Creative support for ad copy editors using lexical ambiguity

Patent metadata

Field	Value
Publication number	US-2019095960-A1
Application number	US-201715714590-A
Country	US
Kind code	A1
Filing date	Sep 25, 2017
Priority date	Sep 25, 2017
Publication date	Mar 28, 2019
Grant date	—

Abstract

According to one embodiment, a method, computer system, and computer program product for generating an advertisement is provided. The present invention may include receiving domain corpora; identifying domain-specific terms within each domain comprising the received domain corpora; scoring the identified domain-specific terms based on each of the domains; generating candidate short phrases comprising at least one of the scored domain-specific terms based on the scoring of the scored domain-specific terms; scoring the candidate short phrases; and selecting candidate short phrases based on the scoring of the candidate short phrases.

First claim

What is claimed is:

1 . A processor-implemented method for advertisement generation, the method comprising:
receiving one or more domain corpora; identifying one or more domain-specific terms within each of a plurality of domains comprising the one or more received domain corpora; scoring the one or more identified domain-specific terms based on each of the plurality of domains; generating one or more candidate short phrases comprising one or more scored domain-specific terms based on the scoring of the one or more scored domain-specific terms; scoring the one or more generated candidate short phrases; and selecting a scored candidate short phrase of the one or more scored candidate short phrases based on the scoring of the one or more candidate short phrases.

2 . The method of claim 1 , further comprising:
searching for one or more media elements in the plurality of domains that match the one or more scored domain-specific terms within the one or more candidate short phrases; scoring the one or more searched media elements; and selecting one or more combinations of one or more of the scored media elements and one or more of the scored candidate short phrases based on the scoring of both the one or more scored candidate short phrases and the one or more scored media elements.

3 . The method of claim 2 , wherein the one or more searched media elements are scored using a supervised machine learning model which is trained to assign a score to the one or more searched media elements based on a degree to which each one or more searched media elements is ad-like.

4 . The method of claim 1 , wherein the one or more candidate short phrases are further refined using a text simplification process.

5 . The method of claim 1 , wherein the scoring of the one or more of domain-specific terms and the scoring of the one or more candidate short phrases is performed by one or more supervised machine learning models.

6 . The method of claim 1 , wherein the plurality of domains comprise at least one product domain and at least one host domain.

7 . The method of claim 1 , wherein the one or more generated candidate short phrases are scored based on one or more factors, wherein the one or more factors are selected from a group consisting of advertising effectiveness, lexical humor, phrase length, coherence markers, action-oriented verbiage, and one or more word embeddings.

8 . A computer system for advertisement generation, the computer system comprising:
one or more processors, one or more computer-readable memories, one or more computer-readable tangible storage medium, and program instructions stored on at least one of the one or more tangible storage medium for execution by at least one of the one or more processors via at least one of the one or more memories, wherein the computer system is capable of performing a method comprising: receiving one or more domain corpora; identifying one or more domain-specific terms within each of a plurality of domains comprising the one or more received domain corpora; scoring the one or more identified domain-specific terms based on each of the plurality of domains; generating one or more candidate short phrases comprising one or more scored domain-specific terms based on the scoring of the one or more scored domain-specific terms; scoring the one or more generated candidate short phrases; and selecting a scored candidate short phrase of the one or more scored candidate short phrases based on the scoring of the one or more candidate short phrases.

9 . The computer system of claim 8 , further comprising:
searching for one or more media elements in the plurality of domains that match the one or more scored domain-specific terms within the one or more candidate short phrases; scoring the one or more searched media elements; and selecting one or more combinations of one or more of the scored media elements and one or more of the scored candidate short phrases based on the scoring of both the one or more scored candidate short phrases and the one or more scored media elements.

10 . The computer system of claim 9 , wherein the one or more searched media elements are scored using a supervised machine learning model which is trained to assign a score to the one or more searched media elements based on a degree to which each one or more searched media elements is ad-like.

11 . The computer system of claim 8 , wherein the one or more candidate short phrases are further refined using a text simplification process.

12 . The computer system of claim 8 , wherein the scoring of the one or more of domain-specific terms and the scoring of the one or more candidate short phrases is performed by one or more supervised machine learning models.

13 . The computer system of claim 8 , wherein the plurality of domains comprise at least one product domain and at least one host domain.

14 . The computer system of claim 8 , wherein the one or more generated candidate short phrases are scored based on one or more factors, wherein the one or more factors are selected from a group consisting of advertising effectiveness, lexical humor, phrase length, coherence markers, action-oriented verbiage, and one or more word embeddings.

15 . A computer program product for advertisement generation, the computer program product comprising:
one or more computer-readable tangible storage medium and program instructions stored on at least one of the one or more tangible storage medium, the program instructions executable by a processor to cause the processor to perform a method comprising: receiving one or more domain corpora; identifying one or more domain-specific terms within each of a plurality of domains comprising the one or more received domain corpora; scoring the one or more identified domain-specific terms based on each of the plurality of domains; generating one or more candidate short phrases comprising one or more scored domain-specific terms based on the scoring of the one or more scored domain-specific terms; scoring the one or more generated candidate short phrases; and selecting a scored candidate short phrase of the one or more scored candidate short phrases based on the scoring of the one or more candidate short phrases.

16 . The computer program product of claim 15 , further comprising:
searching for one or more media elements in the plurality of domains that match the one or more scored domain-specific terms within the one or more candidate short phrases; scoring the one or more searched media elements; and selecting one or more combinations of one or more of the scored media elements and one or more of the scored candidate short phrases based on the scoring of both the one or more scored candidate short phrases and the one or more scored media elements.

17 . The computer program product of claim 16 , wherein the one or more searched media elements are scored using a supervised machine learning model which is trained to assign a score to the one or more searched media elements based on a degree to which each one or more searched media elements is ad-like.

18 . The computer program product of claim 15 , wherein the one or more candidate short phrases are further refined using a text simplification process.

19 . The computer program product of claim 15 , wherein the scoring of the one or more of domain-specific terms and the scoring of the one or more candidate short phrases is performed by one or more supervised machine learning models.

20 . The compu

Assignees

• IBM

Inventors

• Byron Donna K
• Pinel Florian

Classifications

• G06F16/3347

  using vector based model · CPC title

• G06F40/295

  Named entity recognition · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/951

  Indexing; Web crawling techniques · CPC title