Human emotion is expressed in many communication modalities and media formats and so their computational study is equally diversified into natural language processing, speech processing, computer vision, etc. Similarly, a large variety of representation formats has been used in previous research to describe emotions (polarity scales, basic emotion categories, dimensional approaches, appraisal theory, etc.) and has led to an ever proliferating diversity of datasets, predictive models, and software tools for emotion analysis. Because of these two distinct types of heterogeneity, at the expressional and representational level, there is a dire need to unify previous work on increasingly diverging data and label types. This article presents such a unifying computational model. We propose a training procedure that learns a shared latent representation for emotions, so-called emotion embeddings, independent of different natural languages, communication modalities, media or representation label formats, and even disparate model architectures. Experiments on a wide range of heterogeneous affective datasets indicate that this approach yields the desired interoperability for the sake of reusability, interpretability and flexibility, without penalizing prediction quality. Code and data are archived under DOI:10.5281/zenodo.7405327.

Despite remarkable advances in the development of language resources over the recent years, there is still a shortage of annotated, publicly available corpora covering (German) medical language. With the initial release of the German Guideline Program in Oncology NLP Corpus (GGPONC), we have demonstrated how such corpora can be built upon clinical guidelines, a widely available resource in many natural languages with a reasonable coverage of medical terminology. In this work, we describe a major new release for GGPONC. The corpus has been substantially extended in size and re-annotated with a new annotation scheme based on SNOMED CT top level hierarchies, reaching high inter-annotator agreement (γ=.94). Moreover, we annotated elliptical coordinated noun phrases and their resolutions, a common language phenomenon in (not only German) scientific documents. We also trained BERT-based named entity recognition models on this new data set, which achieve high performance on short, coarse-grained entity spans (F1=.89), while the rate of boundary errors increases for long entity spans. GGPONC is freely available through a data use agreement. The trained named entity recognition models, as well as the detailed annotation guide, are also made publicly available.

The exploding amount of user-generated content has spurred NLP research to deal with documents from various digital communication formats (tweets, chats, emails, etc.). Using these texts as language resources implies complying with legal data privacy regulations. To protect the personal data of individuals and preclude their identification, we employ pseudonymization. More precisely, we identify those text spans that carry information revealing an individual’s identity (e.g., names of persons, locations, phone numbers, or dates) and subsequently substitute them with synthetically generated surrogates. Based on CodE Alltag, a German-language email corpus, we address two tasks. The first task is to evaluate various architectures for the automatic recognition of privacy-sensitive entities in raw data. The second task examines the applicability of pseudonymized data as training data for such systems since models learned on original data cannot be published for reasons of privacy protection. As outputs of both tasks, we, first, generate a new pseudonymized version of CodE Alltag compliant with the legal requirements of the General Data Protection Regulation (GDPR). Second, we make accessible a tagger for recognizing privacy-sensitive information in German emails and similar text genres, which is trained on already pseudonymized data.

To track different levels of formality in written discourse, we introduce a novel type of lexicon for the German language, with entries ordered by their degree of (in)formality.We start with a set of words extracted from traditional lexicographic resources, extend it by sentence-based similarity computations, and let crowdworkers assess the enlarged set of lexical items on a continuous informal-formal scale as a gold standard for evaluation.We submit this lexicon to an intrinsic evaluation related to the best regression models and their effect on predicting formality scores and complement our investigation by an extrinsic evaluation of formality on a German-language email corpus.

Research in emotion analysis is scattered across different label formats (e.g., polarity types, basic emotion categories, and affective dimensions), linguistic levels (word vs. sentence vs. discourse), and, of course, (few wellresourced but much more under-resourced) natural languages and text genres (e.g., product reviews, tweets, news).The resulting heterogeneity makes data and software developed under these conflicting constraints hard to compare and challenging to integrate.To resolve this unsatisfactory state of affairs we here propose a training scheme that learns a shared latent representation of emotion independent from different label formats, natural languages, and even disparate model architectures.Experiments on a wide range of datasets indicate that this approach yields the desired interoperability without penalizing prediction quality.Code and data are archived under

Emotion lexicons describe the affective meaning of words and thus constitute a centerpiece for advanced sentiment and emotion analysis.Yet, manually curated lexicons are only available for a handful of languages, leaving most languages of the world without such a precious resource for downstream applications.Even worse, their coverage is often limited both in terms of the lexical units they contain and the emotional variables they feature.In order to break this bottleneck, we here introduce a methodology for creating almost arbitrarily large emotion lexicons for any target language.Our approach requires nothing but a source language emotion lexicon, a bilingual word translation model, and a target language embedding model.Fulfilling these requirements for 91 languages, we are able to generate representationally rich high-coverage lexicons comprising eight emotional variables with more than 100k lexical entries each.We evaluated the automatically generated lexicons against human judgment from 26 datasets, spanning 12 typologically diverse languages, and found that our approach produces results in line with state-of-the-art monolingual approaches to lexicon creation and even surpasses human reliability for some languages and variables.Code and data are available at github.

