Case: 21-1179 Document: 48 Page: 1 Filed: 05/19/2022
`
`
`
`
`United States Court of Appeals
`for the Federal Circuit
`______________________
`
`GOOGLE LLC,
`Appellant
`
`v.
`
`IPA TECHNOLOGIES INC.,
`Appellee
`______________________
`
`2021-1179, 2021-1180, 2021-1185
`______________________
`
`Appeals from the United States Patent and Trademark
`Office, Patent Trial and Appeal Board in Nos. IPR2019-
`00728, IPR2019-00730, IPR2019-00731.
`______________________
`
`Decided: May 19, 2022
`______________________
`
`NAVEEN MODI, Paul Hastings LLP, Washington, DC,
`argued for appellant. Also represented by ARVIND JAIRAM,
`STEPHEN BLAKE KINNAIRD, JOSEPH PALYS, DANIEL
`ZEILBERGER; MICHAEL C. HENDERSHOT, Jones Day, Palo
`Alto, CA.
`
` STEVEN WAYNE HARTSELL, Skiermont Derby LLP, Dal-
`las, TX, argued for appellee. Also represented by JAIME
`OLIN, PAUL SKIERMONT, SARAH ELIZABETH SPIRES; MIEKE
`K. MALMBERG, Los Angeles, CA.
` ______________________
`
`
`

`

`Case: 21-1179 Document: 48 Page: 2 Filed: 05/19/2022
`
`2
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`Before DYK, SCHALL, and TARANTO, Circuit Judges.
`Dyk, Circuit Judge.
` Google LLC (“Google”) appeals three inter partes re-
`view (“IPR”) decisions of the Patent Trial and Appeal Board
`(“Board”) concluding that Google had not shown the chal-
`lenged claims in U.S. Patent Nos. 6,851,115 (“the ’115 pa-
`tent”) and 7,069,560 (“the ’560 patent”) to be unpatentable.
`Because the Board failed to resolve fundamental testimo-
`nial conflicts in concluding that the relied-upon reference
`was not prior art, we vacate the decisions and remand for
`further proceedings consistent with this opinion.
`BACKGROUND
`The ’115 and ’560 patents relate to a “software-based
`architecture . . . for supporting cooperative task completion
`by flexible, dynamic configurations of autonomous elec-
`tronic agents.” J.A. 95; see also J.A. 131. In particular, the
`patents disclose that “[c]ommunications and cooperation
`between agents are brokered by one or more facilitators”
`and that “[t]he facilitators employ strategic reasoning for
`generating a goal satisfaction plan to fulfill arbitrarily com-
`plex goals by users and service requesting agents.” J.A. 95.
`Both patents list David L. Martin and Adam J. Cheyer as
`inventors. Claim 22 of the ’560 patent, which is representa-
`tive for purposes of the discussion in this opinion, requires,
`among other things, a “facilitator agent” that performs var-
`ious functions. ’560 patent, col. 31, l. 61–col. 32, l. 16.
`The patent applications resulting in the ’115 and
`’560 patents were filed on January 5, 1999, and March 17,
`1999 respectively. The underlying technology, known as
`the Open Agent Architecture (“OAA”), was conceived at
`SRI International (“SRI”) in the 1990s. Martin, Cheyer,
`(both SRI employees) and a third SRI employee, Dr. Doug-
`las B. Moran, had earlier co-authored an academic paper
`entitled “Building Distributed Software Systems with the
`Open Agent Architecture” (“the Martin reference”) that
`
`

