Combining Information Extraction, Deductive Reasoning and Machine Learning for Relation Prediction

Suicide remains a leading cause of preventable death worldwide, despite advances in research and decreases in mental health stigma through government health campaigns. Machine learning (ML), a type of artificial intelligence (AI), is the use of algorithms to simulate and imitate human cognition. Given the lack of improvement in clinician-based suicide prediction over time, advancements in technology have allowed for novel approaches to predicting suicide risk. This systematic review and meta-analysis aimed to synthesize current research regarding data sources in ML prediction of suicide risk, incorporating and comparing outcomes between structured data (human interpretable such as psychometric instruments) and unstructured data (only machine interpretable such as electronic health records). Online databases and gray literature were searched for studies relating to ML and suicide risk prediction. There were 31 eligible studies. The outcome for all studies combined was AUC = 0.860, structured data showed AUC = 0.873, and unstructured data was calculated at AUC = 0.866. There was substantial heterogeneity between the studies, the sources of which were unable to be defined. The studies showed good accuracy levels in the prediction of suicide risk behavior overall. Structured data and unstructured data also showed similar outcome accuracy according to meta-analysis, despite different volumes and types of input data.

Linked Open Data (LOD) Cloud is a mesh of open datasets coming from different domains. Among these datasets, a notable amount of datasets belong to the life sciences domain linked together forming an interlinked “Life Sciences Linked Open Data (LSLOD) Cloud”. One of the key challenges for data publishers is to identify and establish links between newly generated domain specific datasets and LSLOD Cloud. While a number of publishing tools exist for creating links from new to existing datasets, tools to detect domain-specific relevant datasets for linking purposes are missing. In this paper, we propose an extended technique for automatically identifying relevant datasets in LSLOD Cloud for inner-ear anatomical and clinical terminologies. We validate the proposed technique with experiments over the publicly accessible LSLOD Cloud using realworld terminologies and datasets provided by clinical organizations.

Existing content management systems (CMSes) usually do not offer flexible, customizable means to create semantic, domain specific indexing and search mechanisms. Therefore, they either do not provide any semantic retrieval, search, browsing functionalities at all on the managed content or the semantic search functionality provided is limited as it depends on the manual annotation of content by users. So, in this study we describe a semantic content management flow by extracting implicit knowledge from both the structure of the CMSes and actual content within them. The task of additional semantic knowledge gathering and providing semantic operations on the content is a challenging task which includes adoption of several latest advancements in information extraction (IE), information retrieval (IR) and Semantic Web areas. In this study, we propose a new approach which provides automatic annotation of content managed in CMSes with the information retrieved from the Linked Open Data (LOD) cloud and several semantic operations on the content in terms of storage and search. We use a simple RDF path language to create custom indexes and retrive semantic knowledge from the LOD cloud suitable for specific use cases. All additional knowledge is materialized along with the actual content of document in dedicated indexes. This semantix indexing infrastructure allows semantically meaningful search facilities on top of it. We realize our approach in the scope of Apache Stanbol project, which is a subproject developed in the scope of IKS project, by focusing on document storage and retrieval. We evaluate our approach in healthcare domain with different domain ontologies (SNOMED/CT, ART, RXNORM) in addition to DBpedia as parts of LOD cloud which are used to annotate documents and content obtained from different health portals.

The linked open data cloud, with a huge volume of data from heterogeneous and interlinked datasets, has turned the Web into a large data store. It has attracted the attention of both developers and researchers in the last few years, opening up new dimensions in machine learning and knowledge discovery. Information extraction procedures for these processes use different approaches, e.g., template-based, federated to multiple sources, fixed depth link traversal, etc. They are limited by problems in online access to datasets' SPARQL endpoints, such as servers being down for maintenance, bandwidth narrowing, limited numbers of access points to datasets in particular time slots, etc., which may result in imprecise and incomplete sets of feature vector generation, affecting the quality of knowledge discovered. The work presented here addresses the disadvantages of online data retrieval by proposing a simple and automatic way to extract features from the linked open data cloud using a linked traversal approach in a local environment with previously identified and known sets of interlinked RDF datasets. The user is given the flexibility to determine the depth of the neighboring properties to be traversed for information extraction to generate the feature vector, which can be used for machine learning and knowledge discovery. The experiment is performed locally with Virtuoso Triple Store for storage of datasets and an interface developed to build the feature vector. The evaluation is performed by comparing the obtained feature vector with gold standard instances annotated manually and with a case study for estimating the effects of demography in movie production for a country. The advantage of the proposed approach lies in overcoming problems with online access of data from the linked data cloud, RDF dataset integration in both local and web environments to build feature vectors for machine learning, and generating background knowledge from the linked data cloud.

