International Business Machines Corporation

Other

Robert Christian Sizemore

Sterling Richardson Smith

David Blake Werts

Kristin E McNeil

David Contreras

Corville O Allen

Roberto Delima

Brendan Bull

Aysu Ezen Can

Keith P Biegert

Erich Shannon Magee

James Edward Fox

David Gerald Herbeck

Mario J Lorenzo

Richard T Brandle

Jordan T Liggitt

Don L Goodliffe

Donald E Keith

Randy A Robinette

Garry J Smithwick

Anthony J Zappavigna

Scott Robert Carrier

Robert Charles Leah

Jennifer Lynn La Rocca

Richard A Salmon

Heng Chu

David Gerard Herbeck

Daniel Reed Drake

John A Scott

Jennifer L La Rocca

A mechanism is provided in a data processing system to implement an annotator for annotating content using context-based surface forms. The mechanism receives a dictionary data structure of surface forms comprising a plurality of regular expressions and input content. The mechanism compares a given span of text in the input content to each regular expression in the dictionary data structure. Responsive to the given span of text matching a given regular expression, an annotator annotates the span of text with a content indicator corresponding to a content category associated with the dictionary data structure. The mechanism performs a natural language processing operation on the input content based on results of the annotation.

Context based surface form generation for cognitive system dictionaries

Aspects of the present disclosure provide a computer-implemented method that includes obtaining the medical condition of the individual based at least in part on a personal identifier associated with the individual, the medical condition being determined by performing natural language processing of medical information associated with the individual. The method further includes obtaining characteristics of the plurality of items. The method further includes comparing the characteristics of the plurality of items to a restriction associated with the medical condition of the individual. The method further includes, responsive to determining that at least one of the characteristics of the plurality of items violates the restriction associated with the medical condition of the individual, preventing the individual from obtaining the item of the plurality of items associated with the violation.

Filtering items based on a medical condition of an individual

Mechanisms are provided to implement a bias identification engine that identifies bias in the operation of a trained cognitive computing system. A bias risk annotator is configured to identify a plurality of bias triggers in inputs and outputs of the trained cognitive computing system based on a bias risk trigger data structure that specifies terms or phrases that are associated with a bias. An annotated input and an annotated output of the trained cognitive computing system is received and processed by the bias risk annotator to determine if they comprise a portion of content that contains a bias trigger. In response to at least one of the annotated input or annotated output comprising a portion of content containing a bias trigger a notification is transmitted, to an administrator computing device, that specifies the presence of bias in the operation of the trained cognitive computing system.

Bias identification in cognitive computing systems

A mechanism is provided in a data processing system comprising least one processor and at least one memory, the at least one memory comprising instructions executed by the at least one processor to cause the at least one processor to implement a clinical decision support system. The mechanism receives a plurality of patient electronic medical records (EMRs) for a patient from a plurality of different sources. For a portion of a patient EMR record of the plurality of patient EMRs, the mechanism detects entities and analyzes a document structure of the portion of the patient EMR to identify a hierarchical structure of the portion of the patient EMR. The mechanism generates a container representation of the portion of the patient EMR based on the hierarchical structure. The mechanism placing each of the one or more sentences within the container representation based on relative position within the hierarchical structure. The mechanism generates a knowledge graph using the detected entities and the container representation.

Container-based knowledge graphs for determining entity relations in non-narrative text

Mechanisms are provided to perform selective deep parsing of natural language content. A targeted deep parse natural language processing system is configured to recognize one or more triggers that specify elements within natural language content that indicate a portion of natural language content that is to be targeted with a deep parse operation. A portion of natural language content is received and a pre-deep parse scan operation is performed on the natural language content based on the one or more triggers to identify one or more sub-portions of the natural language content that contain at least one of the one or more triggers. A deep parse is performed on only the one or more sub-portions of the portion of natural language content that contain at least one of the one or more triggers, while other sub-portions of the portion of natural language content are not deep parsed.