`

`Case: 21-1179 Document: 48 Page: 3 Filed: 05/19/2022
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`3
`
`was published in the Proceedings of the Third Interna-
`tional Conference on the Practical Application of Intelli-
`gent Agents and Multi-Agent Technology, which took place
`March 23–25, 1998. The Martin reference describes the
`OAA project developed at SRI and, significantly for present
`purposes, at least some of the technology embodied in the
`claims of the ’115 and ’560 patents.
`During the prosecution of the ’115 patent, various
`claims were rejected based on the Martin reference, which
`the examiner identified as being prior art. In response, SRI
`contested the prior art status of the reference by submit-
`ting inventor declarations by Martin and Cheyer under 37
`C.F.R. § 1.132, asserting that Dr. Moran was “not a co-in-
`ventor of the subject matter described in the subject matter
`disclosed and claimed in the instant application[s].”
`J.A. 11172–75. If Dr. Moran was not a co-inventor of the
`Martin reference, the Martin reference was not prior art
`because it was made by the same inventive entity as the
`’115 and ’560 patents and not “by others.” 35 U.S.C.
`§ 102(a) (pre-AIA).1 After receiving the declarations, the
`examiner withdrew the rejections based on the Martin ref-
`erence and continued examining the applications. The pa-
`tents were granted and ultimately assigned to appellee IPA
`Technologies, Inc. (“IPA”).
`In February 2019, Google petitioned the Board for inter
`partes review of various claims of the ’115 and ’560 patents,
`relying primarily on the Martin reference to argue that the
`claims would have been obvious. Google contended that
`the Martin reference was prior art as work “by others” be-
`cause it described the work of an inventive entity (Martin,
`Cheyer and Dr. Moran) different from the inventive entity
`of the challenged patents (Martin and Cheyer). The Board
`
`1 Because the patents at issue were filed before
`March 16, 2013, pre-AIA provisions apply. 35 U.S.C. § 100
`(note).
`
`

`

`Case: 21-1179 Document: 48 Page: 4 Filed: 05/19/2022
`
`4
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`instituted review but concluded after the trial proceedings
`that Google “ha[d] not provided sufficient support to ex-
`plain how Dr. Moran’s contribution [wa]s sufficient to es-
`tablish he [wa]s an inventive entity with respect to the
`Martin reference by a preponderance of the evidence” and
`that Google thus failed to “establish[] that Martin was
`prior art under § 102(a) to the ’560 Patent.” J.A. 26.2 Be-
`cause each of Google’s grounds in its petition relied on the
`Martin reference, the Board concluded that Google did not
`establish that any of the challenged claims was unpatent-
`able. Google appeals. We have jurisdiction under 28
`U.S.C. § 1295(a)(4)(A).
`
`DISCUSSION
`I
`We review the Board’s legal determination de novo and
`any underlying factual findings for substantial evidence.
`Duncan Parking Techs., Inc. v. IPS Grp., Inc., 914 F.3d
`1347, 1357 (Fed. Cir. 2019). “[W]hether a reference is a
`work of others for the purposes of § 102(a) is, like that of
`inventorship, a question of law based on underlying facts.”
`Allergan, Inc. v. Apotex Inc., 754 F.3d 952, 969 (Fed. Cir.
`2014) (citing Ethicon, Inc. v. U.S. Surgical Corp., 135 F.3d
`1456, 1460 (Fed. Cir. 1998)).
`Joint inventors need “not physically work together or
`at the same time, . . . make the same type or amount of con-
`tribution, or . . . make a contribution to the subject matter
`of every claim of the patent.” 35 U.S.C. § 116(a). A joint
`inventor must simply:
`
`
`2 Because Google’s arguments regarding the prior
`art status of the Martin reference were the same across the
`three IPR proceedings and that is the sole issue on appeal,
`unless otherwise noted, we cite only the materials corre-
`sponding to IPR2019-728 challenging the ’560 patent.
`
`

`

`Case: 21-1179 Document: 48 Page: 5 Filed: 05/19/2022
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`5
`
`(1) contribute in some significant manner to the
`conception or reduction to practice of the invention,
`(2) make a contribution to the claimed invention
`that is not insignificant in quality, when that con-
`tribution is measured against the dimension of the
`full invention, and (3) do more than merely explain
`to the real inventors well-known concepts and/or
`the current state of the art.
`In re VerHoef, 888 F.3d 1362, 1366 (Fed. Cir. 2018) (quoting
`Pannu v. Iolab Corp., 155 F.3d 1344, 1351 (Fed. Cir. 1998)).
`Accordingly, “to decide whether a reference patent is ‘by
`another’ . . . , the Board must”:
`(1) determine what portions of the reference patent
`were relied on as prior art to anticipate the claim
`limitations at issue, (2) evaluate the degree to
`which those portions were conceived ‘by another,’
`and (3) decide whether that other person’s contri-
`bution is significant enough, when measured
`against the full anticipating disclosure, to render
`him a joint inventor of the applied portions of the
`reference patent.
`Duncan Parking, 914 F.3d at 1358 (quoting pre-AIA
`35 U.S.C. § 102(e)).
` Google argues that the Board improperly imposed a
`burden on Google to prove that the Martin reference has a
`different inventive entity than the challenged patents. In
`response, IPA argues that the Board correctly placed the
`burden on Google to show what Dr. Moran contributed to
`the Martin reference.
`
`The term “burden of proof” has been used to describe
`two distinct concepts: the burden of persuasion and the
`burden of production. The burden of persuasion is “the ul-
`timate burden assigned to a party who must prove some-
`thing to a specified degree of certainty.” Tech. Licensing
`Corp. v. Videotek, Inc., 545 F.3d 1316, 1326 (Fed. Cir.
`
`