Geospatial datasets in the Linking Open Data (LOD) Cloud are rather of static nature and mainly consist of information such as a name, geo coordinates, an address, or opening hours. There is no linked dataset providing dynamic information about the "popularity" of certain places or the "Visiting frequency" of users in specific contextual situations. This type of information within the LOD Cloud, however, would enable a variety of new applications based on semantically enriched location analytics. In this paper, we present Linked Crowdsourced Data as a dataset, which links real user location preferences (e.g., check-ins, ratings, or comments) as well as specific context situations (e.g., weather conditions, holiday information, or measured networks) collected via crowdsourcing to static location data. We showcase the applicability of this dataset for location analytics use cases through a map visualization and highlight its added value with exemplary SPARQL queries that allow for location requests depending on historic context information.

The Linked Open Data (LOD) cloud is a promising reality since the major content producers are offering their data on an open and linked network, through RDF (Resource Description Framework), with the aim of providing Semantic Web applications with a single global database for retrieval of related content and to perform inferences over the network. However, bases with Portuguese-language content are still incipient. In this paper we present the process of inclusion of the InferenceNet – the first resource with common sense and inferentialist knowledge in Portuguese language – on the LOD. Our main goal is to leverage the use and development of Semantic Web applications by content producers in Portuguese language. We develop and evaluated a platform, called SemWidgets, for the creation and execution of widgets able to access and reason over InferenceNet and the open linked data, like DBPedia, Yago, and Article Search API of the New York Times.KeywordsLinguistic ResourcePortuguese LanguageWeb SemanticLinked Open Data

The primary focus of technical communication (TC) in the past decade has been the system-assisted generation and utilization of standardized, structured, and classified content for dynamic output solutions. Nowadays, machine learning (ML) approaches offer a new opportunity to integrate unstructured data into existing knowledge bases without the need to manually organize information into topic-based content enriched with semantic metadata. To make the field of artificial intelligence (AI) more accessible for technical writers and content managers, cloud-based machine learning as a service (MLaaS) solutions provide a starting point for domain-specific ML modelling while unloading the modelling process from extensive coding, data processing and storage demands. Therefore, information architects can focus on information extraction tasks and on prospects to include pre-existing knowledge from other systems into the ML modelling process. In this paper, the capability and performance of a cloud-based ML service, IBM Watson, are analysed to assess their value for semantic context analysis. The ML model is based on a supervised learning method and features deep learning (DL) and natural language processing (NLP) techniques. The subject of the analysis is a corpus of scientific publications on the 2019 Coronavirus disease. The analysis focuses on information extractions regarding preventive measures and effects of the pandemic on healthcare workers.

There is currently a large amount of publicly accessible structured data available as RDF data sets. For example, the Linked Open Data (LOD) cloud now consists of thousands of RDF data sets with over 30 billion triples, and the number and size of the data sets is continuously growing. Many of the data sets in the LOD cloud provide public SPARQL endpoints to allow issuing queries over them. These end-points enable users to retrieve data using precise and highly expressive SPARQL queries. However, in order to do so, the user must have sufficient knowledge about the data sets that she wishes to query, that is, the structure of data, the vocabulary used within the data set, the exact values of literals, their data types, etc. Thus, while SPARQL is powerful, it is not easy to use. An alternative to SPARQL that does not require as much prior knowledge of the data is some form of keyword search over the structured data. Keyword search queries are easy to use, but inherently ambiguous in describing structured queries. This demonstration introduces Sapphire, a system for querying RDF data that strikes a middle ground between ambiguous keyword search and difficult-to-use SPARQL. Our system does not replace either, but utilizes both where they are most effective. Sapphire helps the user construct expressive SPARQL queries that represent her information needs without requiring detailed knowledge about the queried data sets. These queries are then executed over public SPARQL endpoints from the LOD cloud. Sapphire guides the user in the query writing process by showing suggestions of query terms based on the queried data, and by recommending changes to the query based on a predictive user model.

