Case: 21-1412 Document: 53 Page: 1 Filed: 04/01/2022
NOTE: This disposition is nonprecedential.
United States Court of Appeals for the Federal Circuit ______________________
MICROSOFT CORPORATION, Appellant
v.
IPA TECHNOLOGIES INC., Cross-Appellant ______________________
2021-1412, 2021-1413, 2021-1414, 2021-1416, 2021-1417, 2021-1418, 2021-1419, 2021-1420, 2021-1421, 2021-1422, 2021-1423, 2021-1424, 2021-1440, 2021-1442 ______________________
Appeals from the United States Patent and Trademark Office, Patent Trial and Appeal Board in Nos. IPR2019- 00810, IPR2019-00811, IPR2019-00812, IPR2019-00813, IPR2019-00814, IPR2019-00835, IPR2019-00836, IPR2019-00837. ______________________
Decided: April 1, 2022 ______________________
JOSEPH A. MICALLEF, Sidley Austin LLP, Washington, DC, argued for appellant. Also represented by SCOTT BORDER; RICHARD ALAN CEDEROTH, Chicago, IL.
SARAH ELIZABETH SPIRES, Skiermont Derby LLP, Dal- las, TX, argued for cross-appellant. Also represented by Case: 21-1412 Document: 53 Page: 2 Filed: 04/01/2022
STEVEN WAYNE HARTSELL, JAIME OLIN, PAUL SKIERMONT; MIEKE K. MALMBERG, Los Angeles, CA. ______________________
Before DYK, SCHALL, and TARANTO, Circuit Judges. TARANTO, Circuit Judge. IPA Technologies Inc. owns U.S. Patent Nos. 6,851,115 and 7,069,560, which address computing systems with dis- tributed electronic agents. Microsoft Corporation chal- lenged various claims of the patents in eight inter partes reviews (IPRs) in the Patent and Trademark Office. The Office’s Patent Trial and Appeal Board held all challenged claims unpatentable for obviousness, except for claims 8– 10, 29–47, 63, and 86–89 of the ’115 patent and claims 10– 11, 28, 50–51, and 53–55 of the ’560 patent. Microsoft and IPA both appeal. As to Microsoft’s ap- peals: We affirm the Board’s determination of no proven unpatentability of claims 29–47 of the ’115 patent and claims 50–51 and 53–55 of the ’560 patent; but we reverse certain Board findings regarding claims 8–10, 63, and 86– 89 of the ’115 patent and claims 10–11 and 28 of the ’560 patent, and we remand for any further proceedings that may be necessary and appropriate to address those claims. As to IPA’s appeals (including cross-appeals): We affirm all of the Board’s determinations challenged here by IPA. I The ’115 patent is titled “Software-Based Architecture for Communication and Cooperation Among Distributed Electronic Agents,” and the (child) ’560 patent has a simi- lar title—“Highly Scalable Software-Based Architecture for Communication and Cooperation Among Distributed Electronic Agents.” The patents share a specification and have an effective filing date of January 5, 1999. Both de- scribe, in the words of the Abstract of the ’560 patent, a “software-based architecture . . . for supporting Case: 21-1412 Document: 53 Page: 3 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 3
cooperative task completion by flexible, dynamic configura- tions of autonomous electronic agents.” ’560 patent, Ab- stract; see ’115 patent, Abstract (similar). Figure 4 depicts the structure of an exemplary system, which is organized around a “facilitator agent” (equiva- lently here, a “facilitator”)—a specialized server agent that coordinates other specialized agents in the system, in part by keeping an agent registry of all the other agents’ capa- bilities. ’560 patent, Fig. 4; id., col. 6, lines 38–58; id., col. 7, lines 30–32, lines 51–53; see also id., Fig. 7. Cooperative task completion can be achieved in the exemplary system as follows. After receiving a request for service from an agent in the system, also called a “goal,” the facilitator (1) parses and interprets the goal, (2) creates a goal satis- faction plan by dividing the goal into sub-goals, and (3) del- egates the task of performing each sub-goal to an agent having the specialized capability to do so. See id., col. 18, line 54, through col. 19, line 13; see also id., Fig. 11. Com- munication between agents within this system is prefera- bly accomplished via a common language, which the patents call the Interagent Communication Language (ICL). Id., col. 10, line 54, through col. 11, line 7. On March 19, 2019, Microsoft filed eight IPR peti- tions—five involving the ’115 patent and three involving the ’560 patent. Microsoft argued that the subject matter of all the challenged claims would have been obvious over combinations including at least the following references: (1) Kiss (U.S. Patent No. 6,484,155) and (2) FIPA97 (a doc- ument created by the Foundation for Intelligent Physical Agents, J.A. 7519–832). Kiss, titled “Knowledge Management System for Per- forming Dynamic Distributed Problem Solving,” describes a “knowledge management system that supports inquiries of distributed knowledge resources,” in which “[i]nteraction between a user and the knowledge resources is mediated by a collection of cooperative intelligent agents.” Kiss, Case: 21-1412 Document: 53 Page: 4 Filed: 04/01/2022
