Systems and methods for predicting qualitative ratings for advertisements based on machine learning

1. A computer-implemented method comprising:

determining, by a computing system, a representation of an advertisement based on a first machine learning model, wherein the representation is provided to a second machine learning model;

providing, by the computing system, the representation to the second machine learning model, wherein the second machine learning model is trained based on training data including representations of a plurality of advertisements and associated qualitative ratings; and

determining, by the computing system, one or more qualitive ratings associated with the advertisement based on the second machine learning model, wherein the one or more qualitative ratings include a rating associated with a call-to-action (CTA).

2. The computer-implemented method of claim 1 , wherein the first machine learning model is a neural network that includes a plurality of layers, and the representation of the advertisement includes a set of features determined based on the neural network.

3. The computer-implemented method of claim 2 , wherein the set of features includes an output of each neuron of a fully connected layer of a convolutional neural network.

4. The computer-implemented method of claim 2 , wherein each feature of the set of features indicates a likelihood of an attribute being associated with the advertisement.

5. The computer-implemented method of claim 4 , wherein the attribute being associated with the advertisement is a visual attribute, the visual attribute including at least one of an object, a concept, a theme, an animal, or a person depicted in the advertisement.

6. The computer-implemented method of claim 4 , wherein the attribute being associated with the advertisement is a nonvisual attribute, the nonvisual attribute including metadata information associated with the advertisement.

7. The computer-implemented method of claim 2 , wherein the representation of the advertisement is a feature vector including the set of features expressed as values.

8. The computer-implemented method of claim 1 , wherein the one or more qualitative ratings is-further include a rating associated with at least one of noticeability, a focal point, interesting information, or an emotional reward.

9. The computer-implemented method of claim 1 , wherein a qualitative rating is associated with visual content of an advertisement.

10. The computer-implemented method of claim 1 , wherein the second machine learning model is based at least in part on one or more of: a linear regression model, a random forest, or a neural network.

11. The computer-implemented method of claim 1 , wherein the determining the representation of the advertisement includes providing pixel data for the advertisement to the first machine learning model as input.

12. The computer-implemented method of claim 1 , wherein the first machine learning model is trained based on training data including pixel data for a plurality of advertisements and associated representations of the plurality of advertisements.

13. A system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the system to perform:

determining a representation of an advertisement based on a first machine learning model, wherein the representation is provided to a second machine learning model;

providing the representation to the second machine learning model, wherein the second machine learning model is trained based on training data including representations of a plurality of advertisements and associated qualitative ratings; and

determining one or more qualitive ratings associated with the advertisement based on the second machine learning model, wherein the one or more qualitative ratings include a rating associated with a call-to-action (CTA).

14. The system of claim 13 , wherein the first machine learning model is a neural network that includes a plurality of layers, and the representation of the advertisement includes a set of features determined based on the neural network.

15. The system of claim 13 , wherein the second machine learning model is based at least in part on one or more of: a linear regression model, a random forest, or a neural network.

16. The system of claim 13 , wherein the first machine learning model is trained based on training data including pixel data for a plurality of advertisements and associated representations of the plurality of advertisements.

17. A non-transitory computer readable medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:

determining a representation of an advertisement based on a first machine learning model, wherein the representation is provided to a second machine learning model;

providing the representation to the second machine learning model, wherein the second machine learning model is trained based on training data including representations of a plurality of advertisements and associated qualitative ratings; and

determining one or more qualitive ratings associated with the advertisement based on the second machine learning model, wherein the one or more qualitative ratings include a rating associated with a call-to-action (CTA).

18. The non-transitory computer readable medium of claim 17 , wherein the first machine learning model is a neural network that includes a plurality of layers, and the representation of the advertisement includes a set of features determined based on the neural network.

19. The non-transitory computer readable medium of claim 17 , wherein the second machine learning model is based at least in part on one or more of: a linear regression model, a random forest, or a neural network.

20. The non-transitory computer readable medium of claim 17 , wherein the first machine learning model is trained based on training data including pixel data for a plurality of advertisements and associated representations of the plurality of advertisements.