RINX: A system for information and knowledge extraction from resumes

Information Extraction (IE) systems assist analysts to assimilate information from electronic documents. This paper focuses on IE tasks designed to support information discovery applications. Since information discovery implies examining large volumes of heterogeneous documents for situations that cannot be anticipated a priori, they require IE systems to have breadth as well as depth. This implies the need for a domain-independent IE system that can easily be customized for specific domains: end users must be given tools to customize the system on their own. It also implies the need for defining new intermediate level IE tasks that are richer than the subject-verb-object (SVO) triples produced by shallow systems, yet not as complex as the domain-specific scenarios defined by the Message Understanding Conference (MUC). This paper describes InfoXtract, a robust, scalable, intermediate-level IE engine that can be ported to various domains. It describes new IE tasks such as synthesis of entity profiles, and extraction of concept-based general events which represent realistic near-term goals focused on deriving useful, actionable information. Entity profiles consolidate information about a person/organization/location etc. within a document and across documents into a single template; this takes into account aliases and anaphoric references as well as key relationships and events pertaining to that entity. Concept-based events attempt to normalize information such as time expressions (e.g., yesterday) as well as ambiguous location references (e.g., Buffalo). These new tasks facilitate the correlation of output from an IE engine with structured data to enable text mining. InfoXtract's hybrid architecture comprised of grammatical processing and machine learning is described in detail. Benchmarking results for the core engine and applications utilizing the engine are presented.

The volume of data in the world is growing very fast and generated from verity of sources like social media, sensors airline industry or scientific data in different formats. Biggest challenge is how to infer meaningful insights from such a varietyful and big data along with concern of data storage and management of fast growing data. The size of the databases used in today’s enterprises has been growing at exponential rates day by day. Hence, industries requirement to quickly process and analyze the big data volumes for business decision making and customer insights has also grown exponentially. Data pouring from various sources may be can be structured or unstructured in nature. Structured data refers to a relatively well-organized information, which can be further inserted into traditional RDBMS. As Traditional RDBMS are efficient and easy queries by simple, straightforward search algorithms or SQL queries. In contrast to structured data, unstructured data can be considered as information, which does not, comes in a pre-defined data format, well organized data storage model, or cannot be stored well into relational tables. It is assumed to be fastest growing type of data, e.g. image, sensors data, web chats, social networking messaging data, video, documents, log files, and email data. There are many techniques and software available, which can process and provide efficient storage of unstructured data and help organization to perform analytics on unstructured data. Unstructured data does not well-organized and not stored in predefined manner e.g. logs, web chats. The variety and on ordered nature of data makes storage methods and structure makes execution a time and resource-consuming affair. Advancement into technology has open floodgates to push huge volume of unstructured type of data. Multimedia data is one of the example of unstructured big data, which spans all over the Internet. This needs high execution capability to extract useful information. Rapid processing of multimedia data such as video is important for e.g. criminal investigations, surveillance monitoring, news analysis, sports analytics domain, emotion extraction, etc. Hence, analysis of multimedia data in minimum timeframe is one of the latest research areas. Therefore, we have researched techniques for analyzing unstructured data to extract meaningful information hidden in the big data. In addition, we will describe about various techniques and software used to Manage, process unstructured big data in efficient manner, and increases the performance of complexity analysis.

SummaryBackground:Clinical Data Warehouses (CDW) reuse Electronic health records (EHR) to make their data retrievable for research purposes or patient recruitment for clinical trials. However, much information are hidden in unstructured data like discharge letters. They can be preprocessed and converted to structured data via information extraction (IE), which is unfortunately a laborious task and therefore usually not available for most of the text data in CDW.Objectives:The goal of our work is to provide an ad hoc IE service that allows users to query text data ad hoc in a manner similar to querying structured data in a CDW. While search engines just return text snippets, our systems also returns frequencies (e.g. how many patients exist with “heart failure” including textual synonyms or how many patients have an LVEF < 45) based on the content of discharge letters or textual reports for special investigations like heart echo. Three subtasks are addressed: (1) To recognize and to exclude negations and their scopes, (2) to extract concepts, i.e. Boolean values and (3) to extract numerical values.Methods:We implemented an extended version of the NegEx-algorithm for German texts that detects negations and determines their scope. Furthermore, our document oriented CDW PaDaWaN was extended with query functions, e.g. context sensitive queries and regex queries, and an extraction mode for computing the frequencies for Boolean and numerical values.Results:Evaluations in chest X-ray reports and in discharge letters showed high F1-scores for the three subtasks: Detection of negated concepts in chest X-ray reports with an F1-score of 0.99 and in discharge letters with 0.97; of Boolean values in chest X-ray reports about 0.99, and of numerical values in chest X-ray reports and discharge letters also around 0.99 with the exception of the concept age.Discussion:The advantages of an ad hoc IE over a standard IE are the low development effort (just entering the concept with its variants), the promptness of the results and the adaptability by the user to his or her particular question. Disadvantage are usually lower accuracy and confidence.This ad hoc information extraction approach is novel and exceeds existing systems: Roogle [1] extracts predefined concepts from texts at preprocessing and makes them retrievable at runtime. Dr. Warehouse [2] applies negation detection and indexes the produced subtexts which include affirmed findings. Our approach combines negation detection and the extraction of concepts. But the extraction does not take place during preprocessing, but at runtime. That provides an ad hoc, dynamic, interactive and adjustable information extraction of random concepts and even their values on the fly at runtime.Conclusions:We developed an ad hoc information extraction query feature for Boolean and numerical values within a CDW with high recall and precision based on a pipeline that detects and removes negations and their scope in clinical texts.