Abstract. Kiss’s agents include meta agents, user agents, and knowledge agents (or intelligent agents), which are or- ganized into layers. Id., Fig. 1; id., col. 5, line 1, through col. 7, line 19. A registry, in the agent service layer, lists the capabilities of each agent that is active in the system. Id., col. 6, line 66, through col. 7, line 19. After a meta agent receives a “formulated question” from a user agent, it “analyzes the formulated question to formulate a solu- tion plan.” Id., col. 10, lines 12–15. “By interacting with the agent service layer,” including the registry, the meta agent “allocates and assigns tasks to knowledge agents . . . based on the interests and capabilities of those knowledge agents.” Id., col. 10, lines 15–18; id., col. 6, line 66, through col. 7, line 2. Undisputedly, however, Kiss does not disclose an interagent communication language. Microsoft Open- ing Br. 12–13. FIPA97, created to “provide[] specification of basic agent technologies that can be integrated by agent systems developers to make complex systems with a high degree of interoperability,” is split into seven parts. J.A. 7523–25. Part 1 describes a framework for agent management, de- fined by Agent Platforms and Agent Domains. J.A. 7524; J.A. 7536 (Fig. 2). Each Agent Domain has a Directory Fa- cilitator that maintains an agent registry for agents within the domain. J.A. 7532–33. Part 2 describes an Agent Com- munication Language (ACL), a common communication protocol that allows agents to communicate with one an- other. J.A. 7524; J.A. 7575. Part 3 discusses agent-soft- ware integration, and Parts 4–7 provide exemplary applications of the system (e.g., for personal travel assis- tance or for audio/video entertainment and broadcasting). J.A. 7524–25. The Board instituted all IPRs and ultimately issued eight final written decisions that declared all challenged claims unpatentable for obviousness, except for claims 8– 10, 29–47, 63, and 86–89 of the ’115 patent and claims 10– 11, 28, 50–51, and 53–55 of the ’560 patent. All eight final Case: 21-1412 Document: 53 Page: 5 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 5
written decisions are now before us on timely appeals (or cross-appeals) by Microsoft and IPA. We have jurisdiction under 28 U.S.C. § 1295(a)(4)(A) and 35 U.S.C. §§ 141(c), 319. Further details about the challenged patents and claims, prior art, and procedural history are discussed as relevant below. II Microsoft presents five arguments on appeal. We re- view the Board’s legal conclusions (including claim con- structions based solely on intrinsic evidence) de novo and its factual findings (including those concerning the prior art’s teachings and the motivation to combine) for substan- tial-evidence support. Hamilton Beach Brands, Inc. v. f’real Foods, LLC, 908 F.3d 1328, 1339 (Fed. Cir. 2018); Redline Detection, LLC v. Star Envirotech, Inc., 811 F.3d 435, 449 (Fed. Cir. 2015). Substantial evidence exists if “‘a reasonable fact finder could have arrived at the agency’s decision’” in light of the record as a whole. Intelligent Bio- Systems, Inc. v. Illumina Cambridge Ltd., 821 F.3d 1359, 1366 (Fed. Cir. 2016) (quoting In re Gartside, 203 F.3d 1305, 1312 (Fed. Cir. 2000)). We review the Board’s judg- ments concerning what arguments have been adequately presented in a petition and other pleadings for abuse of dis- cretion. See Ericsson Inc. v. Intellectual Ventures I LLC, 901 F.3d 1374, 1379 (Fed. Cir. 2018); Altaire Pharms., Inc. v. Paragon Bioteck, Inc., 889 F.3d 1274, 1284 (Fed. Cir. 2018), remand order modified by stipulation, 738 F. App’x 1017 (Fed. Cir. 2018); Intelligent Bio-Systems, 821 F.3d at 1367; see also MModal LLC v. Nuance Commc’ns, Inc., 846 F. App’x 900, 906 (Fed. Cir. 2021). “An abuse of discretion is found if the decision: (1) is clearly unreasonable, arbi- trary, or fanciful; (2) is based on an erroneous conclusion of law; (3) rests on clearly erroneous fact finding; or (4) in- volves a record that contains no evidence on which the Board could rationally base its decision.” Intelligent Bio- Systems, 821 F.3d at 1367 (citation omitted). Case: 21-1412 Document: 53 Page: 6 Filed: 04/01/2022
A Microsoft’s first argument on appeal is that, for claims 8–10 and 63 of the ’115 patent and for claim 28 of the ’560 patent, the Board erred in finding that Microsoft had not proven how or why Kiss and FIPA97 would have been com- bined by a relevant artisan. Microsoft Opening Br. 49–57. Claims 8 and 9 of the ’115 patent, which are representative for purposes of this issue, read: 8. A computer-implemented method as recited in claim 5 wherein the agent registry data struc- ture includes at least one trigger declaration for one active agent. 9. A computer-implemented method as recited in claim 5 wherein the agent registry data struc- ture includes at least one task declaration, and pro- cess characteristics for each active agent. ’115 patent, col. 30, lines 10–16 (emphases added). For con- text, the undisputed interpretation of “trigger” adopted by the Board is “a general mechanism for requesting some ac- tion be taken when one or more conditions is met.” Mi- crosoft Corp. v. IPA Techs. Inc., No. IPR2019-00810, 2020 WL 6110974, at *50 (P.T.A.B. Oct. 15, 2020) (IPR810 Deci- sion). And the specification describes a “trigger declara- tion” as follows: “Each registered agent may be optionally associated with one or more triggers, which preferably could be referenced through their associated Trigger Dec- laration fields . . . in the parent facilitator Agent Regis- try . . . .” ’115 patent, col. 17, lines 5–9. “Process characteristics,” as construed by the Board, are “character- istics of a process for an agent.” IPR810 Decision, 2020 WL 6110974, at *11; see also ’115 patent, col. 17, lines 12–21 (“Each registered agent may be optionally associated with one or more Process Characteristics, which preferably could be referenced through their associated Process Char- acteristics Declaration fields . . . in the parent facilitator Agent Registry . . . .”). Case: 21-1412 Document: 53 Page: 7 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 7
