Claims for Analyzing Twitter Posts Held Unpatentable by Judge Castel
On October 29, 2018, United States District Judge P. Kevin Castel (S.D.N.Y.) issued a decision granting Defendant Bloomberg's Rule 12(b)(6) motion to dismiss iSentium's patent infringement claim because it is directed to subject matter that is not eligible for patenting under 35 U.S.C. § 101.
iSentium's suit alleges infringement of United States Patent No. 8,856,056, a patent directed to a method for evaluating social media posts that discuss publicly traded assets. In particular, the claimed method purports to evaluate each post's polarity, i.e., whether the post expresses a positive, negative or neutral opinion, and strength, i.e., how strongly the opinion is held. The specification discloses that determination of polarity and strength is based on databases of known phrases and further grammatical analysis, and contends that changes in an asset’s price can be anticipated based on this analysis.
Bloomberg, argued that the claims of the '056 patent are directed to an abstract idea that is not patentable under 35 U.S.C. § 101. Under step one of Alice's two-part test for patentable subject matter, Judge Castel held that the '056 patent was directed to the ineligible abstract idea of "collecting statements from social media and identifying opinions through the use of algorithms and databases." In support, the decision cited Federal Circuit decisions holding that data analysis is an abstract idea, even when performed by high speed computers.
Judge Castel next proceeded to step two of the Alice test, the search for an inventive concept that would transform an abstract idea into a patent-eligible invention. For this step, iSentium urged the Court to consider claim 1, a means-plus-function claim to a sentiment calculator comprising means for determining a messages polarity and strength "based upon pairs of lexical items in local syntactic context." According to iSentium, the corresponding algorithm disclosed in the specification contained the required inventive concept that conferred patent eligibility. Judge Castel held otherwise.
At oral argument, counsel for iSentium described the purported inventive algorithm of claim 1 as “the consideration of words in relation to each other.” But when asked for further detail, counsel responded: "It is something more complicated and we have a patent that does not disclose the way of performing this" (emphasis in opinion). Judge Castel concluded that the consideration of words in relation to each other and the consideration of words other than adjectives, without further specificity, is conventional, and not inventive, even assuming that this form of language processing was an improvement over pure word-search based prior art algorithms.
Case: iSentium, LLC v. Bloomberg Finance LP, 1-17-cv-07601(PKC), Dkt. No. 32 (S.D.N.Y. October 29, 2018).