The vast amount of social communication distributed over various electronic media channels (tweets, blogs, emails, etc.), so-called user-generated content (UGC), creates entirely new opportunities for today’s NLP research. Yet, data privacy concerns implied by the unauthorized use of these text streams as a data resource are often neglected. In an attempt to reconciliate the diverging needs of unconstrained raw data use and preservation of data privacy in digital communication, we here investigate the automatic recognition of privacy-sensitive stretches of text in UGC and provide an algorithmic solution for the protection of personal data via pseudonymization. Our focus is directed at the de-identification of emails where personally identifying information does not only refer to the sender but also to those people, locations, dates, and other identifiers mentioned in greetings, boilerplates and the content-carrying body of emails. We evaluate several de-identification procedures and systems on two hitherto non-anonymized German-language email corpora (CodE AlltagS+d and CodE AlltagXL), and generate fully pseudonymized versions for both (CodE Alltag 2.0) in which personally identifying information of all social actors addressed in these mails has been camouflaged (to the greatest extent possible).

Genes and proteins constitute the fundamental entities of molecular genetics. We here introduce ProGene (formerly called FSU-PRGE), a corpus that reflects our efforts to cope with this important class of named entities within the framework of a long-lasting large-scale annotation campaign at the Jena University Language & Information Engineering (JULIE) Lab. We assembled the entire corpus from 11 subcorpora covering various biological domains to achieve an overall subdomain-independent corpus. It consists of 3,308 MEDLINE abstracts with over 36k sentences and more than 960k tokens annotated with nearly 60k named entity mentions. Two annotators strove for carefully assigning entity mentions to classes of genes/proteins as well as families/groups, complexes, variants and enumerations of those where genes and proteins are represented by a single class. The main purpose of the corpus is to provide a large body of consistent and reliable annotations for supervised training and evaluation of machine learning algorithms in this relevant domain. Furthermore, we provide an evaluation of two state-of-the-art baseline systems — BioBert and flair — on the ProGene corpus. We make the evaluation datasets and the trained models available to encourage comparable evaluations of new methods in the future.

The massive digitization efforts related to historical newspapers over the past decades have focused on mass media sources and ordinary people as their primary recipients. Much less attention has been paid to newspapers published for a more specialized audience, e.g., those aiming at scholarly or cultural exchange within intellectual communities much narrower in scope, such as newspapers devoted to music criticism, arts or philosophy. Only some few of these specialized newspapers have been digitized up until now, but they are usually not well curated in terms of digitization quality, data formatting, completeness, redundancy (de-duplication), supply of metadata, and, hence, searchability. This paper describes our approach to eliminate these drawbacks for a major German-language newspaper resource of the Romantic Age, the Allgemeine Musikalische Zeitung (General Music Gazette). We here focus on a workflow that copes with a posteriori digitization problems, inconsistent OCRing and index building for searchability. In addition, we provide a user-friendly graphic interface to empower content-centric access to this (and other) digital resource(s) adopting open-source software for the purpose of Web presentation.

From 2017 to 2019 the Text REtrieval Conference (TREC) held a challenge task on precision medicine using documents from medical publications (PubMed) and clinical trials. Despite lots of performance measurements carried out in these evaluation campaigns, the scientific community is still pretty unsure about the impact individual system features and their weights have on the overall system performance. In order to overcome this explanatory gap, we first determined optimal feature configurations using the Sequential Model-based Algorithm Configuration (SMAC) program and applied its output to a BM25-based search engine. We then ran an ablation study to systematically assess the individual contributions of relevant system features: BM25 parameters, query type and weighting schema, query expansion, stop word filtering, and keyword boosting. For evaluation, we employed the gold standard data from the three TREC Precision Medicine (TREC-PM) installments to evaluate the effectiveness of different features using the commonly shared infNDCG metric.

Emotion Representation Mapping (ERM) has the goal to convert existing emotion ratings from one representation format into another one, e.g., mapping Valence-Arousal-Dominance annotations for words or sentences into Ekman’s Basic Emotions and vice versa. ERM can thus not only be considered as an alternative to Word Emotion Induction (WEI) techniques for automatic emotion lexicon construction but may also help mitigate problems that come from the proliferation of emotion representation formats in recent years. We propose a new neural network approach to ERM that not only outperforms the previous state-of-the-art. Equally important, we present a refined evaluation methodology and gather strong evidence that our model yields results which are (almost) as reliable as human annotations, even in cross-lingual settings. Based on these results we generate new emotion ratings for 13 typologically diverse languages and claim that they have near-gold quality, at least.

