Value scorer in an automated disclosure assessment system

A computer program product comprising a computer readable storage medium having stored thereon:

• program instructions programmed to extract a set of topics from an idea description using topic extraction;
• program instructions programmed to search a set of databases to identify a set of works related to the set of topics, wherein each work of the set of works corresponds to one or more respective topics of the set of topics;
• program instructions programmed to search the set of databases to identify a set of randomly selected works related to a set of randomly determined topics, wherein each randomly selected work of the set of randomly selected works corresponds to one or more respective randomly determined topics of the set of randomly determined topics;
• program instructions programmed to automatically generate a set of related publication date densities of the set of works by, for each topic of the set of topics:
• weighing the works of the set of works that correspond to the topic by, for each work:
• (i) determining a distance between a publication date of the work and the present date, (ii) multiplying the determined distance by a predetermined value, (iii) calculating the exponential of the difference between one and the multiplied distance, and (iv) using the calculated exponential as the weight for the work, and
• calculating a sum of the weights for the works that correspond to the topic;
• program instructions programmed to automatically generate a set of baseline publication date densities of the set of randomly selected works by, for each randomly determined topic of the set of randomly determined topics:
• weighing the randomly selected works of the set of randomly selected works that correspond to the randomly determined topic by, for each randomly selected work:
• (i) determining a distance between a publication date of the randomly selected work and the present date, (ii) multiplying the determined distance for the randomly selected work by the predetermined value, (iii) calculating the exponential of the difference between one and the multiplied distance for the randomly selected work, and (iv) using the calculated exponential for the randomly selected work as the weight for the randomly selected work, and
• calculating a sum of the weights for the randomly selected works that correspond to the randomly determined topic;
• program instructions programmed to generate a density score for the idea description based, at least in part, on a comparison between the set of baseline publication date densities and the set of related publication date densities;
• program instructions programmed to compute a similarity of the idea description to the works of the set of works using distributional semantics;
• and
• program instructions programmed to assess a value of the idea description as a patent application based, at least in part, on the density score and the computed similarity to the works of the set of works, wherein the density score and the computed similarity to the works of the set of works are each weighed by respective weights that have been optimized through training of an artificial neural network.

A computer system comprising:

• a processor(s) set;
• and
• a computer readable storage medium;
• wherein:
• the processor set is structured, located, connected and/or programmed to run program instructions stored on the computer readable storage medium;
• and
• the stored program instructions include:
• program instructions programmed to extract a set of topics from an idea description using topic extraction;
• program instructions programmed to search a set of databases to identify a set of works related to the set of topics, wherein each work of the set of works corresponds to one or more respective topics of the set of topics;
• program instructions programmed to search the set of databases to identify a set of randomly selected works related to a set of randomly determined topics, wherein each randomly selected work of the set of randomly selected works corresponds to one or more respective randomly determined topics of the set of randomly determined topics;
• program instructions programmed to automatically generate a set of related publication date densities of the set of works by, for each topic of the set of topics:
• weighing the works of the set of works that correspond to the topic by, for each work:
• (i) determining a distance between a publication date of the work and the present date, (ii) multiplying the determined distance by a predetermined value, (iii) calculating the exponential of the difference between one and the multiplied distance, and (iv) using the calculated exponential as the weight for the work, and
• calculating a sum of the weights for the works that correspond to the topic;
• program instructions programmed to automatically generate a set of baseline publication date densities of the set of randomly selected works by, for each randomly determined topic of the set of randomly determined topics:
• weighing the randomly selected works of the set of randomly selected works that correspond to the randomly determined topic by, for each randomly selected work:
• (i) determining a distance between a publication date of the randomly selected work and the present date, (ii) multiplying the determined distance for the randomly selected work by the predetermined value, (iii) calculating the exponential of the difference between one and the multiplied distance for the randomly selected work, and (iv) using the calculated exponential for the randomly selected work as the weight for the randomly selected work, and
• calculating a sum of the weights for the randomly selected works that correspond to the randomly determined topic;
• program instructions programmed to generate a density score for the idea description based, at least in part, on a comparison between the set of baseline publication date densities and the set of related publication date densities;
• program instructions programmed to compute a similarity of the idea description to the works of the set of works using distributional semantics;
• and
• program instructions programmed to assess a value of the idea description as a patent application based, at least in part, on the density score and the computed similarity to the works of the set of works, wherein the density score and the computed similarity to the works of the set of works are each weighed by respective weights that have been optimized through training of an artificial neural network.