In its IPR2019-00810 petition, which included chal- lenges to claims 8–10 of the ’115 patent, Microsoft relied on a combination of Kiss and FIPA97 as disclosing the claimed agent registry, see, e.g., J.A. 1428–30, and Kiss as disclos- ing the agent registry data structure, see, e.g., IPR810 De- cision, 2020 WL 6110974, at *59. For trigger declarations, Microsoft alleged that FIPA97 Part 6, through its descrip- tion of the “preconditions2” trigger condition, disclosed an agent registry with trigger declarations. J.A. 1455–56. For process characteristics, Microsoft alleged that FIPA97 Part 1, through its description of an “interaction-protocols” pa- rameter, disclosed an agent registry with process charac- teristics. J.A. 1457. Regarding how and why the relevant artisan would combine Kiss and FIPA97 to yield agent registry data structures with trigger declarations or process characteris- tics, Microsoft, in the general art-description section of its petition, explained: In [the Kiss and FIPA97] combination, FIPA97 dis- closes the inter-agent communication language re- quirements of the claims not expressly disclosed in Kiss, as well as other administrative functionality (i.e., Parts 1–3) and exemplary practices (i.e., Parts 4–7). In particular, FIPA97 provides a common communications protocol and language (FIPA ACL) between the agents of Kiss, and also adds its administrative functionality and exemplary prac- tices to the Kiss system, as described with specificity below. These techniques are used to implement the functionality described in Kiss, including facilitat- ing agent collaboration, agent registry, and inter- agent messaging, as well [as] adding functionality that is disclosed in FIPA97. Where the two systems disclose analogous functionality, such as facilitat- ing cooperation and agent registry, their techniques are combined, as a Skilled Artisan would under- stand that to be an efficient and common sense way Case: 21-1412 Document: 53 Page: 8 Filed: 04/01/2022
to implement the combined system in order to ob- tain the benefits of both, and therefore be moti- vated to do so. J.A. 1423 (emphases added and citations omitted). Mi- crosoft also stated that the relevant artisan would include FIPA97’s administrative functionality and exemplary practices to obtain the most advantageous implementation of FIPA97’s common communication protocol, ACL. J.A. 1426 (also discussing standardization motivation). Mi- crosoft’s expert, Dr. Lieberman, made similar statements. See J.A. 7227–42 (Lieberman Decl. ¶¶ 217–31). In response, IPA contended that there was no explana- tion of why a relevant artisan would be motivated to incor- porate FIPA97’s agent registry’s trigger declarations and process characteristics into Kiss’s agent registry data structure. J.A. 1865–67. Microsoft replied by pointing to its petition’s discussion of why a relevant artisan would want to add the administrative functionality and exem- plary practices of FIPA97. J.A. 2029–34. In its final decision in IPR2019-00810 for claims 8–10 of the ’115 patent, the Board found that Microsoft did not prove “how or why Kiss’s agent registry data structure would have been combined with FIPA97’s” analogous agent registry to obtain the subject matter of the claims. IPR810 Decision, 2020 WL 6110974, at *61 (claims 9–10); id. at *62 (claim 8). The Board recognized that “above, we discussed [Microsoft’s] proposed combination of Kiss’s electronic agents and their general functions and operations, includ- ing agent collaboration, agent registry, and inter-agent messaging, with FIPA97’s ACL,” i.e., the combination of Kiss’s registry with FIPA97’s communication protocol, but the Board stated that, for the present claims, Microsoft “is proposing a different (or additional) combination—that is, combining the registry of Kiss with the registry of FIPA97.” Id. at *61 (claims 9–10) (emphasis added); see also id. at *62 (claim 8). The Board addressed claim 63 of the ’115 Case: 21-1412 Document: 53 Page: 9 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 9
patent and claim 28 of the ’560 patent similarly in IPR2019-00814 and IPR2019-00835. Microsoft Corp. v. IPA Techs., Inc., No. IPR2019-00814, 2020 WL 6532192, *57–59 (P.T.A.B. Nov. 5, 2020) (IPR814 Decision); Mi- crosoft Corp. v. IPA Techs., Inc., No. IPR2019-00835, 2020 WL 6106141, *46–47 (P.T.A.B. Oct. 15, 2020) (IPR835 De- cision). The Board’s determinations at issue are not supported by substantial evidence. Concerning why the pertinent as- pects of the registries of Kiss and FIPA97 would be com- bined, Microsoft presented significant evidence of motivation, see J.A. 7227–42 (Lieberman Decl. ¶¶ 217– 31)—which the Board credited elsewhere in its decisions, see IPR810 Decision, 2020 WL 6110974, at *39–40. The Board stated that this evidence did not explain why Kiss’s registry and FIPA97’s registry (with its trigger declarations and/or process characteristics) would be combined, as op- posed to why Kiss’s registry and FIPA97’s ACL would be combined, but that determination is unreasonable on the evidence. The Lieberman declaration did account for why relevant artisans starting with a Kiss-like system would look to FIPA97 and borrow not just its ACL, but also its other administrative functionalities and exemplary prac- tices. See, e.g., J.A. 7240–41 (Lieberman Decl. ¶¶ 229–31). Concerning how Kiss and FIPA97’s registries would be combined, Microsoft presented evidence, with little elabo- ration for what Microsoft said a skilled artisan would know how to do. See, e.g., J.A. 1423–26; J.A. 7227–29, 7240 (Lieberman Decl. ¶¶ 217–19, 228). IPA’s response, which focused almost entirely on the why question, was at most a simple, unelaborated assertion that Microsoft’s evidence as to how was not enough, with no identification at all of what necessary details were missing. See, e.g., J.A. 1865–67 (ob- jecting to why, not how, the combination would be made); J.A. 4846–48 (making single-sentence assertion about the how issue); J.A. 12861 (Medvidovic Decl. ¶ 261) (same). The Board, for its part, did not meaningfully address the Case: 21-1412 Document: 53 Page: 10 Filed: 04/01/2022