`

`Case: 21-1179 Document: 48 Page: 6 Filed: 05/19/2022
`
`6
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`2008). In an IPR, “the burden of persuasion is on the peti-
`tioner to prove ‘unpatentability by a preponderance of the
`evidence,’ and that burden never shifts to the patentee.”
`Dynamic Drinkware, LLC v. Nat’l Graphics, Inc., 800 F.3d
`1375, 1378 (Fed. Cir. 2015) (quoting 35 U.S.C. § 316(e)). In
`contrast, the burden of production, or “going forward with
`evidence,” is a shifting one, “the allocation of which de-
`pends on where in the process of trial the issue arises.” Id.
`at 1379 (quoting Tech. Licensing, 545 F.3d at 1327). The
`burden of production may be met either by “producing ad-
`ditional evidence” or by “presenting persuasive argument
`based on new evidence or evidence already of record.” Tech.
`Licensing, 545 F.3d at 1327.
` With respect to the burden of production, the record
`contains evidence and arguments from Google and IPA in
`support of their respective positions on the prior art status
`of the Martin reference. Google, as the petitioner, had the
`ultimate burden of persuasion to prove unpatentability by
`a preponderance of the evidence and this burden never
`shifted. We see no error with the Board’s requiring that
`Google establish the Martin reference was prior art “by an-
`other” by showing that Dr. Moran made a significant
`enough contribution to the portions relied on to invalidate
`the challenged patents to qualify as a joint inventor of
`those portions. We next address whether Google satisfied
`its burden.
`
`II
` Dr. Moran was the most senior computer scientist on
`the OAA team, worked on the OAA team for five years, au-
`thored five papers relating to OAA, and was named as a
`joint inventor on U.S. Patent No. 6,859,931 (“the ’931 pa-
`tent”), which is a continuation-in-part of the application
`that resulted in the ’115 patent. The ’931 patent claims
`appear to be directed at the same computer architecture as
`the ’115 and ’560 patents with an additional “bridge agent”
`capable of communicating with and incorporating other
`
`

`

`Case: 21-1179 Document: 48 Page: 7 Filed: 05/19/2022
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`7
`
`distributed component systems. E.g., ’931 patent, col. 33,
`l. 4–col. 34, l. 10. Although Dr. Moran did “little program-
`ming for OAA after 1995” due to a nerve injury, he claims
`that he “conducted code reviews and design sessions” on
`top of his primary responsibility for “presentations, demon-
`stration scenarios, funding proposals and research publica-
`tions.” J.A. 3665. According to Cheyer, Dr. Moran “helped
`solve specific problems, mostly at the application level” in
`the OAA project. J.A. 8790. The Board found that Dr. Mo-
`ran contributed “supporting foundational concepts” to the
`OAA project. J.A. 17. It is evident that Dr. Moran made
`general technical contributions to the OAA project. But
`that is not the relevant inquiry. To be a joint inventor of
`the Martin reference, as recognized in Duncan Parking, he
`must have made an inventive contribution to the portions
`of the reference relied on and relevant to establishing obvi-
`ousness. See 914 F.3d at 1358.3
`Here, Google claims that the description in the Martin
`reference of using a facilitator is, in part, grounds for find-
`ing that the Martin reference rendered the claims obvious,
`and that Dr. Moran made an inventive contribution to the
`facilitator concept recited in the Martin reference. Specifi-
`cally, Dr. Moran claimed to “play[] a significant role regard-
`ing the distributed agent-based approach, and in particular
`using a facilitator, as described in the [Martin reference] at
`
`
`3 The portions of the reference being considered
`must be relied upon and relevant to establishing obvious-
`ness. Otherwise, a party challenging a patent could artifi-
`cially alter the inventive entity for comparison by citing
`extraneous portions of a multi-inventor prior art reference,
`thereby making it “by others” even if the portions of the
`reference necessary for establishing obviousness had the
`same inventive entity as the challenged patent.
`
`

