as “Geography” and “Leisure and Recreation,” as described in
`
`Fig. 4 of Wical, cannot be related because they represent independent
`
`ontologies. This is incorrect because Wical expressly discloses relationships
`
`2
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`between categories in different ontologies. For example, regarding the
`
`example described in Fig. 4, Wical expressly discloses that the category
`
`“France,” listed in the “Geography” ontology, is “cross referenced and/or
`
`linked” to categories in the “Leisure and Recreation” ontology. Ex. 1006 at
`
`11:36-45.
`
`6.
`
`I understand IV and Dr. Papakonstantinou argue that creating a relationship
`
`between two top level categories in different ontologies would undermine
`
`Wical’s purpose of establishing separate ontologies. This is also incorrect.
`
`Wical’s system seeks to “construct a search and retrieval system that . . .
`
`generates a similar response for different queries that have similar
`
`meanings.” Ex. 1006 at 1:48-56. Wical achieves this goal in part by
`
`organizing
`
`information using multiple classification hierarchies or
`
`ontologies. Ex. 1006 at 6:7-11, 11:14-35, Fig. 8C. This gives the search and
`
`retrieval system multiple perspectives and allows for more accurate
`
`searches. Recognizing relationships among categories across different
`
`classification hierarchies or ontologies are essential to this result.
`
`7.
`
`Nor does Wical’s arrangement of classification categories in separate
`
`ontologies preclude relationships among high-level categories. In fact,
`
`Wical expressly discloses traveling up the relationship hierarchy when
`
`identifying categories relevant to a user’s search. Ex. 1006 at 14:1-36. For
`
`3
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`example, using Fig. 6, Wical describes a method of identifying additional
`
`related categories relevant to a user’s search. Ex. 1006 at 14:1-36. Here,
`
`Wical examines the strength of the relationships between the category
`
`labeled “NodeA” and other categories in the directed graph. Ex. 1006 at
`
`14:1-36. Specifically, Wical discloses traveling up the relationship
`
`hierarchy from NodeY to NodeX and considers whether NodeX should be
`
`included in a search starting from NodeA. Ex. 1006 at 14:1-36.
`
`
`
`In this example, NodeX is excluded from the search because the relationship
`
`between it and NodeA is too weak. Ex. 1006 at 14:1-36. But Wical clearly
`
`discloses a relationship path between NodeA and NodeX. By recognizing
`
`relationships with categories higher in the classification hierarchy, Wical’s
`
`system ensures that documents classified more generally will also be located
`
`4
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`when they are relevant to a user’s search. Thus a person of ordinary skill at
`
`the time of the alleged invention claimed in the ’434 patent would have
`
`understood that Wical discloses relationships between all levels of the
`
`directed graph and does not preclude relationships between higher level
`
`categories through their sub-categories.
`
`8.
`
`I understand that Dr. Papakonstantinou also argues that relationships
`
`between Wical’s high-level categories would create inefficiencies because
`
`such a relationship would create a loop in Wical’s directed graph. This is
`
`incorrect because Wical’s disclosed system is not incompatible with cross-
`
`references and/or links between categories that result in loops across the
`
`directed graph. In fact, Wical discloses that its directed graph may contain
`
`loops. Ex. 1006 at Fig. 4. In Fig. 4 for example, Wical explains that “the
`
`categories ‘France’, ‘art galleries and museums’, and ‘places of interest’ are
`
`cross referenced and/or linked.” Ex. 1006 at 11:39-41. As can be seen in
`
`Fig. 4 below, the cross-references and links between these categories create
`
`a loop between “France” and “Places of Interest” through the category “Art
`
`Galleries & Museums.”
`
`5
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`9.
`
`The concept of directed graphs has been well known since before December
`
`
`
`1999. A person having ordinary skill in the art would have been familiar
`
`with directed graph data structures as well as the algorithms used to traverse
`
`directed graphs in a variety of computing solutions. A person of ordinary
`
`skill would have also known how to detect and account for loops in a
`
`directed graph, including the directed graph described in Wical.
`
`10.
`
`I understand that IV and Dr. Papakonstantinou argue that the high-level
`
`categories “Geography” and “Leisure and Recreation” cannot be related
`
`because the categories are too remote. This is incorrect because one of
`
`ordinary skill in the art would consider these categories related. Wical’s
`
`6
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`system differentiates between relationships between categories and the
`
`strength of those relationships. 1006 at 12:18-22. The links and cross-
`
`references between categories “further include distance weights,” and these
`
`“distance weights provide a quantitative value to indicate the strength of the
`
`linguistic, semantic, or usage relationship between two categories/terms.”
`
`Ex. 1006 at 12:18-22. More documents associating two categories results in
`
`higher distance weights between those two categories. Ex. 1006 at 12:24-
`
`45. When there are not enough documents in the document set being
`
`processed that associate any two categories, Wical discloses using an
`
`“association marker” to denote a weak relationship between those two
`
`categories. Ex. 1006 at 12:46-62. In Wical’s system, a variety of variables
`
`may be used to determine the strength of the relationship between two
`
`categories, including the distance weight between them, association markers,
`
`and even whether the relationship was established based on manual
`
`identification by a linguist or through automated document processing. Ex.