how issue separately from the why issue. It did not identify any concrete unanswered questions about how to achieve the combination (which is not a matter of bodily incorpora- tion, but of combining the features required to arrive at the claim). It did not even say whether the combination re- quired more than simply adding one more field to Kiss’s registry data structure. In these circumstances, we see no basis for a reasona- ble determination rejecting Microsoft’s evidence of why and how the combination at issue would be made. Accordingly, we reverse the Board’s finding that Microsoft has not per- suasively shown how or why Kiss and FIPA97 would have been combined to generate the subject matter of claims 8– 10 and 63 of the ’115 patent and claim 28 of the ’560 patent. We remand for the Board to address any remaining obvi- ousness disputes concerning these claims. See J.A. 1865– 67 (IPA in its Patent Owner Response making non-disclo- sure arguments beyond lack of motivation for claims 8–10 of the ’115 patent); J.A. 1867 (same for claim 63 of the ’115 patent); J.A. 4846–48 (same for claim 28 of the ’560 patent); see also J.A. 5101–03 (IPA’s sur-reply on claim 28 of the ’560 patent). B Microsoft’s second argument on appeal is that, for claims 29–47 of the ’115 patent, the Board erred in finding the petition insufficient. Microsoft Opening Br. 57–64. Claim 29, which is representative for this issue, is repro- duced below, as is claim 1, which is relevant to our analy- sis. 1. A computer-implemented method for com- munication and cooperative task completion among a plurality of distributed electronic agents, comprising the acts of: . . . constructing a goal satisfaction plan wherein the goal satisfaction plan includes: Case: 21-1412 Document: 53 Page: 11 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 11
a suitable delegation of sub-goal requests to best complete the requested service request—by using reasoning that includes one or more of do- main-independent coordination strategies, do- main-specific reasoning, and application-specific reasoning comprising rules and learning algo- rithms; and dispatching each of the sub-goals to a selected client agent for performance, based on a match be- tween the sub-goal being dispatched and the regis- tered functional capabilities of the selected client agent. ’115 patent, col. 29, lines 10–43. 29. A computer program stored on a computer readable medium, the computer program executa- ble to facilitate cooperative task completion within a distributed computing environment, the distrib- uted computing environment including a plurality of autonomous electronic agents, the distributed computing environment supporting an Interagent Communication Language, the computer program comprising computer executable instructions for: . . . constructing a base goal satisfaction plan in- cluding the sub-acts of: determining whether the request service is available, determining sub-goals required in completing the base goal by using reasoning that includes one or more of domain-independent coordination strat- egies, domain-specific reasoning, and application- specific reasoning comprising rules and learning algorithms, Case: 21-1412 Document: 53 Page: 12 Filed: 04/01/2022
selecting service-providing electronic agents from the agent registry suitable for performing the determined sub-goals, and ordering a delegation of sub-goal requests com- plete the requested service . . . . Id., col. 31, line 48, through col. 32, line 22. In its IPR2019-00810 petition, for the claim 29 limita- tions concerning “constructing a [base] goal satisfaction plan,” Microsoft stated only: “Kiss/FIPA97 satisfies this claim for the reasons stated in §§ VI.A.1.h.” J.A. 1464 (re- ferring to the petition’s discussion of the claim 1 limitation “constructing a goal satisfaction plan,” J.A. 1445–49). In its institution decision, the Board found that Microsoft had not demonstrated a reasonable likelihood of prevailing, ob- serving that “there exist material differences between these limitations on the face of the claims” but that Mi- crosoft “does not address the claim language of claim 29 or explain why its analysis for claim 1 is sufficient to demon- strate a reasonable likelihood of prevailing with respect to claim 29.” J.A. 1693–95. IPA echoed this view in its patent owner’s response, J.A. 1845–47, and, in reply, Microsoft ar- gued that there were no material differences between claims 1 and 29, explaining for each alleged difference why the respective claim language had the same meaning and how the alleged disclosure for claim 1 equally applied to claim 29, J.A. 2019–27. In its final decision in IPR2019-00810, the Board deter- mined that Microsoft had failed to meet its burden in the petition to show with particularity the evidence that sup- ported the grounds for the challenge to claim 29. IPR810 Decision, 2020 WL 6110974, at *64–66. Additionally, the Board declined to consider Microsoft’s arguments “pre- sented belatedly in the Petitioner Reply.” Id. at *66. The Board found dependent claims 30–32 and 38–39 unpatent- able because they “depend directly or indirectly from claim 29” and Microsoft’s “arguments and evidence presented . . . Case: 21-1412 Document: 53 Page: 13 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 13
