Interthinx Inc v. CoreLogic Solutions LLC

1. Case Identification

• Patent #: 5,361,201
• Filed: September 18, 2012
• Petitioner(s): Interthinx, Inc.
• Patent Owner(s): CoreLogic Information Solutions, Inc.
• Challenged Claims: 1, 5, 6, 9, and 10

2. Patent Overview

• Title: Real Estate Appraisal Using Predictive Modeling
• Brief Description: The ’201 patent discloses computer-implemented methods for appraising real estate. The methods involve using predictive modeling and algorithms to estimate property values, aiming to reduce the human bias found in conventional appraisals by applying statistical techniques.

3. Grounds for Unpatentability

Ground 1: Ineligibility under §101 - Claims 1, 5, 6, 9, and 10 are invalid under 35 U.S.C. §101.

• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner argued that the challenged claims were directed to the abstract idea of real estate appraisal, a fundamental economic practice and mental process. It was asserted that the claims merely automated a classic business method and that the steps—collecting data, analyzing it with statistical models, and determining a value—could be performed in the human mind. Petitioner cited Federal Circuit precedent, including Bilski and CyberSource, to argue that implementing such a mental process on a generic computer does not confer patent eligibility. The claim's recitation of a computer was argued to be insignificant post-solution activity, as the core of the invention was the mathematical algorithm, not a technological improvement.
  • Key Aspects: A central contention was that the lead inventor admitted the claimed process imitates the steps a manual appraiser would take, and that the invention was simply the "statistical process of developing the [value] estimate."

Ground 2: Anticipation by Tay - Claims 1, 5, 9, and 10 are anticipated by Tay under 35 U.S.C. § 102(a).

• Prior Art Relied Upon: Tay et al., “Artificial Intelligence and the Mass Appraisal of Residential Apartments,” 10 Journal of Property Valuation and Investment 525-540 (1991-1992) (“Tay”).
• Core Argument for this Ground:
  • Prior Art Mapping: Petitioner asserted that Tay disclosed every element of the challenged claims. Tay described a computer-implemented system using back-propagation artificial neural networks (ANNs) and multiple regression analysis (MRA) for valuing residential apartments. Petitioner argued Tay taught collecting training data (apartment sales data), using that data to develop predictive models (the ANN and MRA models), storing the models in computer memory, and generating signals indicative of an appraised value. Tay also allegedly disclosed developing and storing an error model by comparing the estimated sale price to the actual price and using a "backward pass" to correct weights, thereby generating a signal for an error range.

Ground 3: Obviousness over Tay - Claims 5, 6, 9, and 10 are obvious over Tay under 35 U.S.C. § 103.

• Prior Art Relied Upon: Tay

• Core Argument for this Ground:

  • Prior Art Mapping: To the extent any claim limitations were not explicitly taught by Tay, Petitioner argued they would have been obvious. For instance, claim 5’s step of aggregating property training data into area-specific data sets was argued to be an obvious modification. Petitioner contended it was well-known at the time of filing to select and group real estate data according to geographic areas to improve appraisal accuracy. Similarly, claim 9’s steps of selecting a geographic area and obtaining area data were presented as obvious techniques for a person of ordinary skill in the art (POSITA) to apply.
  • Motivation to Combine: The motivation was to apply a known technique (geographic data aggregation) to a known method (Tay’s automated appraisal system) to improve its performance, which Petitioner argued was a simple matter of design choice.
  • Expectation of Success: A POSITA would have had a high expectation of success because applying geographic-based analysis to real estate valuation was a conventional and predictable method for improving model accuracy.

• Additional Grounds: Petitioner asserted numerous additional anticipation and obviousness challenges against the claims, including grounds based on Lu et al. (“Neurocomputing Approach to Residential Property Valuation”), Boyle (“An Expert System for Valuation of Residential Properties”), Jensen (“Artificial Intelligence in Computer-Assisted Mass Appraisal”), Carbone (“A Feedback Model for Automated Real Estate Assessment”), Des Rosiers (“Integrating Geographic Information Systems to Hedonic Price Modeling”), and Eckert (“Property Appraisal and Assessment Administration”).

4. Key Claim Construction Positions

• "computer-implemented process": Petitioner argued for the construction provisionally adopted in related litigation: "Plain and ordinary meaning, which does not require a general purpose computer and which does not exclude human interaction or input." This construction was critical to Petitioner's §101 argument that the claims covered unpatentable mental processes that could be performed with or without a computer.
• "predictive model": Petitioner supported the construction: "Plain and ordinary meaning, which is not limited to a neural network and which does not exclude a regression model." This broad construction allowed Petitioner to map prior art references that used various statistical techniques, such as the ANNs in Tay and the MRA in Boyle, to the claims.

5. Key Technical Contentions (Beyond Claim Construction)

• Petitioner’s central technical contention was that the ’201 patent did not claim a true technological invention. The petition argued that the patent’s claimed contribution was merely the application of known mathematical and statistical techniques to the field of real estate appraisal. It contended that the process steps were not tied to any specific machine and that the use of a computer was simply as a tool to automate mental processes, which does not solve a technical problem with a technical solution.

6. Relief Requested

• Petitioner requested the institution of a Transitional Post-Grant Review for a Covered Business Method Patent and the cancellation of claims 1, 5, 6, 9, and 10 of the ’201 patent as unpatentable under 35 U.S.C. §§ 101, 102, and 103.