A multicore apparatus determines fitness of a candidate for a role. The apparatus includes a multicore system processing device, a plurality of parallel multicore graphics processing devices, a network interface device, a storage device, and a system interface bus. The network interface device provides remote connection to the multicore system processing device. The storage device stores training data including positive and negative examples. The positive examples represent candidates who would be invited to an interview, and the negative examples represent candidates who would not be invited to an interview. The positive and negative examples are used by the plurality of parallel multicore graphics processing devices to train a deep learning model, which is used by the multicore system processing device to determine fitness of the candidate for the role while eliminating unwanted bias.

At an artificial intelligence system, a respective feature set is generated from individual text collections pertaining to an item, using a first machine learning model which is trained to perform character-level analysis. Using at least a portion of a second machine learning model, a score associated with a semantic criterion is generated for an item; the training input to the second model is based on the feature sets. A recommendation associated with the item is generated based on the score.

At an artificial intelligence system, a respective feature set is generated from individual text collections pertaining to an item, using a first machine learning model which is trained to perform character-level analysis. Using at least a portion of a second machine learning model, a score associated with a semantic criterion is generated for an item; the training input to the second model is based on the feature sets. A recommendation associated with the item is generated based on the score.

Provided is a method of constructing a neural network translation model. The method includes generating a first neural network translation model learning a feature of source domain data used in an unspecific field, generating a second neural network translation model learning a feature of target domain data used in a specific field, generating a third neural network translation model learning a common feature of the source domain data and the target domain data; and generating a combiner combining translation results of the first to third neural network translation models.

Provided is a method of constructing a neural network translation model. The method includes generating a first neural network translation model learning a feature of source domain data used in an unspecific field, generating a second neural network translation model learning a feature of target domain data used in a specific field, generating a third neural network translation model learning a common feature of the source domain data and the target domain data; and generating a combiner combining translation results of the first to third neural network translation models.

A method for generating a dual-class dataset is disclosed. A single-class dataset and a context dataset are obtained. The context dataset can be labeled. A model can be trained using the combination of the single-class dataset and the labeled context dataset. The model can be run on the context dataset. The data points that are classified the same as the data points included in the single-class dataset, can be removed from the labeled context dataset and added to the single-class dataset. These steps can be repeated until no data points are classified by the model.

A method for generating a dual-class dataset is disclosed. A single-class dataset and a context dataset are obtained. The context dataset can be labeled. A model can be trained using the combination of the single-class dataset and the labeled context dataset. The model can be run on the context dataset. The data points that are classified the same as the data points included in the single-class dataset, can be removed from the labeled context dataset and added to the single-class dataset. These steps can be repeated until no data points are classified by the model.

An annotation corresponding to a first range of text of a first document may be received. Based on the annotation, comparisons may be performed between a text string that comprises the first range of text and a group of text of a second document at different positions in the group of text. Based on the comparisons, similarity scores between the text string and the group of text may be determined at the different positions in the group of text. A position for the annotation in the group of text may be selected based on the similarity scores at the different positions. The annotation may be associated with a second range of text in the group of text that corresponds to the position.

A system and a method are provided for displaying message history while composing a message. The method includes displaying a message composition application comprising a first recipient field, a message history display area, and a message composition field; detecting an input into the first recipient field for specifying a recipient; and upon detecting that the recipient has been specified: displaying a second recipient field; navigating application focus from the first recipient field to the second recipient field; and displaying a message history associated with the recipient in the message history display area.

A system and a method are provided for displaying message history while composing a message. The method includes displaying a message composition application comprising a first recipient field, a message history display area, and a message composition field; detecting an input into the first recipient field for specifying a recipient; and upon detecting that the recipient has been specified: displaying a second recipient field; navigating application focus from the first recipient field to the second recipient field; and displaying a message history associated with the recipient in the message history display area.