Abstract: Local abbreviation expansion is provided through context correlation. In various embodiments, an abbreviation within a phrase is identified. The abbreviation is surrounded by a plurality of words. The words surrounding the abbreviation are provided to a trained neural network. The neural network includes a projection layer adapted to map inputs of the neural network onto a continuous vector space. An expansion is received from the trained neural network. The expansion corresponds to the abbreviation based on the surrounding plurality of words.

Abstract: Methods and systems of summarizing medical data. One system includes an electronic processor configured to analyze medical data to extract a medical concept and a plurality of additional attributes of the medical concept and store the medical concept and the plurality of additional attributes. The electronic processor is configured to generate a first medical summary associated with the patient, where the first medical summary is based on the stored medical concept and at least a first additional attribute included in the stored plurality of additional attributes. The electronic processor is configured to receive a user interaction with the first medical summary. The electronic processor is configured to generate a second medical summary associated with the patient based on the user interaction, the second medical summary is based on the stored medical concept and at least a second additional attribute included in the stored plurality of additional attributes.

Abstract: A system configured to predict fine-grained affective states. The system comprising a processor configured to execute instructions to create training data comprising content conveying emotions, and to create a trained model by performing an emotion vector space model training process using the training data to train a model using a feed forward neural network that converts discrete emotions into emotion vector representations. The system uses the trained model to predict fine-grained affective states for text conveying an emotion.

Abstract: Negation scope analysis for negation detection is provided. In various embodiments, a phrase is read from a report collection. The phrase is searched for at least one of a predetermined set of negation keywords. A dependency parse tree is generated of the phrase. The dependency parse tree is traversed starting with the at least one of the predetermined set of negation keywords. Based on the traversal, a plurality of words of the phrase are determined that are spanned by the at least one of the predetermined set of negation keywords.

Abstract: A computer-implemented method for fine-grained affective states prediction. The computer-implemented method creates training data comprising content conveying emotions. The method creates a trained model by performing an emotion vector space model training process using the training data to train a model using a feed forward neural network that converts discrete emotions into emotion vector representations. The trained model can be used to predict fine-grained affective states for text conveying an emotion.

Abstract: Semi-automatic annotation of medical images is provided. In various embodiments, a classifier is applied to each of a first plurality of medical images to generate a label and an associated confidence value for each of the first plurality of medical images. The classifier is pre-trained using a manually labeled set of medical images. Those of the first plurality of medical images having an associated confidence value below a predetermined threshold are selected. The selected medical images are provided to a user. Updated labels are received from the user for the selected medical images. The classifier is retrained using the first plurality of medical images, with the updated labels for the selected medical images and the generated labels for medical images not selected.

Abstract: Negation scope analysis for negation detection is provided. In various embodiments, a phrase is read from a report collection. The phrase is searched for at least one of a predetermined set of negation keywords. A dependency parse tree is generated of the phrase. The dependency parse tree is traversed starting with the at least one of the predetermined set of negation keywords. Based on the traversal, a plurality of words of the phrase are determined that are spanned by the at least one of the predetermined set of negation keywords.

Abstract: Mechanisms are provided to implement a natural language request processing engine (NLRPE). The NRLPE performs natural language processing on a portion of unstructured text in an electronic data structure to generate textual characteristics of the portion of unstructured text. The NRLPE annotates at least one phrase in the portion of unstructured text at least by linking the at least one phrase to one or more concepts specified in at least one ontological data structure based on the textual characteristics of the portion of unstructured text. The NRLPE generates a model of the portion of unstructured text based on the one or more concepts linked to the at least one phrase. The NRLPE processes a request for information specifying a concept of interest based on the model of the portion of unstructured text by retrieving the at least one phrase or the at least one merged phrase as a response.

Abstract: A mechanism is provided that implements a cognitive data processing system for automatically processing ambiguously labeled data associated with a medical image. The cognitive data processing system receives an ambiguously labeled set of training data in which the ambiguously labeled set of training data comprises portions of data and associated labels, and wherein at least one portion of data in the ambiguously labeled set of training data has a plurality of different labels that together render the portion of data ambiguously labeled. The cognitive data processing system configures an implementation of a model that comprises a loss term, a maximizing term, and a sparsity term. The cognitive data processing system processes the ambiguously labeled set of training data based on the model to identifying a mapping that minimizes a loss function and thereby train the cognitive data processing system.