only address the additionally recited limitations . . . and, therefore, do not remedy the deficiencies in [Microsoft’s] analysis of independent claim 29.” Id. at *66, *70. The Board addressed claim 29’s other challenged dependent claims similarly in IPR2019-00811 and IPR2019-00812. Microsoft Corp. v. IPA Techs. Inc., No. IPR2019-00811, 2020 WL 6112246, at *56–58, *62 (P.T.A.B. Oct. 15, 2020) (IPR811 Decision); Microsoft Corp. v. IPA Techs. Inc., No. IPR2019-00812, 2020 WL 6106303, at *58–61, *66, *74 (P.T.A.B. Oct. 15, 2020). We affirm the Board’s determination that Microsoft in- sufficiently pleaded the unpatentability of claims 29–47. Microsoft’s petition did not grapple with the claim 29 lan- guage, simply stating “Kiss/FIPA97 satisfies this claim for the reasons stated in” the section pertaining to the claim 1 limitation “constructing a goal satisfaction plan.” J.A. 1464. That statement is true, however, only if there are no differences between the claims that might affect the Kiss and FIPA97 combination’s disclosure. Yet that central, necessary premise is one that Microsoft wholly failed to de- fend in its petition, even though there are potentially meaningful differences in claim language that warrant dis- cussion. 1 The Board did not abuse its discretion in
1 Compare ’115 patent, col. 29, lines 34–39 (claim 1 requiring “suitable delegation of subgoal requests to best compete the requested service request” by using certain reasoning techniques), with id., col. 32, lines 11–16 (claim 29 requiring “determining whether the request service is available” and “determining sub-goals required in complet- ing the base goal” by using certain reasoning techniques). Compare also id., col. 29, lines 40–43 (claim 1 requiring “dispatching each of the sub-goals to a selected client agent for performance, based on a match between the sub-goal being dispatched and the registered functional capabilities of the selected client agent”), with id., col. 32, lines 17–21 Case: 21-1412 Document: 53 Page: 14 Filed: 04/01/2022
concluding that Microsoft’s complete failure to address the language differences and explain why they do not matter was a failure to meet the particularity requirements for what must be in the petition and cannot be left to be filled in on reply. See 35 U.S.C. § 312(a)(3); 37 C.F.R. §§ 42.22(a)(2) (requiring, among other things, “a detailed explanation of the significance of the evidence”), 42.104(b)(4)–(5) (requiring that petition “specify where each element of the claim is found in the prior art patents or printed publication relied upon” and, regarding submit- ted evidence, state “the relevance of the evidence to the challenge raised, including identifying specific portions of the evidence that support the challenge”); Intelligent Bio- Systems, 821 F.3d at 1369; Microsoft Corp. v. Biscotti, Inc., 878 F.3d 1052, 1074–75 (Fed. Cir. 2017); Harmonic Inc. v. Avid Tech., Inc., 815 F.3d 1356, 1363–64 (Fed. Cir. 2016). Microsoft effectively introduced a new issue when, on reply, it discussed claim differences for the first time. Cf. Cham- berlain Grp., Inc. v. One World Techs., Inc., 944 F.3d 919, 925 (Fed. Cir. 2019); see also AMC Multi-Cinema, Inc. v. Fall Line Patents, LLC, No. 2021-1051, 2021 WL 4470062, at *6 (Fed. Cir. Sept. 30, 2021) (collecting cases). Microsoft argues for a contrary conclusion by observing that two claims using different words can cover the same subject matter and that the same prior-art disclosures or combinations can satisfy two differently worded claims. See, e.g., Microsoft Opening Br. 59–60 (citing Tandon Corp. v. US Int’l Trade Comm’n, 831 F.2d 1017, 1023 (Fed. Cir. 1987)). But those observations do not cure the problem. Microsoft, in its petition, said nothing at all to defend the crucial premise that the different wording did not matter;
(claim 29 requiring “selecting service-providing electronic agents from the agent registry suitable for performing the determined sub-goals” and “ordering a delegation of sub- goal requests complete the requested service”). Case: 21-1412 Document: 53 Page: 15 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 15
and the reply came too late to make up for that complete omission from the petition. Accordingly, we affirm the Board’s determinations for claims 29–47 of the ’115 patent in IPR2019-00810, IPR2019-00811, and IPR2019-00812. C Microsoft’s third argument on appeal is that, for claims 86–89 of the ’115 patent, the Board erred in determining that the petition was insufficient. Microsoft Opening Br. 64–67. Claim 86, which is representative for this issue, re- cites: 86. A data wave carrier providing a transport mechanism for information communication in a distributed computing environment having at least one facilitator agent . . . wherein said at least one facilitator agent is op- erable to construct a goal satisfaction plan by using reasoning . . . . ’115 patent, col. 37, line 10, through col. 38, line 7 (empha- ses added). In its IPR2019-00811 petition, Microsoft said the fol- lowing about the claim 86 preamble: Kiss/FIPA97 discloses “information communica- tion in a distributed computing environment hav- ing at least one facilitator agent and at least one active client agent” for the reasons set forth above in §§ VI.[A.]1.a. EX1003, ¶¶ 200–205, 347–351. In particular, Kiss discloses various communications between agents. EX1005, 12:21–14:30, Figs. 8–20. In the combined system, moreover, the details con- cerning the interests and capabilities of that new agent are communicated to the agent service layer, which combined with the meta-agent of Kiss consti- tutes “a facilitator agent” because it facilitates the cooperation in resolving requests by bi-directionally Case: 21-1412 Document: 53 Page: 16 Filed: 04/01/2022
