Abstract: Systems and methods include: implementing a first machine learning model to generate an output of a global digital threat score for an online activity based on an input of the collected digital event data; implementing a second machine learning model that generates a category inference of a category of digital fraud or a category of digital abuse from a plurality of digital fraud or digital abuse categories; selecting a third machine learning model from an ensemble of digital fraud or digital abuse machine learning models based on the category inference generated by the second machine learning model, wherein the ensemble of digital fraud or digital abuse machine learning models comprise a plurality of disparate digital fraud or digital abuse category-specific machine learning models; and implementing the selected third machine learning model to generate a digital fraud or digital abuse category-specific threat score based on the digital event data.

Abstract: Systems and methods include: implementing a first machine learning model to generate an output of a global digital threat score for an online activity based on an input of the collected digital event data; implementing a second machine learning model that generates a category inference of a category of digital fraud or a category of digital abuse from a plurality of digital fraud or digital abuse categories; selecting a third machine learning model from an ensemble of digital fraud or digital abuse machine learning models based on the category inference generated by the second machine learning model, wherein the ensemble of digital fraud or digital abuse machine learning models comprise a plurality of disparate digital fraud or digital abuse category-specific machine learning models; and implementing the selected third machine learning model to generate a digital fraud or digital abuse category-specific threat score based on the digital event data.

Abstract: Systems and methods include: implementing a first machine learning model to generate an output of a global digital threat score for an online activity based on an input of the collected digital event data; implementing a second machine learning model that generates a category inference of a category of digital fraud or a category of digital abuse from a plurality of digital fraud or digital abuse categories; selecting a third machine learning model from an ensemble of digital fraud or digital abuse machine learning models based on the category inference generated by the second machine learning model, wherein the ensemble of digital fraud or digital abuse machine learning models comprise a plurality of disparate digital fraud or digital abuse category-specific machine learning models; and implementing the selected third machine learning model to generate a digital fraud or digital abuse category-specific threat score based on the digital event data.

Abstract: Systems and methods include: implementing a first machine learning model to generate an output of a global digital threat score for an online activity based on an input of the collected digital event data; implementing a second machine learning model that generates a category inference of a category of digital fraud or a category of digital abuse from a plurality of digital fraud or digital abuse categories; selecting a third machine learning model from an ensemble of digital fraud or digital abuse machine learning models based on the category inference generated by the second machine learning model, wherein the ensemble of digital fraud or digital abuse machine learning models comprise a plurality of disparate digital fraud or digital abuse category-specific machine learning models; and implementing the selected third machine learning model to generate a digital fraud or digital abuse category-specific threat score based on the digital event data.

Abstract: Systems and methods include: implementing a first machine learning model to generate an output of a global digital threat score for an online activity based on an input of the collected digital event data; implementing a second machine learning model that generates a category inference of a category of digital fraud or a category of digital abuse from a plurality of digital fraud or digital abuse categories; selecting a third machine learning model from an ensemble of digital fraud or digital abuse machine learning models based on the category inference generated by the second machine learning model, wherein the ensemble of digital fraud or digital abuse machine learning models comprise a plurality of disparate digital fraud or digital abuse category-specific machine learning models; and implementing the selected third machine learning model to generate a digital fraud or digital abuse category-specific threat score based on the digital event data.

Abstract: Systems and methods include: collecting digital event data from at least one remote source of digital event data; using the collected digital event data as input into primary machine learning ensemble that predicts the likelihood of digital fraud and/or digital abuse; generating by the machine learning system the global digital threat score; identifying a sub-request for a specific digital threat score for a digital abuse type; in response to identifying the sub-request, providing the input of the collected digital event data to a secondary machine learning model ensemble of the machine learning system that predicts a likelihood of the identified digital abuse type; generating by the secondary machine learning ensemble the specific digital threat score for the digital abuse type based on the input of the collected digital event data; and transmitting the global digital threat score and the specific digital threat score for the identified digital abuse type.

