“Method, Apparatus, And Computer Program Product For Identifying Hazardous Conditions And Predicting Policy Transaction Behavior” in Patent Application Approval Process (USPTO 20190080415)

2019 MAR 29 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- A patent application by the inventors Kim, Sung Hoon (Boston, MA); Gadgil, Shweta (Cambridge, MA); Harvey, Joshua (Cambridge, MA); Li, Lingge (Irvine, CA), filed on September 7, 2018, was made available online on March 14, 2019, according to news reporting originating from Washington, D.C., by NewsRx correspondents.

This patent application is assigned to Liberty Mutual Insurance Company (Boston, Massachusetts, United States).

The following quote was obtained by the news editors from the background information supplied by the inventors: “External physical plots of land and structures have hazards associated with them. Programmatically making decisions to provide policies to insure such external plots and structures can be difficult in view of such hazards. Through applied effort, ingenuity, and innovation, many of these identified problems have been solved by developing solutions that are included in embodiments of the present disclosure, many examples of which are described in detail herein.”

In addition to the background information obtained for this patent application, NewsRx journalists also obtained the inventors’ summary information for this patent application: “Embodiments of the present disclosure relate to concepts for predicting policy transaction behavior. In one embodiment, a computing entity or apparatus is configured to receive, from a remote computing device, a plurality of digital image files associated with the external plot of physical land. In embodiments, each digital image file of the plurality of digital image files is representative of a different field of view of the external plot of physical land. In embodiments, the computing entity or apparatus is further configured to extract, from each digital image file of the plurality of digital image files, one or more image-based plot properties associated with the external plot of physical land. In embodiments, the extracting is performed on a pixel by pixel basis and using a first trained machine learning model. In embodiments, the first trained machine learning model outputs a predicted image-based plot property based on one or more pixels of a digital image file and a confidence measure associated with the predicted image-based plot property. In embodiments, the computing entity or apparatus is further configured to generate, for each image-based plot property and using a second trained machine learning model, a policy transaction prediction. In embodiments, the policy transaction prediction being a programmatically generated likelihood that the image-based plot property will lead to a policy transaction occurrence. In embodiments, the computing entity or apparatus is further configured to, based on the policy transaction predictions, generate an aggregated policy transaction prediction for the external plot of physical land. In embodiments, the aggregated policy transaction prediction being a programmatically generated likelihood that a policy transaction will occur subsequent an insurance policy issuance for the external plot of physical land. In embodiments, the computing entity or apparatus is further configured to transmit, to a requesting computing device, the aggregated policy transaction prediction.

“The details of one or more embodiments of the subject matter described in this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.”

The claims supplied by the inventors are:

“1. An apparatus for predicting policy transaction behavior, the apparatus comprising at least one processor and at least one memory including computer program code, the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to: receive, from a remote computing device, a plurality of digital image files associated with the external plot of physical land, each digital image file of the plurality of digital image files representative of a different field of view of the external plot of physical land; extract, from each digital image file of the plurality of digital image files, one or more image-based plot properties associated with the external plot of physical land, wherein the extracting is performed on a pixel by pixel basis and using a first trained machine learning model, and wherein the first trained machine learning model outputs a predicted image-based plot property based on one or more pixels of a digital image file and a confidence measure associated with the predicted image-based plot property; generate, for each image-based plot property and using a second trained machine learning model, a policy transaction prediction, the policy transaction prediction being a programmatically generated likelihood that the image-based plot property will lead to a policy transaction occurrence; based on the policy transaction predictions, generate an aggregated policy transaction prediction for the external plot of physical land, the aggregated policy transaction prediction being a programmatically generated likelihood that a policy transaction will occur subsequent an insurance policy issuance for the external plot of physical land; and transmit, to a requesting computing device, the aggregated policy transaction prediction.

“2. The apparatus of claim 1, further configured to generate the aggregated policy transaction prediction for the external plot of physical land prior to an issuance of an insurance policy for the external plot of physical land.

“3. The apparatus of claim 2, wherein the policy transaction is a policy termination.

“4. The apparatus of claim 3, wherein the policy termination is initiated by an entity issuing the insurance policy for the external plot of physical land.

“5. The apparatus of claim 1, wherein the external plot of physical land includes one or more structures.