communicating with knowledge agents and assign- ing tasks (sub-goals) to them in order the coordinate the cooperative completion of tasks. EX1005, 12:1– 14:30, Figs. 8–20; EX1003, ¶¶ 200–205, 347–351. FIPA97 also discloses a Directory Facilitator (“fa- cilitator agent”), which provides analogous func- tionality, EX1006, 6–7, and for that reason would in the combined system be implemented in the meta agent. EX1003, ¶¶ 347–351. J.A. 2706–07 (original emphases omitted and emphases added). When later analyzing the claim 86 limitation “wherein said at least one facilitator agent is operable to construct a goal satisfaction plan,” Microsoft stated, “This claim element is satisfied for the same reasons described above in § VI.[A.]1.h.” J.A. 2708 (referring to the petition’s claim 1 discussion of how Kiss’s meta agent constructs a goal satisfaction plan, J.A. 2693–97). The Board determined in its final decision that Mi- crosoft had failed to show with particularity what evidence supported the grounds for the challenge to claim 86, and therefore its dependent claims 87–89. In particular, the Board found that Microsoft could not rely on the single sen- tence “[t]his claim element is satisfied for the same reasons described above in § VI.[A.]1.h,” cross-referencing the peti- tion’s discussion of claim 1, to explain the prior art’s teach- ing of the claim 86 limitation “wherein said at least one facilitator agent is operable to construct a goal satisfaction plan,” because claim 1 “does not recite ‘facilitator agent,’ nor does [Microsoft] explain in the context of its analysis of claim 1 how the combination of Kiss and FIPA97 teaches ‘at least one facilitator agent is operable to construct a goal satisfaction plan.’” IPR811 Decision, 2020 WL 6112246, at *58 (citation omitted). The Board’s determination on this point was an abuse of discretion. The Board’s reasoning neglects the petition’s discussion of the claim 86 preamble, which explained how Case: 21-1412 Document: 53 Page: 17 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 17
the combination of Kiss and FIPA97 disclosed a facilitator agent, constructed in part from Kiss’s meta agent. J.A. 2706–07. As a result, when addressing the more refined limitation “wherein said at least one facilitator agent is op- erable to construct a goal satisfaction plan,” Microsoft did not need to refer to claim 1’s discussion to disclose the fa- cilitator agent, having already explained how the combina- tion of Kiss and FIPA97 disclosed this feature. Instead, Microsoft used the cross-reference to explain how Kiss’s meta agent (part of the facilitator agent, according to the petition) had the capability of constructing a goal satisfac- tion plan. J.A. 2708. The Board’s failure to consider Mi- crosoft’s discussion of the claim 86 preamble, which expressly fills the hole the Board identified in the petition, and the Board’s resulting misapprehension of the purpose of the cross-reference, was error. Although IPA relies on Microsoft Corp. v. FG SRC, LLC, to argue otherwise, see IPA Opening & Response Br. 72–74, that case does not help IPA because it involved something quite different from this case—a Board refusal to “deduce the existence of an obvi- ousness argument from within [a petition’s] anticipation argument,” 860 F. App’x 708, 713 (Fed. Cir. 2021). Accordingly, we reverse the Board’s finding that Mi- crosoft’s petition did not sufficiently demonstrate with par- ticularity the evidentiary support for the grounds for the challenge to claims 86–89 of the ’115 patent, and we re- mand for the Board to reach the obviousness merits. It may be that obviousness of claims 86–89 follows from find- ings that (1) the combination of Kiss and FIPA97 teaches a facilitator and (2) Kiss’s meta agent constructs a goal sat- isfaction plan, J.A. 1818–47—two facts that either we or the Board have found, see, e.g., supra Section II.A; IPR814 Decision, 2020 WL 6532192, at *53–54; IPR835 Decision, 2020 WL 6106141, at *31–33. But we leave the merits de- termination to the Board. Case: 21-1412 Document: 53 Page: 18 Filed: 04/01/2022
D Microsoft’s fourth argument on appeal is that, for claims 10 and 11 of the ’560 patent, the Board erred in its implicit claim construction of “advisory suggestions.” Mi- crosoft Opening Br. 67–72. Claim 10, which is representa- tive for this issue, recites: 10. A computer architecture as recited in claim 8 wherein the ICL syntax supports explicit task completion advisory suggestions within goal ex- pressions. ’560 patent, col. 31, lines 10–12 (emphasis added). The specification, while not using the language of “advisory suggestions,” does discuss “advice parameters,” stating: Advice parameters preferably give constraints or guidance to the facilitator in completing and inter- preting the goal. For example, a solution_limit pa- rameter preferably allows the requester to say how many solutions it is interested in; the facilitator and/or service providers are free to use this infor- mation in optimizing their efforts. Similarly, a time_limit is preferably used to say how long the requester is willing to wait for solutions to its re- quest, and, in a multiple facilitator system, a level_limit may preferably be used to say how re- mote the facilitators may be that are consulted in the search for solutions. A priority parameter is preferably used to indicate that a request is more urgent than previous requests that have not yet been satisfied. Other preferred advice parameters include but are not limited to parameters used to tell the facilitator whether parallel satisfaction of the parts of a goal is appropriate, how to combine and filter results arriving from multiple solver agents, and whether the requester itself may be considered a candidate solver of the sub-goals of a request. Case: 21-1412 Document: 53 Page: 19 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 19
