Context Directions, a Jeffrey M. Gross-controlled NPE who had asserted two patents in the Western District of Texas ostensibly related to detecting the context in which a mobile device is used, recently had the USPTO enter a substantially narrow construction of one of those patents, likely rendering it (and its family members) useless for future litigation. The office found the claims were means-plus-function claims, and the patent owner effectively conceded the claims were abstract and lacked algorithmic support—in other words, Context Directions conceded that the claims are indefinite.

On March 29, 2021, Unified Patents filed an ex parte reexamination against certain claims of U.S. Patent 10,142,791, assigned to Context Directions, LLC, a Jeffrey M. Gross NPE. A month later, the office granted reexamination and later entered a non-final office action rejecting all of the challenged claims, including a finding that the claims contained means-plus-function limitations subject to 35 U.S.C. § 112 ¶ 6. App. No. 90//014,712, Non-Final Office Action, 10- (June 11, 2021).

First, the examiner found that limitations referencing a “classifier” performing the function of “evaluat[ing] one or more contexts of the mobile device” lacked sufficient structure on their own to perform the claimed function and, therefore, were subject to § 112 ¶ 6. Id., 11-13; see also 24. The examiner also confirmed that the claimed classifier was software, and that the corresponding three disclosures in the specification as the corresponding disclosure:

"[T]he classifiers 116, 126, 136 may be implemented in a form of electronic circuits performing specific function in the non-volatile memory 103 of the device 100, loaded into RAM 102 when needed and executed by the processor 101." C8:L9-14, 791 Patent.

"The selection of determined signal features can be made in many ways- arbitrarily or objectively, for example, using the method of principal component analysis (PCA)." C8:L36-39.

"The classifier maps the features vector to a value, that indicates whether the given context is present or not." C8:L50-53.

Id., 15-16. Notably, none of the identified structures are algorithms, rendering the claims invalid before any court.

The examiner also concluded that the claimed “mobile device,” configured to activate a classification by classifiers assigned to first and second sensor groups and adapt a configuration of a classifier assigned to the first sensor group, was subject to § 112 ¶ 6. Id., 23. The examiner identified no corresponding algorithm for these limitations, and instead only identified a figure depicting a mobile device. Id., 23-24; see also 26. Again, the examiner was unable to identify a corresponding algorithm to perform the claimed concept of “evaluating” a context of a mobile device, and instead only identified a figure of a mobile device and a list of devices that might qualify, such as a phone or camera. This is not enough.

In response, the applicant argued that the claims did not invoke § 112 ¶ 6, and that a classifier was a term of art having broad meaning. App. No. 90/014,712, Am/Req. for Reconsideration after Non-Final Reject., 10-15 (Aug. 10, 2021), and explained that “many classifiers are quantizers of numerical values that assign a label designating one of several possible classes to a range of numerical values,” giving examples of medical tests, school grades, hurricane/tornado severity, and automotive warning lights. Id., 15-16.  Of note, Patent Owner’s own statements here suggest the claims are unpatentably abstract, as claiming they cover all of these conventional classification activities simply implemented by software.

The examiner rejected the applicant’s characterization of the means-plus-function limitations, reaching the same conclusions as it had in its office action regarding claim construction. App. No. 90//014,712, Notice of Intent to Issue Ex Parte Reexamination Certificate, 10-15 (Aug. 10, 2021). It then changed course from its findings regarding obviousness and issued a notice of intent to issue a reexamination certificate, with little explanation.

Regardless, it is axiomatic that, under Aristocrat Technologies v. Intern. Game Techs. and Williamson v. Citrix, software means-plus-function limitations are limited to algorithms disclosed in the specification. Although examiners in reexamination are unable to cancel original claims for indefiniteness (see 37 CFR 1.552), given that the patent’s disclosure (as cited by the examiner) lacks algorithmic support, it would be objectively baseless for Context Directions to argue otherwise in court. What’s more, the ’791 patent clearly lacks any disclosure of an algorithm sufficient to support the broad functional language of the claims, and the patent owner’s own statements suggest the claims are abstract. The issuing claims are thus invalid under at least Citrix, even if not obvious over the prior art cited during reexamination. Any further enforcement activity on this patent could be considered sanctionable as objectively baseless, particularly if Context Directions refuses to identify any algorithms or how the accused systems map to a given alleged algorithm in the specification.

Not only that, the examiner’s late change of heart was in error, as the Office overlooked that a reference in Unified’s request that disclosed separate groups of hardware and software components working together to perform the same classifications as exemplary embodiments disclosed in the specification of the ’791 patent; additionally addition, another reference in the request provided additional examples of how to perform classifications of the user’s actions based on a comparison to threshold values (e.g., classifying the user as driving or not based on readings from different acceleration and GPS circuits). See Request, Exhibit AA, 11-34 (mapping both Lindquist and Dietz to the classifier limitations).

The ’791 patent, and likely its family members, are defective on their face as a result of the examiner’s conclusions and the applicant’s arguments during reexamination. Indeed, even if the examiner is wrong about the means-plus-function limitations, the applicant’s characterizations of the classifier concepts raise abstractness concerns under 35 U.S.C. § 101. Any assertion efforts may justify fees under F.R.C.P. 11 and 35 U.S.C. § 285 and may constitute inequitable conduct.