`
`1006 at 13:45-14:55. But Wical does not suggest that categories that share a
`
`weak relationship are unrelated. For example, as I explained in ¶ 7 above,
`
`even where categories share a weak relationship, Wical discloses a
`
`relationship path between them.
`
`7
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`11.
`
`I understand that IV argues that it is my position that to consider Wical’s
`
`high-level categories to be related would render all relationships in Wical
`
`meaningless. That is not my position. As I explained in my earlier
`
`declaration, Wical discloses related high-level categories. Ex. 1001 at ¶ 136.
`
`For example, the high-level categories “Geography” and “Leisure and
`
`Recreation” in Wical’s Fig. 4 are related through their respective directed
`
`graph connections. Ex. 1001 at ¶ 136.
`
`12. During my deposition I further clarified that Wical’s system provides some
`
`flexibility in how one may traverse the directed graph. For example, one
`
`could implement an algorithm to traverse the directed graph that limits the
`
`number of steps taken from the starting category. But such implementation
`
`choices do not alter the basic architecture disclosed by Wical where high
`
`level categories can be related through their respective directed graph
`
`connections as is described in Wical’s Fig. 4 for the high level categories
`
`“Geography” and “Leisure and Recreation.”
`
`III. Morita Discloses Related Domain Tags
`
`13.
`
`I understand that IV and Dr. Papakonstantinou argue that Morita fails to
`
`disclose related domain tags. This is incorrect. As I explained in my earlier
`
`Declaration, Morita discloses creating keyword connection tables, which
`
`store keywords (i.e., tags) and information about how different keywords are
`
`8
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`related to each other. Ex. 1001 at ¶¶ 192-94, 220. Morita’s keyword
`
`connection table includes domain tags. Ex. 1001 at ¶¶ 192-94. For
`
`example, in Fig. 5, Morita shows a keyword connection table with
`
`relationship entries for the domain tag KWM-1 (“ENTERPRISE”). Ex. 1007
`
`at Fig. 5; Ex. 1001 at ¶ 220. All relationships for this domain tag are entered
`
`into the keyword connection table, including its relationships with other
`
`domain tags. Ex. 1007 at 3:36-4:3; Ex. 1001 at ¶ 237.
`
`14. Morita’s system does not prevent relationships between different domain
`
`tags. Morita includes relationship types such as “also known as,”
`
`“synonym,” “IS-A,” “IS-PART-OF,” etc. Ex. 1007 at 3:43-48. Nothing in
`
`Morita suggests that a keyword such as “ENTERPRISE” cannot also be
`
`related to another keyword that may itself be a grouping of other keywords.
`
`All relationships for a given keyword are recorded in the keyword
`
`connection table, including keywords that are domain tags. Ex. 1007 at
`
`3:36-4:3. Thus even though Morita’s figures show only one example of a
`
`domain tag, a person of ordinary skill would have understood that other
`
`domain tags related to the “ENTERPRISE” domain would be added to the
`
`keyword connection table.
`
`9
`
`
`
`
`
`Responsive Declaration of H. V. Jagadish
`
`IV. Conclusion
`
`15.
`
`I hereby declare under penalty of perjury under the laws of the United States
`
`of America that the foregoing 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. See 18 U.S.C.
`
`§ 1001.
`
`
`
`
`
`Date: June 20, 2016
`
`Respectfully submitted,
`
`
`
`
`
`
`
`10
Accessing this document will incur an additional charge of $.
After purchase, you can access this document again without charge.
Accept $ Charge
Still Working On It
This document is taking longer than usual to download. This can happen if we need to contact the court directly to obtain the document and their servers are running slowly.
Give it another minute or two to complete, and then try the refresh button.
A few More Minutes ... Still Working
It can take up to 5 minutes for us to download a document if the court servers are running slowly.
Thank you for your continued patience.
This document could not be displayed.
We could not find this document within its docket. Please go back to the docket page and check the link. If that does not work, go back to the docket and refresh it to pull the newest information.
Your account does not support viewing this document.
You need a Paid Account to view this document. Click here to change your account type.
Your account does not support viewing this document.
Set your membership
status to view this document.
With a Docket Alarm membership, you'll
get a whole lot more, including:
- Up-to-date information for this case.
- Email alerts whenever there is an update.
- Full text search for other cases.
- Get email alerts whenever a new case matches your search.
One Moment Please
The filing “” is large (MB) and is being downloaded.
Please refresh this page in a few minutes to see if the filing has been downloaded. The filing will also be emailed to you when the download completes.
Your document is on its way!
If you do not receive the document in five minutes, contact support at support@docketalarm.com.
Sealed Document
We are unable to display this document, it may be under a court ordered seal.
If you have proper credentials to access the file, you may proceed directly to the court's system using your government issued username and password.
Access Government Site