’560 patent, col. 16, lines 23–41. In its IPR2019-00836 petition, Microsoft did not pro- pose a claim construction for “advisory suggestions,” and it argued that this claim limitation was taught by FIPA97’s “Constraints+,” including the Director Facilitator Depth constraint, which Microsoft contended was similar to the “level_limit” parameter discussed in the specification. J.A. 5668–69; see also J.A. 8839–40 (Lieberman Decl. ¶¶ 430– 33). Microsoft thus equated “advisory suggestions” with advice parameters, inclusive of constraints. In response, IPA also did not propose a construction, but argued that FIPA97’s “Constraints+” did not disclose “advisory sugges- tions” because “advisory suggestions” are only those advice parameters which constitute guidance, not those which constitute constraints. J.A. 4844–46. In its final decision, the Board agreed with IPA, implic- itly adopting the construction that “advisory suggestions” are advice parameters that are guidance, exclusive of con- straints. The Board stated: Although the ’560 patent could more clearly differ- entiate “constraints” from “guidance” or “advisory suggestions,” the quoted passage from the ’560 pa- tent first describes some parameters that are quan- titative in nature (i.e., solution_limit, time_limit, and level_limit) and then describes some that are not (i.e., priority, whether parallel satisfaction is appropriate, how to combine and filter results). In the context of the quoted passage, solution_limit, time_limit, and level_limit, which require numeri- cal values, are more likely than the other parame- ters to be considered “constraints” rather than “guidance.” [Microsoft] asserts that FIPA97 teaches solution_limit, time_limit, and level_limit Case: 21-1412 Document: 53 Page: 20 Filed: 04/01/2022
parameters, 2 but [Microsoft] does not attempt to explain why these are disclosed examples of “advi- sory suggestions” rather than “constraints.” Under these circumstances, [Microsoft] has not met its burden to show that the prior art teaches the “ad- visory suggestions” limitation. Microsoft Corp. v. IPA Techs., Inc., No. IPR2019-00836, 2020 WL 6260642, *47 (P.T.A.B. Oct. 22, 2020) (IPR836 De- cision) (citations omitted). We reverse the Board’s implicit claim construction. IPA does not dispute that FIPA97’s “Constraints+” are con- straints as used in the ’560 specification. J.A. 4844–46; IPA Opening & Response Br. 74–76. Rather, the dispute is over whether the claim term “advisory suggestions” means advice parameters (and thus includes constraints) or, in- stead, means guidance only (and thus excludes con- straints). Compare Microsoft Opening Br. 67–72, with IPA Opening & Response Br. 74–76. In the present IPRs, we look for the correct claim construction under Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc), not for the broadest reasonable interpretation. Nothing in the Board’s decision or the parties’ briefs supply any significant basis for claim construction in this matter except for the specification passage quoted above.
2 Microsoft calls this clause “confused,” since Mi- crosoft never asserted that FIPA97 teaches the solu- tion_limit, time_limit, and level_limit parameters. Microsoft Opening Br. 69. But the likely Board intent was to say only that Microsoft asserts that FIPA97 teaches con- straints similar to the solution_limit, time_limit, and level_limit parameters. J.A. 5668–69; see also J.A. 8839– 40 (Lieberman Decl. ¶¶ 430–33). We need not pursue this matter further, because it does not affect the basis for our reversal. Case: 21-1412 Document: 53 Page: 21 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 21
See, e.g., IPR836 Decision, 2020 WL 6260642, at *46–47; IPA Opening & Response Br. 74–76. This is not a case in- volving a plain meaning of claim language that could chan- nel the consideration of the specification to an inquiry into disclaimer or redefinition. We look to the specification for how it resolves uncertainty left by the claim language when read in isolation. See, e.g., World Class Tech. Corp. v. Ormco Corp., 769 F.3d 1120, 1123–24 (Fed. Cir. 2014). The specification does not use the phrase “advisory suggestions” or even the term “suggestion,” and it therefore does not supply an explanation of the meaning of the phrase or term in this context. But the claim phrase “ad- visory suggestions” has an evident similarity to the speci- fication phrase “advice parameter,” and that similarity, in this context, suggests comparable scope. We see no suffi- cient basis for a different construction—specifically, for the narrowing to a “guidance” subset within the class of what the specification identifies as “advice parameters.” All the examples, whether expressly numerical or not, can easily be understood as “constraints,” and we see no sound basis for the Board’s apparently contrary view. Moreover, the specification does not make clear what, if any, distinction there is between “guidance” and “constraints,” and the Board did not define that distinction, on which its (and IPA’s) claim construction rests. In these circumstances, we conclude that the fairest meaning of “advisory suggestions” is “advice parameter,” not a meaning that excludes “con- straints.” Accordingly, we reverse the Board’s implicit claim con- struction, on which the Board rested its determination that Microsoft did not prove unpatentability of claims 10 and 11 of the ’560 patent. We remand to the Board to take the necessary next steps, which depend on what, if any, issues remain under the correct claim construction. See J.A. 4844–46. Case: 21-1412 Document: 53 Page: 22 Filed: 04/01/2022