Abstract: A mechanism is provided that implements a cognitive data processing system for automatically processing ambiguously labeled data associated with a medical image. The cognitive data processing system receives an ambiguously labeled set of training data in which the ambiguously labeled set of training data comprises portions of data and associated labels, and wherein at least one portion of data in the ambiguously labeled set of training data has a plurality of different labels that together render the portion of data ambiguously labeled. The cognitive data processing system configures an implementation of a model that comprises a loss term, a maximizing term, and a sparsity term. The cognitive data processing system processes the ambiguously labeled set of training data based on the model to identifying a mapping that minimizes a loss function and thereby train the cognitive data processing system.

Abstract: Mechanisms are provided to implement a natural language request processing engine (NLRPE). The NRLPE performs natural language processing on a portion of unstructured text in an electronic data structure to generate textual characteristics of the portion of unstructured text. The NRLPE annotates at least one phrase in the portion of unstructured text at least by linking the at least one phrase to one or more concepts specified in at least one ontological data structure based on the textual characteristics of the portion of unstructured text. The NRLPE generates a model of the portion of unstructured text based on the one or more concepts linked to the at least one phrase. The NRLPE processes a request for information specifying a concept of interest based on the model of the portion of unstructured text by retrieving the at least one phrase or the at least one merged phrase as a response.

Abstract: Semi-automatic annotation of medical images is provided. In various embodiments, a classifier is applied to each of a first plurality of medical images to generate a label and an associated confidence value for each of the first plurality of medical images. The classifier is pre-trained using a manually labeled set of medical images. Those of the first plurality of medical images having an associated confidence value below a predetermined threshold are selected. The selected medical images are provided to a user. Updated labels are received from the user for the selected medical images. The classifier is retrained using the first plurality of medical images, with the updated labels for the selected medical images and the generated labels for medical images not selected.

Abstract: Local abbreviation expansion is provided through context correlation. In various embodiments, an abbreviation within a phrase is identified. The abbreviation is surrounded by a plurality of words. The words surrounding the abbreviation are provided to a trained neural network. The neural network includes a projection layer adapted to map inputs of the neural network onto a continuous vector space. An expansion is received from the trained neural network. The expansion corresponds to the abbreviation based on the surrounding plurality of words.

Abstract: Negation scope analysis for negation detection is provided. In various embodiments, a phrase is read from a report collection. The phrase is searched for at least one of a predetermined set of negation keywords. A dependency parse tree is generated of the phrase. The dependency parse tree is traversed starting with the at least one of the predetermined set of negation keywords. Based on the traversal, a plurality of words of the phrase are determined that are spanned by the at least one of the predetermined set of negation keywords.

Abstract: A cross-modality neural network transform for semi-automatic medical image annotation is provided. In various embodiments, an input medical image is mapped to a first vector in a text vector space. The first vector corresponds to the features of the medical image. A set of predetermined vectors is searched for a closest one of the predetermined vectors to the first vector. From the closest one of the predetermined vectors, one or more keywords is determined describing the input medical image.

Abstract: Automatic detection of disease presence from combining disease-specific measurements with textual descriptions of disease and its severity in unstructured textual reports is provided. In various embodiments, a knowledge graph of clinical concepts is read. Based on the knowledge graph, a plurality of associations are determined between disease names, symptoms, anatomical abnormalities, and qualifiers. A corpus of clinical reports is read. Based on the plurality of associations, a plurality of portions indicative of a disease condition are located within the corpus of clinical reports. Within each of the plurality of portions, name/value pairs are detected corresponding to measurements indicative of the disease condition. The measurements indicative of the disease condition are extracted.

Abstract: Cohort detection from multimodal data by machine learning is provided. In various embodiments, a plurality of patient records associated with a patient are read from a plurality of data sources. A plurality of disease-specific features are extracted from the plurality of patient records. The plurality of disease-specific features are provided to a classifier. An indicator of a likely disease condition of the patient is received from the classifier.