Abstract: Aspects of the present invention provide devices that subtitle streaming video with audio and identify a speaker in a streaming video with audio according to words spoken by the speaker matched to a cognitive print. The cognitive print includes traits classified according a hierarchical long short term model (LSTM). The hierarchical LSTM includes layers of LSTMs and each layer corresponds to the classification of one trait. A processor annotates a subtitle of the words spoken by the speaker, which decorates the subtitle with a label representative of the identified speaker, and streams the decorated subtitle with the streaming video with audio.

Abstract: Aspects of the present invention provide devices that subtitle streaming video with audio and identify a speaker in a streaming video with audio according to words spoken by the speaker matched to a cognitive print. The cognitive print includes traits classified according a hierarchical long short term model (LSTM). The hierarchical LSTM includes layers of LSTMs and each layer corresponds to the classification of one trait. A processor annotates a subtitle of the words spoken by the speaker, which decorates the subtitle with a label representative of the identified speaker, and streams the decorated subtitle with the streaming video with audio.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: collecting various word data from cross-domain sources and subject websites; assessing relevancy of feature vectors from external domains, live content of subject websites, and secondary terms derived from the live contents; expanding a language model for a domain by relevance passing a logistic regression threshold.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: determining that one or more word of a feature vector more supports than negates a language model corresponding to the domain based on a sensitivity of respective word. Words having acceptable sensitivities are added to the language model, and the language model is enhanced by use of machine learning in order to accurately and comprehensively model the language specific for the domain.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: collecting various word data from cross-domain sources and subject websites; assessing relevancy of feature vectors from external domains, live content of subject websites, and secondary terms derived from the live contents; expanding a language model for a domain by relevance passing a logistic regression threshold.

Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: determining that one or more word of a feature vector more supports than negates a language model corresponding to the domain based on a sensitivity of respective word. Words having acceptable sensitivities are added to the language model, and the language model is enhanced by use of machine learning in order to accurately and comprehensively model the language specific for the domain.