`

`Case: 21-1179 Document: 48 Page: 8 Filed: 05/19/2022
`
`8
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`Section 4 and 4.1–4.5.” J.A. 3692.4 Section 4.5 of the Mar-
`tin reference states:
`Facilitation plays a central role in OAA. At its core,
`our notion of facilitation is similar to that proposed
`by Genesereth [] and others. In short, a facilitator
`maintains a knowledge base that records the capa-
`bilities of a collection of agents, and uses that
`knowledge to assist requesters and providers of
`services in making contact. But our notion of facil-
`itation is also considerably stronger in three re-
`spects.
`J.A. 3954 (internal citation omitted). The reference then
`identifies three ways in which its facilitator differs from
`the prior art: “transparent delegation,” “handling of com-
`pound goals” via delegation, optimization, and interpreta-
`tion, and using “strategies and advice given by the
`requesting agent.” J.A. 3954–55. These concepts found
`
`
`4 Dr. Moran also claimed to “play[] a significant role
`in the approach of using recursion to decompose base goals
`into subgoals that were then dispatched to agents, e.g., as
`described in the [Martin reference] at Sections 2.5 and 4.1–
`4.2.” J.A. 3692. For our purposes, it is unnecessary to ad-
`dress this claim because a joint inventor need not contrib-
`ute to every aspect of a prior art reference, but rather need
`only make a “contribution [] significant enough, when
`measured against the full anticipating disclosure, to render
`him a joint inventor of the applied portions of the [prior art]
`reference.” Duncan Parking, 914 F.3d at 1358. If, on re-
`mand, joint inventorship is not found as to the relied-on
`portions of the Martin reference for the facilitator claim
`limitation, the Board may have to address the recursion
`portions of the reference, including Dr. Moran’s role in
`those portions and what bearing those portions have on the
`challenged patent claims.
`
`

`

`Case: 21-1179 Document: 48 Page: 9 Filed: 05/19/2022
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`9
`
`their way into the specification of the ’560 patent. See ’560
`patent, col. 19, ll. 14–62.
`When asked about Dr. Moran’s contributions to OAA,
`Cheyer acknowledged Dr. Moran’s assistance with tech-
`nical problems “at the application level” but stated that Dr.
`Moran did not “influenc[e] core OAA architecture or struc-
`ture.” J.A. 8790–91. Cheyer describes Dr. Moran’s state-
`ments about “having multiple facilitator architectures and
`a recursive structure” as a “mischaracterization of his role”
`because those concepts “clearly existed before he ever be-
`came involved with the project,” dating back to “the very
`first OAA paper.” J.A. 8819–22.
`Martin testified that he and Cheyer were responsible
`for the technical details in the Martin reference, and that
`Dr. Moran’s role on OAA was “administrative.” J.A. 8889.
`When presented with other articles relating to technical as-
`pects of OAA listing Dr. Moran as an author, Martin
`acknowledged that Dr. Moran “may have made contribu-
`tions to the technological development of either interfaces
`for use with OAA, agents used with OAA, or systems based
`on OAA” but with the caveat that these were topics that
`were “not necessarily part of OAA per se.” J.A. 8907.
`
`The testimony of Dr. Moran, if credited, might well es-
`tablish that he was a co-inventor of the particular portions
`of the Martin reference relied on by Google in, and relevant
`to, the challenge to particular claims. However, the Board
`did not complete the full Duncan analysis. Instead, it ap-
`pears to have concluded that Dr. Moran’s testimony was
`insufficiently corroborated, “agree[ing] with Patent Owner
`that Dr. Moran’s [claimed contributions to the Martin ref-
`erence were] . . . not supported by additional evidence at
`trial,” and that the “record lack[ed] sufficient supporting
`evidence to establish the contributions of Dr. Moran and
`Messrs. Cheyer and Martin, beyond the Rule 1.132 decla-
`rations.” J.A. 22–23.
`
`