Selective deep parsing of natural language content

Embodiments include methods, systems and computer program products for performing named entity recognition. Aspects include obtaining a text having a plurality of words and comparing each of the plurality of words to a dictionary. Aspects also include creating, based on the comparison, an annotation for at least one of the plurality of words that the least one of the plurality of words refers to a named entity. Aspects further include parsing the text to identify a part of speech for each of the plurality of words and removing the annotations from each of the at least one of the plurality of words that has a part of speech that is not one or a noun and a noun supporting adjective.

Named entity recognition

A mechanism is provided to implement a cognitive dictionary builder. The mechanism configures the cognitive dictionary builder with a set of selection criteria comprising a set of rules. The mechanism performs natural language processing on an input document in a corpus of information to analyze a context for each term or phrase in the input document and applies the set of rules to each term or phrase in the input document with respect to its context. The mechanism adds a term or phrase to at least one corresponding dictionary data structure based on a result of applying the set of rules.

Dynamic creation/expansion of cognitive model dictionaries based on analysis of natural language content

A computer-implemented method includes receiving, by a natural language processing (NLP) annotator, an input text that is to be annotated. The method further includes determining, by the NLP annotator, a user setting that indicates an aggressiveness level of annotation to be used to annotate the input text. The method further includes selecting, by the NLP annotator, from a plurality of dictionaries, a first dictionary based at least in part on the aggressiveness level. The method further includes generating, by the NLP annotator, annotated text of the input text based at least in part on the first dictionary. The method further includes outputting, by the NLP annotator, the annotated text.

Configuring metrics and recall levels for natural language processing annotator

A mechanism is provided for re-ranking search results based on a personal medical condition. One or more medical conditions associated with a patient are identified based on an analysis of patient information associated with the patient. The one or more medical conditions are correlated with a set of medical condition content indicator data structures. A search query corresponding to the patient is processed to generate initial search results that are ranked with an initial ranking. The initial search results are analyzed based on a patient specific dictionary data structure corresponding to the one or more medical conditions associated with the patient. The content of the initial search results are re-ranked to generate re-ranked search results having a modified ranking that is based on the one or more medical conditions of the patient. The re-ranked search results are output to the patient in accordance with the modified ranking.

Search results ranking based on a personal medical condition

A medical condition based content filter mechanism is provided that analyzes patient information associated with a patient to identify medical condition(s) associated with the patient and correlating the medical condition(s) with one or more medical condition content indicator data structures specifying negative content indicators and/or positive content indicators. A user specific content indicator dictionary data structure (USCID) is generated based on the correlation and used to process received content to filter out portions of the content matching the negative content indicators or present portions of the content matching the positive content indicators, thereby generating modified content that is output to the patient via a content access application executing on a computing device associated with the patient.

Automated content avoidance based on medical conditions

A mechanism is provided for implementing an emergency response cognitive computing system. The emergency response cognitive computing system identifies a first party about which a communication is being performed between a caller and the emergency response system. The emergency response cognitive computing system generates, in real-time, a transcript of the communication as the communication is being conducted. The emergency response cognitive computing system performs natural language processing on the transcript of the communication to identify portions of content corresponding to medical concepts. The emergency response cognitive computing system processes patient information to identify elements of the patient information referencing concepts corresponding to medical concepts. The emergency response cognitive computing system outputs the elements of the patient information to a second party involved in the communication.

Automatic identification of medical information pertinent to a natural language conversation

A method for de-identifying protected health information (PHI) associated with electronic medical records (EMRs) based on a common analysis structure (CAS) is provided. The method may include detecting a system event associated with a system comprising the EMRs. The method may further include in response to detecting the system event, detecting a first CAS associated with the EMRs. The method may further include extracting first CAS data associated with the first CAS, wherein the first CAS data comprises unstructured data associated with the EMRs and normalized annotations based on CAS objects that are associated with the unstructured data. The method may further include obfuscating the unstructured data associated with the first CAS. The method may also include generating a second CAS comprising the obfuscated unstructured data and a copied version of the normalized annotations, wherein the copied version of normalized annotations are correlated with the obfuscated unstructured data.

Growing community of inventors

Fuquay-Varina, NC, United States of America

Robert Christian Sizemore