Abstract: A computerized method of cleaning a data set, configured for training of a machine learning model, includes data instances, each including a label. It comprises: computing coreset(s) from the data set, generating an importance measure for each coreset sample, indicative of an importance/weight value; selecting, from the coresets, a coreset(s) (a representative sub-set) that is representative of the data set; and selecting a number of samples, based on importance-related criteria, indicative of corresponding importance measures of the selected coreset(s) having high importance/low weight values. The set of selected samples has high probability of high importance/low weight values, compared to a probability associated with selecting data instances from the data set. The set facilitates determining whether each sample requires an action. The action gives an increased-quality representative sub-set. This facilitates quicker/higher-accuracy training of the model.

Abstract: A computerized method of cleaning a data set comprises: (a) providing the data set, configured for training of a machine learning model, and comprising data instances. Each instance comprises a label; (b) selecting a number of data instances from the set based on importance-related criteria, indicative of corresponding high importance values associated with the data instances. This generates high importance set(s), giving rise to a set of selected data instances having a high probability of high importance values, as compared to a second probability associated with selecting data instances from the data set. The selected set facilitates determining, for each data instance of, whether it requires an action. This facilitates performing the action, which brings about increased quality data set(s), which facilitate quicker and/or higher-accuracy training of the machine learning model, as compared to a second training of the machine learning model performed utilizing the data set.

Abstract: A method of machine learning model training comprises: (a) providing a data set, configured for model training, comprising data sub-sets, each comprising data instances; (b) providing a coreset tree, computed from the data set, comprising coresets, wherein each leaf coreset based on a corresponding data sub-set; (c) removing data instance(s) from data sub-set(s), generating reduced sub-set(s); (d) recomputing coreset(s) that is an ancestor of the reduced sub-set(s), to reflect removal of instance(s), giving a modified coreset tree; (e) selecting, from the modified tree, coresets that are representative of the data set exclusive of the removed instance(s); (f) training the model utilizing the selected coresets, giving a trained model; (g) performing prediction(s) for the removed instance(s), utilizing the model; (g) evaluating model predictability metric(s), associated with the prediction(s), indicative of the model's ability to correctly predict the removed instance(s).

Abstract: A method for allowing enhanced interaction between an IFrame or a web page of a website that corresponds to a domain and an embedded IFrame from a different domain. Accordingly, two IFrames from the same domain that are placed on the same page are allowed to run a JavaScript code one on the other. The website is allowed to place a site tag, from a domain different than the web page's domain that provided to the website by a trusted third party on the top page that is not being nested within an IFrame. The site tag that is delivered on the top page is allowed to perform actions on behalf of an ad tag from the same domain that is nested within the IFrames.

Abstract: Method for automatically monitoring and verifying advertising content during a campaign, delivered over a data network. Accordingly, advertisers submit a list of sites, on which the advertising content should be placed according to desired insertion order. Mapping crawlers visit these sites and locate pages with advertisements that belong to required sections, pages that do not belong to the required sections or pages with high probability for incidents. A list of pages to visit per every site is generated and autonomous or plug-in visual crawlers are allowed to visit the list of pages, according to predetermined site visiting plan. A crawlers' manager allocates the pages between visual crawlers, for obtaining adequate incident coverage and load on the visual crawlers. An incident identifier compares the insertion orders with the delivery data and whenever an insertion order and its corresponding delivery data do not match, an incident report is generated.

Abstract: Method for automatically monitoring and verifying advertising content during a campaign, delivered over a data network. Accordingly, advertisers submit a list of sites, on which the advertising content should be placed according to desired insertion order. Mapping crawlers visit these sites and locate pages with advertisements that belong to required sections, pages that do not belong to the required sections or pages with high probability for incidents. A list of pages to visit per every site is generated and autonomous or plug-in visual crawlers are allowed to visit the list: of pages, according to predetermined site visiting plan. A crawlers' manager allocates the pages between visual crawlers, for obtaining adequate incident coverage and load on the visual crawlers. An incident identifier compares the insertion orders with the delivery data and whenever an insertion order and its corresponding delivery data do not match, an incident report is generated.

Abstract: Method for automatically monitoring and verifying advertising content during a campaign, delivered over a data network. Accordingly, one or more advertisers submit, via a user interface, a list (that may be generated manually or by the mapping crawlers) of sites or of sections per site, on which the advertising content should be placed according to a desired insertion order (the insertion order information may modified at any time point). In addition, one or more mapping crawlers are activated to visit these sites and locate pages with advertisements that belong to required sections, pages that do not belong to the required sections or pages with high probability for incidents. A list of pages to visit per every site is generated and autonomous or Plug-in visual crawlers are allowed to visit the list of pages, according to a predetermined site visiting plan. A crawlers' manager allocates the pages between visual crawlers, for obtaining required adequate incident coverage and load on the visual crawlers.

Abstract: A real-time system for verifying advertisements against their campaign rules prior to serving them by an ad server that includes: (a) a decision server for deciding whether or not an advertisement intended to be presented to a user matches its campaign rules; (b) a database, which receives a connection from the decision server, the database comprises campaign rules according to publisher's intents; (c) a Web server comprising query SW modules, the server delivers the query SW modules to the user's browser; and (d) a query SW module executed by the user's browser and adapted to extract parameters that are necessary for the decision server; wherein the parameters are delivered to the decision server which analyzes the campaign rules and the parameters provided, and sends its decision to the ad sever which serves or block the advertisement accordingly.

Abstract: Method for automatically monitoring and verifying advertising content during a campaign, delivered over a data network. Accordingly, one or more advertisers submit, via a user interface, a list (that may be generated manually or by the mapping crawlers) of sites or of sections per site, on which the advertising content should be placed according to a desired insertion order (the insertion order information may modified at any time point). In addition, one or more mapping crawlers are activated to visit these sites and locate pages with advertisements that belong to required sections, pages that do not belong to the required sections or pages with high probability for incidents. A list of pages to visit per every site is generated and autonomous or Plug-in visual crawlers are allowed to visit the list of pages, according to a predetermined site visiting plan. A crawlers' manager allocates the pages between visual crawlers, for obtaining required adequate incident coverage and load on the visual crawlers.