`

`Case: 21-1179 Document: 48 Page: 10 Filed: 05/19/2022
`
`10
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`Both parties appear to agree that the Board held Dr.
`Moran’s testimony was insufficiently corroborated. Con-
`trary to the Board’s decision, the record reveals more than
`adequate corroboration of Dr. Moran’s testimony. While
`the majority of corroboration cases involve issued patents,
`our cases have also required corroboration of testimony
`that an individual is an inventor of a potentially invalidat-
`ing prior art reference that is not a patent. Allergan, 754
`F.3d at 969. While “corroborating an inventor’s testimony
`is a well-established principle in our case law” for purposes
`of determining inventorship, it is not the case that “an in-
`ventor must produce contemporaneous documentary evi-
`dence . . . to support his or her declaration” or that a “high
`degree of corroboration . . . is required across the board.”
`EmeraChem Holdings, LLC v. Volkswagen Grp. of Am.,
`Inc., 859 F.3d 1341, 1346–47 (Fed. Cir. 2017). Although co-
`authorship does not presumptively make a co-author a co-
`inventor, e.g., In re Katz, 687 F.2d 450, 455 (CCPA 1982);
`Allergan, 754 F.3d at 969, it is significant corroborating ev-
`idence that a co-author contributed to the invention. Here,
`Dr. Moran’s testimony as to his technical contributions was
`sufficiently corroborated by his being named as a co-author
`on the Martin reference, his role within the OAA project,
`Cheyer’s acknowledgement of Dr. Moran’s technical contri-
`butions to the OAA project, and Dr. Moran’s being named
`on the related ’931 patent.
`The issue in this case was not lack of corroboration for
`Dr. Moran’s testimony, but rather whether his testimony
`should ultimately be credited over Cheyer and Martin’s
`conflicting testimony during the IPR proceedings.5 Instead
`of resolving the conflicts, the Board stated that it found
`
`
`5 The § 1.132 declarations by Martin and Cheyer,
`standing alone, were insufficiently focused on the Duncan
`question to raise a fact issue. See EmeraChem, 859 F.3d at
`1345–46.
`
`

`

`Case: 21-1179 Document: 48 Page: 11 Filed: 05/19/2022
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`11
`
`“the testimony of Dr. Moran, Mr. Martin, and Mr. Cheyer
`credible with respect to the facts cited herein.” J.A. 17
`n.10; see also IPA Br. at 37 (“The Board found all the wit-
`nesses credible . . . .”); id. at 66 (The Board found that “all
`the witnesses, including [Dr.] Moran, were ‘credible.’”).
`This was not a tenable position for the Board to take. The
`Board was required to resolve this highly relevant eviden-
`tiary conflict and make appropriate findings of fact.
`III
`IPA argues that we can nonetheless affirm because it
`is inconsistent for Google to claim “that [Dr.] Moran alleg-
`edly contributed invalidating subject matter to the Martin
`[reference], yet is correctly excluded as a named inventor
`of the Patents-at-Issue” because “[i]f the Martin [reference]
`discloses the key limitations of the inventions claimed in
`the Patents-at-Issue and [Dr.] Moran, in fact, contributed
`to the inventive subject matter in the Martin [reference],
`then, by extension, [Dr.] Moran would have also contrib-
`uted inventive subject matter to the Patents-at-Issue.” IPA
`Br. at 46. Of course, if Dr. Moran were a joint inventor on
`both the Martin reference and the patents-at-issue, then
`the Martin reference would no longer be prior art “by an-
`other.” But the named inventors on the patents, Cheyer
`and Martin, are presumptively “the true and only inven-
`tors.” Ethicon, 135 F.3d at 1460. IPA cannot raise this
`argument as a defense without actually seeking correction
`of inventorship of the patents, which it has not. See Pannu,
`155 F.3d at 1350 (“[A] patent with improper inventorship
`does not avoid invalidation simply because it might be cor-
`rected under section 256. Rather, the patentee must claim
`entitlement to relief under the statute and the court must
`give the patentee an opportunity to correct the inventor-
`ship.”); Horizon Meds. LLC v. Alkem Labs. Ltd., 2021 WL
`5315424, at *3 (Fed. Cir. 2021) (effort to correct
`
`

`

`Case: 21-1179 Document: 48 Page: 12 Filed: 05/19/2022
`
`12
`
`GOOGLE LLC v. IPA TECHNOLOGIES INC.
`
`inventorship of issued patent to avoid pre-AIA 35 U.S.C.
`§ 102(a) prior art).6
`
`CONCLUSION
`We vacate the Board’s decisions and remand for fur-
`ther proceedings consistent with this opinion.
`VACATED AND REMANDED
`COSTS
`
`No costs.
`
`
`6 The Manual of Patent Examining Procedure
`(“MPEP”) similarly advises applicants that a pre-AIA
`35 U.S.C. § 102(a) rejection can be overcome “by adding the
`coauthors as inventors to the application” assuming the
`statutory requirements are met. MPEP § 2132.01(I) (9th
`ed. Rev. 10, June 2020).
`
`