“6. The apparatus of claim 1, wherein the apparatus is further caused to assign a price value to an insurance policy for the external plot of physical land based at least on the aggregated policy transaction prediction.

“7. The apparatus of claim 1, wherein the apparatus is further caused to assign a risk value to an insurance policy for the external plot of physical land based at least on the aggregated policy transaction prediction.

“8. The apparatus of claim 1, wherein an image-based plot property of the plurality of plot properties is one of a tree overhang on a roof, presence of solar panels on a roof, presence of skylight windows on a roof, presence of unnatural objects on a roof top, roof material, roof surface, roof pitch, detached structures on the external plot of physical land, presence of a pool on the external plot of physical land, a pool shape, presence of barriers around insured property and between adjacent properties, distance to brush and trees, buffer between house or structure and surrounding vegetation, presence of ADT, presence of a “Beware Of Dog” sign, presence of a “Do Not Trespass” sign, cracks in a driveway, peeling paint from a structure, or signs of rust on a dwelling.

“9. The apparatus of claim 1, wherein the apparatus is further caused to: determine that one or more of the image-based plot properties represents one or more hazardous image-based plot properties; assign a severity to each hazardous image-based plot property of the one or more hazardous image-based plot properties; and return an indication of each of the one or more hazardous image-based plot properties.

“10. The apparatus of claim 9, wherein the apparatus is further caused to: generate a hazard score for the external plot of physical land; and use the generated hazard score for generating one or more policy transaction predictions.

“11. The apparatus of claim 10, wherein the hazard score is generated based at least on a count of the hazardous image-based plot properties and a severity of each of the one or more hazardous image-based plot properties.

“12. The apparatus of claim 1, wherein the apparatus is further caused to: receive an underwriting status and input the underwriting status into the second machine learning model.

“13. The apparatus of claim 1, wherein the first machine learning model is a convolutional neural network.

“14. The apparatus of claim 13, wherein the first machine learning model is trained using digital images of known plot properties.

“15. The apparatus of claim 14, wherein the first machine learning model provides image-based plot properties based upon any arbitrary number of received digital image files.

“16. The apparatus of claim 1, wherein the second machine learning model is a tree based machine learning model.

“17. The apparatus of claim 16, wherein the second machine learning model is trained using known plot properties and known image-based plot properties.

“18. The apparatus of claim 17, wherein the second machine learning model generates a programmatically generated prediction representing a likelihood that a hazard exists in any of the image-based plot properties.

“19. The apparatus of claim 18, wherein the second machine learning model uses the existence or lack of a hazard to programmatically generate a policy transaction prediction.

“20. A computer-implemented method for predicting policy transaction behavior, the method comprising: receiving, using a processor and from a remote computing device, a plurality of digital image files associated with the external plot of physical land, each digital image file of the plurality of digital image files representative of a different field of view of the external plot of physical land; extracting, using the processor, from each digital image file of the plurality of digital image files, one or more image-based plot properties associated with the external plot of physical land, wherein the extracting is performed on a pixel by pixel basis and using a first trained machine learning model, and wherein the first trained machine learning model outputs a predicted image-based plot property based on one or more pixels of a digital image file and a confidence measure associated with the predicted image-based plot property; generating, using the processor, for each image-based plot property and using a second trained machine learning model, a policy transaction prediction, the policy transaction prediction being a programmatically generated likelihood that the image-based plot property will lead to a policy transaction occurrence; based on the policy transaction predictions, generating, using the processor, an aggregated policy transaction prediction for the external plot of physical land, the aggregated policy transaction prediction being a programmatically generated likelihood that a policy transaction will occur subsequent an insurance policy issuance for the external plot of physical land; and transmitting, using the processor and to a requesting computing device, the aggregated policy transaction prediction.”

URL and more information on this patent application, see: Kim, Sung Hoon; Gadgil, Shweta; Harvey, Joshua; Li, Lingge. Method, Apparatus, And Computer Program Product For Identifying Hazardous Conditions And Predicting Policy Transaction Behavior. Filed September 7, 2018 and posted March 14, 2019. Patent URL: http://appft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PG01&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.html&r=1&f=G&l=50&s1=%2220190080415%22.PGNR.&OS=DN/20190080415&RS=DN/20190080415

(Our reports deliver fact-based news of research and discoveries from around the world.)