Abstract: A method of training a text-generating model for grammatical error correction (GEC) includes obtaining a multilingual set of text samples where each text sample includes a monolingual textual representation of a respective sentence. The operations also include, for each text sample of the multilingual set of text samples, generating a corrupted synthetic version of the respective text sample where the corrupted synthetic version of the respective text sample includes a grammatical change to the monolingual textual representation of the respective sentence associated with the respective text sample. The operations further include training the text-generating model using a training set of sample pairs. Each sample pair in the training set of sample pairs includes one of the respective text samples of the multilingual set of text samples and the corresponding corrupted synthetic version of the one of the respective text samples of the multilingual set of text samples.

Abstract: Implementations are directed to generating synthetic labeled/preference data by extracting preference pairs from sets of N outputs to a given unlabeled input to a generative model. A plurality of generative outputs are generated by a generative model from a set of input data. A reward model is used to determine a plurality of reward values for the plurality of generative outputs. Based on the reward values, a pair of generative outputs from the plurality of generative outputs is selected for inclusion in a training example. The pair of outputs include a positive training example and a negative training example, where the reward values indicate that the positive training example is preferred over the negative training example. The process can be repeated for a plurality of sets of input data to generate a plurality of training examples for inclusion in a training dataset, which can be used to update reward model(s).

Abstract: A method of training a text-generating model for grammatical error correction (GEC) includes obtaining a multilingual set of text samples where each text sample includes a monolingual textual representation of a respective sentence. The operations also include, for each text sample of the multilingual set of text samples, generating a corrupted synthetic version of the respective text sample where the corrupted synthetic version of the respective text sample includes a grammatical change to the monolingual textual representation of the respective sentence associated with the respective text sample. The operations further include training the text-generating model using a training set of sample pairs. Each sample pair in the training set of sample pairs includes one of the respective text samples of the multilingual set of text samples and the corresponding corrupted synthetic version of the one of the respective text samples of the multilingual set of text samples.

Abstract: Electronic arrangement comprising a data interface for transferring data with external elements, at least one processor for processing instructions and other data, and memory for storing the instructions and other data, said at least one processor being configured, in accordance with the stored instructions, to obtain at least one predictive user model including one or more demographic characteristics as dependent variables to be predicted and usage statistics of applications as explanatory variables, obtain deterministic usage statistics indicative of digital applications a target user has utilized during a monitoring period, and determine, through utilization of the deterministic usage statistics obtained during the monitoring period as input to the at least one established predictive model, an estimate of said one or more of the demographic characteristics of the target user.

Abstract: System for obtaining and analyzing data on online user behavior and other activity having regard to Internet connectable user devices, optionally mobile devices, the system being configured to collect data from a plurality of data sources, wherein the collected data includes at least individual user-level data acquired from user devices and census-level data indicating behavior and demographic characteristics across the entire population of users, active or measured services, with the user-level data being preferably collected using at least one user research panel of controlled constitution, preferably from multiple devices of each participant in the panel, the collected data being indicative of Internet, content, media, application, and/or device usage, organize the obtained user-level and census-level data into a multivalent categorized data set utilizing an ontological metadata schema, determine an electronic deliverable based on an integration of the user-level data and census-level data wherein census-l

Abstract: Electronic arrangement (112) comprising a data interface (210) for transferring data with external elements, at least one processor (202) for processing instructions and other data, and memory (204) for storing the instructions and other data, said at least one processor being configured, in accordance with the stored instructions, to obtain at least one predictive user model (312,314) including one or more demographic characteristics as dependent variables to be predicted and usage statistics of applications as explanatory variables, obtain deterministic usage (316) statistics indicative of digital applications a target user has utilized during a monitoring period, and determine (318), through utilization of the deterministic usage statistics obtained during the monitoring period as input to the at least one established predictive model, an estimate of said one or more of the demographic characteristics of the target user.

Abstract: Electronic system for obtaining data, via one or more digital devices, on user behavior, digital transactions, and exposure relative to digital content and services, or external exposure and associated events between the user and the environment via sensors attached to digital devices, the system being configured to collect data reflecting the content and objects that the user at least potentially perceives as rendered on one or more digital screens attached to smart devices, reconstruct the at least potentially perceived visual landscape based on the collected data, and determine the target and/or level of user attention in view of the reconstruction and associated exposure events detected therein, and to apply locally stored information about rules or fingerprints in the digital object recognition process involving the collected data and validation of the type or identity of user actions, digital content, or external objects, as reflected by the reconstruction recapturing the visual landscape.

Abstract: System for obtaining and analyzing data on online user behavior and other activity having regard to Internet connectable user devices, optionally mobile devices, the system being configured to collect data from a plurality of data sources, wherein the collected data includes at least individual user-level data acquired from user devices and census-level data indicating behavior and demographic characteristics across the entire population of users, active or measured services, with the user-level data being preferably collected using at least one user research panel of controlled constitution, preferably from multiple devices of each participant in the panel, the collected data being indicative of Internet, content, media, application, and/or device usage, organize the obtained user-level and census-level data into a multivalent categorized data set utilizing an ontological metadata schema, determine an electronic deliverable based on an integration of the user-level data and census-level data wherein census-l

Abstract: Electronic system for obtaining data, via one or more digital devices, on user behavior, digital transactions, and exposure relative to digital content and services, or external exposure and associated events between the user and the environment via sensors attached to digital devices, the system being configured to collect data reflecting the content and objects that the user at least potentially perceives as rendered on one or more digital screens attached to smart devices, reconstruct the at least potentially perceived visual landscape based on the collected data, and determine the target and/or level of user attention in view of the reconstruction and associated exposure events detected therein, and to apply locally stored information about rules or fingerprints in the digital object recognition process involving the collected data and validation of the type or identity of user actions, digital content, or external objects, as reflected by the reconstruction recapturing the visual landscape.