Abstract: Documents or document sets may be scored to reflect a value of an action, such as a selection for example, when an ad is served with the document (or a document belonging to a document set). A cost associated with the action with respect to an ad that was served with a document may then be adjusted using the score. For example, ad scores may be accepted or determined, and a document may be scored using the ad scores when served with the document and ad scores across a collection of documents to generate a document score. Each of the ad scores may indicate a value of an action with respect to an ad, such as a conversion rate, or a return on investment for an ad selection for example. Document scores used in this way may help advertisers get a more consistent cost per conversion, or return on investment, without requiring them to enter and manage various offers for various documents and/or various ad serving systems having various conversion rates or returns on investment.

Abstract: A system determines query similarity. The system determines a volume per unit time of an issued first query over a time period and determines a volume per unit time of issued other queries over the time period. The system compares the volume per unit time of each of the issued other queries to the volume per unit time of the issued first query. The system identifies ones of the issued other queries as similar to the first query based on the comparison.

Abstract: Documents or document sets may be scored to reflect a value of an action, such as a selection for example, when an ad is served with the document (or a document belonging to a document set). A cost associated with the action with respect to an ad that was served with a document may then be adjusted using the score. For example, ad scores may be accepted or determined, and a document may be scored using the ad scores when served with the document and ad scores across a collection of documents to generate a document score. Each of the ad scores may indicate a value of an action with respect to an ad, such as a conversion rate, or a return on investment for an ad selection for example. Document scores used in this way may help advertisers get a more consistent cost per conversion, or return on investment, without requiring them to enter and manage various offers for various documents and/or various ad serving systems having various conversion rates or returns on investment.

Abstract: Documents or document sets may be scored to reflect a value of an action, such as a selection for example, when an ad is served with the document (or a document belonging to a document set). A cost associated with the action with respect to an ad that was served with a document may then be adjusted using the score. For example, ad scores may be accepted or determined, and a document may be scored using the ad scores when served with the document and ad scores across a collection of documents to generate a document score. Each of the ad scores may indicate a value of an action with respect to an ad, such as a conversion rate, or a return on investment for an ad selection for example. Document scores used in this way may help advertisers get a more consistent cost per conversion, or return on investment, without requiring them to enter and manage various offers for various documents and/or various ad serving systems having various conversion rates or returns on investment.

Abstract: This disclosure generally relates to assigning and simultaneously running multiple client-side experiments on client devices. A file includes information regarding experiments that are available, including information regarding “layers,” which are logical, imaginary containers in which each experiment “resides.” Each experiment is associated with one layer. For each experiment, the file includes information regarding a location and size of the experiment within the layer. When the client device takes an action, a software module identifies a value of an identifier associated with the action. Each such identifier is associated with one or more of the layers. The software module can calculate, for each of the associated layers, a location within the layer based on the identifier value. The computer software module can identify, based on the information in the file, each experiment that overlaps with the calculated location within each layer and cause each identified experiment to be activated.

Abstract: A system determines query similarity. The system determines a volume per unit time of an issued first query over a time period and determines a volume per unit time of issued other queries over the time period. The system compares the volume per unit time of each of the issued other queries to the volume per unit time of the issued first query. The system identifies ones of the issued other queries as similar to the first query based on the comparison.

Abstract: Documents or document sets may be scored to reflect a value of an action, such as a selection for example, when an ad is served with the document (or a document belonging to a document set). A cost associated with the action with respect to an ad that was served with a document may then be adjusted using the score. For example, ad scores may be accepted or determined, and a document may be scored using the ad scores when served with the document and ad scores across a collection of documents to generate a document score. Each of the ad scores may indicate a value of an action with respect to an ad, such as a conversion rate, or a return on investment for an ad selection for example. Document scores used in this way may help advertisers get a more consistent cost per conversion, or return on investment, without requiring them to enter and manage various offers for various documents and/or various ad serving systems having various conversion rates or returns on investment.