E Microsoft’s fifth and final argument on appeal is that the Board erred in rejecting its challenge to claims 50–51 and 53–55 of the ’560 patent. Microsoft Opening Br. 72– 79. Claim 50, which is representative for purposes of this argument, recites: 50. A computer-implemented method for providing cooperative task completion within a dis- tributed computing environment, the distributed computing environment including a plurality of au- tonomous electronic agents, the distributed compu- ting environment supporting an Interagent Communication Language, the computer-imple- mented method comprising the acts of: providing a plurality of synchronized agent reg- istries each declaring capabilities of service-provid- ing electronic agents currently active within the distributed computing environment, the plurality of synchronized agent registries each resident within a separate computer process . . . . ’560 patent, col. 34, line 47, through col. 35, line 10 (empha- ses added). In its petition in IPR2019-00836, Microsoft offered a construction for the term “computer process,” suggesting that it means “a program, a part of a program, or a list of steps to be completed by the computer.” J.A. 5640; see also J.A. 8743 (Lieberman Decl. ¶ 232); J.A. 8300 (Microsoft Press Computer Dictionary). Applying this construction, Microsoft asserted that FIPA97’s Figure 2 taught several components of the “providing . . .” paragraph limitation of claim 50. J.A. 5684–88. FIPA97’s Figure 2 is reproduced below. Case: 21-1412 Document: 53 Page: 23 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 23
J.A. 7536. Microsoft’s petition did not assert or explain how, for the identified plurality of agent registries, “each” was “resident within a separate computer process.” See J.A. 5684–87; see also J.A. 8937 (Lieberman Decl. ¶ 697). IPA’s patent owner response did not call attention to this omission from the petition but focused on disputing Mi- crosoft’s construction of “computer process.” J.A. 4756–60. The Board adopted its own construction of “computer process” to mean “a process that runs or executes one or more computer programs.” IPR836 Decision, 2020 WL 6260642, at *9–10. Crucially, then, for a reason not de- pendent on resolution of the claim-construction dispute, the Board ruled that Microsoft had not shown that the re- lied-on prior art, specifically FIPA97’s Figure 2, taught an element of Claim 50. Id. at *50–52. Specifically, the Board determined: Even under [Microsoft’s] proposed construc- tion, . . . the Petition fails to address the language of claim 50, which recites “the plurality of agent reg- istries each resident within a separate computer process.” Without providing a detailed explanation Case: 21-1412 Document: 53 Page: 24 Filed: 04/01/2022
tying its argument to the claim language, [Mi- crosoft] has not met its burden of showing how the prior art teaches the recited limitation. Id. at *51 (citations omitted). We affirm that determination, and thus we do not reach the parties’ dispute concerning the proper construc- tion of “computer process.” We need not address whether Microsoft’s petition was adequate because it was readily apparent how FIPA97’s Figure 2 meets the claim require- ment that each of the plurality of agent registries is resi- dent within a separate computer process. This is so because Microsoft’s opening brief in this court provided no coherent explanation of why this part of the Board’s deci- sion was incorrect. See Microsoft Opening Br. 76–79. Its effort to do so with a new explanation in its reply brief was too late. See Microsoft Response & Reply Br. 17–18; Amhil Enters. Ltd. v. Wawa, Inc., 81 F.3d 1554, 1564 (Fed. Cir. 1996) (“A reply brief, which should ‘reply to the brief of the appellee,’ is not the appropriate place to raise, for the first time, an issue for appellate review.” (quoting Fed. R. App. P. 28(c))). For those reasons, we affirm the Board’s unpatentabil- ity determinations for claims 50–51 and 53–55 of the ’560 patent in IPR2019-00836. III We have considered IPA’s challenges to various Board determinations. We find them all unpersuasive. In partic- ular, we see no reversible error, based on the standards for corroboration or otherwise, in the Board’s determination that the version of FIPA97 in evidence was prior art. IV For the foregoing reasons, we affirm as to the Board determinations challenged by IPA. As to the determina- tions challenged by Microsoft, in the respects detailed in Case: 21-1412 Document: 53 Page: 25 Filed: 04/01/2022
MICROSOFT CORPORATION v. IPA TECHNOLOGIES INC. 25
this opinion, we (1) reverse in IPR2019-00810, IPR2019- 00814, and IPR2019-00835 regarding claims 8–10 and 63 of the ’115 patent and claim 28 of the ’560 patent; (2) affirm in IPR2019-00810, IPR2019-00811, and IPR2019-00812 re- garding claims 29–47 of the ’115 patent; (3) reverse in IPR2019-00811 regarding claims 86–89 of the ’115 patent; (4) reverse in IPR2019-00836 regarding claims 10–11 of the ’560 patent; and (5) affirm in IPR2019-00836 regarding claims 50–51 and 53–55 of the ’560 patent. We remand for such further proceedings as may be necessary and appro- priate regarding the claims on which we have reversed cer- tain Board determinations. The parties shall bear their own costs. AFFIRMED IN PART, REVERSED IN PART, AND REMANDED