Yahoo Ad Tech LLC

The present teaching relates to identifying content that matches a query. Train data include queries, advertisement, and hyperlinks associated with query sessions. A plurality of subwords for each of the queries in the training data are identified. A query/ads model is then trained by optimizing vectors associated the plurality of subwords for each of the queries, advertisements, and hyperlinks in the training data with respect to an objective function. At least one vector associated with each of the queries is derived based on the plurality of subword vectors in the query/ads model that represent the plurality of subwords of the query.

The instant system and methods solves the cold start problem through various systems and methods directed to aggregating user interaction data associated with a user over a period of time, generating an embedding model based on the aggregated user interaction data, generating a content embedding vector based on the embedding model, generating an embedding profile vector based on the embedding model, storing the embedding profile vector in a storage device, receiving each of the content embedding vector and embedding vector profile for training a ranking model, and generating a predicted list of one or more content items of interest for recommending to the user.

The present teaching relates to method, system, medium, and implementations for approximating a non-linear relationship between a numerical feature and an output of a model. A value of a numerical feature is received and is transformed, via a transform function, into a transformed value within a fixed range. With respect to each of a plurality of basis functions used for approximating the non-linear relationship, a respective basis function value of the basis function is computed based on the transformed value. An approximated value of the non-linear numeric feature is generated based on a sum of the plurality of basis function values weighted respectively by each corresponding one of a set of the weights, obtained via machine learning.

Embodiments of the present invention are directed to detecting misuse of contextual data, such as user data. In accordance with some embodiments of the present invention, artificial contextual data is created for artificial users and is, thereafter, provided to a bidder(s). Based on reception of artificial contextual data, bidders can place bids for presentation of advertisements. Bids placed by bidders receiving such artificial context data can be collected and analyzed. When bids appear to be atypically increased or decreased based on the artificial contextual data, a determination can be made that the bidder retains and/or misuses contextual data, such as user data.

Disclosed are systems and methods for generating recommendations based on online history information and geospatial data. The disclosed systems and methods provide a novel framework for automating the transmission of electronic recommendations to users by consistently monitoring electronic information regarding a user's interests correlating that data with inventory information and geospatial information. The disclosed framework operates by leveraging available user data, user provided objectives, or various real-time data associated with a user, and using the aforementioned data as inputs for recommendation models, in order to generate and transmit the optimal recommendations to the user.

A facility for processing a search query is described. The facility identifies one or more items that satisfy the query, at least one of which is a media sequence. For each identified media sequence, the facility identifies an advertising message based upon the contents of the query. In response to the query, the facility returns a search result that indicates the identified items. When one of the identified media sequences is selected in the search result, the selected media sequence is provided in conjunction with the advertising message identified for it.

One or more computing devices, systems, and/or methods for generating and providing recommendations of products are provided. For example, content is extracted from a message sent to a user. The content is evaluated to identify a product identifier corresponding to a product title of a product. If the product identifier is a truncated version of the product title, then a database of product titles and frequencies of occurrence of the product titles is used to complete the product title. A model is used to infer a product category for the product title. Matching scores are assigned to products within a product category based upon weighted attributes. A recommendation is provided to the user for a product having a matching score greater than a matching threshold.

In an example, a query for a generative artificial intelligence (AI) tool may be received from a client device. A database including query autosuggestion profiles may be accessed to identify a set of query autosuggestion profiles matching the query. A set of query autosuggestions may be generated based upon the query and the set of query autosuggestion profiles. An autosuggestion interface indicative of the set of query autosuggestions may be provided on the client device. In response to receiving a selection of a first query autosuggestion of the set of query autosuggestions via the autosuggestion interface, the generative AI tool may be used to generate a first content item based upon the first query autosuggestion. The first content item may be provided for presentation on the client device.

A method of providing targeted content across user channels includes receiving a plurality of device activity events on a user device, storing the received plurality of device activity events, creating an audience population based on the stored device activity events, determining links for additional secondary user devices based on the user device, creating a cross-channel advertising campaign targeting the determined audience population, and running the cross-channel advertising campaign based on the device activity and the determined additional secondary user devices.

In an example, sets of event information associated with events may be identified. The events may include intentional click events, accidental click events and/or skip events. Accidental click probabilities associated with the accidental click events and/or the skip events may be determined. Machine learning model training may be performed, using the sets of event information associated with the events and labels associated with the events, to generate a first machine learning model. The labels may include second labels associated with the intentional click events and/or third labels associated with the accidental click events and/or the skip events. The second labels may correspond to an intentional click classification. The third labels may be based upon the accidental click probabilities. Click probabilities associated with content items may be determined using the first machine learning model. A content item may be selected for presentation via a client device based upon the click probabilities.