The Web and its Semantic extension (i.e. Linked Open Data) contain open global-scale knowledge and make it available to potentially intelligent machines that want to benefit from it. Nevertheless, most of Linked Open Data lack ontological distinctions and have sparse axiomatisation. For example, distinctions such as whether an entity is inherently a class or an individual, or whether it is a physical object or not, are hardly expressed in the data, although they have been largely studied and formalised by foundational ontologies (e.g. DOLCE, SUMO). These distinctions belong to common sense too, which is relevant for many artificial intelligence tasks such as natural language understanding, scene recognition, and the like. There is a gap between foundational ontologies, that often formalise or are inspired by pre-existing philosophical theories and are developed with a top-down approach, and Linked Open Data that mostly derive from existing databases or crowd-based effort (e.g. DBpedia, Wikidata). We investigate whether machines can learn foundational distinctions over Linked Open Data entities, and if they match common sense. We want to answer questions such as “does the DBpedia entity for dog refer to a class or to an instance?”. We report on a set of experiments based on machine learning and crowdsourcing that show promising results.

This study aims to develop high-performing Machine Learning and Deep Learning models in predicting hospital length of stay (LOS) while enhancing interpretability. We compare performance and interpretability of models trained only on structured tabular data with models trained only on unstructured clinical text data, and on mixed data. The structured data was used to train fourteen classical Machine Learning models including advanced ensemble trees, neural networks and k-nearest neighbors. The unstructured data was used to fine-tune a pre-trained Bio Clinical BERT Transformer Deep Learning model. The structured and unstructured data were then merged into a tabular dataset after vectorization of the clinical text and a dimensional reduction through Latent Dirichlet Allocation. The study used the free and publicly available Medical Information Mart for Intensive Care (MIMIC) III database, on the open AutoML Library AutoGluon. Performance is evaluated with respect to two types of random classifiers, used as baselines. The best model from structured data demonstrates high performance (ROC AUC = 0.944, PRC AUC = 0.655) with limited interpretability, where the most important predictors of prolonged LOS are the level of blood urea nitrogen and of platelets. The Transformer model displays a good but lower performance (ROC AUC = 0.842, PRC AUC = 0.375) with a richer array of interpretability by providing more specific in-hospital factors including procedures, conditions, and medical history. The best model trained on mixed data satisfies both a high level of performance (ROC AUC = 0.963, PRC AUC = 0.746) and a much larger scope in interpretability including pathologies of the intestine, the colon, and the blood; infectious diseases, respiratory problems, procedures involving sedation and intubation, and vascular surgery. Our results outperform most of the state-of-the-art models in LOS prediction both in terms of performance and of interpretability. Data fusion between structured and unstructured text data may significantly improve performance and interpretability.

The traditional means of extracting information from the Web are keyword-based search and browsing. The Semantic Web adds structured information (i.e., semantic annotations and references) supporting both activities. One of the most interesting recent developments is Linked Open Data (LOD), where information is presented in the form of facts – often originating from published domain-specific databases – that can be accessed both by a human and a machine via specific query endpoints. In this article, we argue that machine learning provides a new way to query web data, in particular LOD, by analyzing and exploiting statistical regularities. We discuss challenges when applying machine learning to the Web and discuss the particular learning approaches we have been pursuing in THESEUS. We discuss a number of applications where the Web is queried via machine learning and describe several extensions to our approaches.