This paper introduces the Jena Document Information System (JeDIS). The focus lies on its capability to partition annotation graphs into modules. Annotation modules are defined in terms of types from the annotation schema. Modules allow easy manipulation of their annotations (deletion or update) and the creation of alternative annotations of individual documents even for annotation formalisms that by design do not support this feature.

Predicting the emotional value of lexical items is a well-known problem in sentiment analysis.While research has focused on polarity for quite a long time, meanwhile this early focus has been shifted to more expressive emotion representation models (such as Basic Emotions or Valence-Arousal-Dominance).This change resulted in a proliferation of heterogeneous formats and, in parallel, often smallsized, non-interoperable resources (lexicons and corpus annotations).In particular, the limitations in size hampered the application of deep learning methods in this area because they typically require large amounts of input data.We here present a solution to get around this language data bottleneck by rephrasing word emotion induction as a multi-task learning problem.In this approach, the prediction of each independent emotion dimension is considered as an individual task and hidden layers are shared between these dimensions.We investigate whether multi-task learning is more advantageous than single-task learning for emotion prediction by comparing our model against a wide range of alternative emotion and polarity induction methods featuring 9 typologically diverse languages and a total of 15 conditions.Our model turns out to outperform each one of them.Against all odds, the proposed deep learning approach yields the largest gain on the smallest data sets, merely composed of one thousand samples.

We assess the reliability and accuracy of (neural) word embeddings for both modern and historical English and German. Our research provides deeper insights into the empirically justified choice of optimal training methods and parameters. The overall low reliability we observe, nevertheless, casts doubt on the suitability of word neighborhoods in embedding spaces as a basis for qualitative conclusions on synchronic and diachronic lexico-semantic matters, an issue currently high up in the agenda of Digital Humanities.

Emotion analysis (EA) and sentiment analysis are closely related tasks differing in the psychological phenomenon they aim to catch. We address fine-grained models for EA which treat the computation of the emotional status of narrative documents as a regression rather than a classification problem, as performed by coarse-grained approaches. We introduce Ekman's Basic Emotions (BE) and Russell and Mehrabian's Valence-Arousal-Dominance (VAD) model—two major schemes of emotion representation following opposing lines of psychological research, i.e., categorical and dimensional models—and discuss problems when BEs are used in a regression approach. We present the first natural language system thoroughly evaluated for fine-grained emotion analysis using the VAD scheme. Although we only employ simple BOW features, we reach correlation values up until r = .65 with human annotations. Furthermore, we show that the prevailing evaluation methodology relying solely on Pearson's correlation coefficient r is deficient which leads us to the introduction of a complementary error-based metric. Due to the lack of comparable (VAD-based) systems, we, finally, introduce a novel method of mapping between VAD and BE emotion representations to create a reasonable basis for comparison. This enables us to evaluate VAD output against human BE judgments and, thus, allows for a more direct comparison with existing BE-based emotion analysis systems. Even with this, admittedly, error-prone transformation step our VAD-based system achieves state-of-the-art performance in three out of six emotion categories, out-performing all existing BE-based systems but one.

This article surveys efforts on text mining of the pharmacogenomics literature, mainly from the period 2008 to 2011. Pharmacogenomics (or pharmacogenetics) is the field that studies how human genetic variation impacts drug response. Therefore, publications span the intersection of research in genotypes, phenotypes and pharmacology, a topic that has increasingly become a focus of active research in recent years. This survey covers efforts dealing with the automatic recognition of relevant named entities (e.g. genes, gene variants and proteins, diseases and other pathological phenomena, drugs and other chemicals relevant for medical treatment), as well as various forms of relations between them. A wide range of text genres is considered, such as scientific publications (abstracts, as well as full texts), patent texts and clinical narratives. We also discuss infrastructure and resources needed for advanced text analytics, e.g. document corpora annotated with corresponding semantic metadata (gold standards and training data), biomedical terminologies and ontologies providing domain-specific background knowledge at different levels of formality and specificity, software architectures for building complex and scalable text analytics pipelines and Web services grounded to them, as well as comprehensive ways to disseminate and interact with the typically huge amounts of semiformal knowledge structures extracted by text mining tools. Finally, we consider some of the novel applications that have already been developed in the field of pharmacogenomic text mining and point out perspectives for future research.

BACKGROUND: Bio-molecular event extraction from literature is recognized as an important task of bio text mining and, as such, many relevant systems have been developed and made available during the last decade. While such systems provide useful services individually, there is a need for a meta-service to enable comparison and ensemble of such services, offering optimal solutions for various purposes. RESULTS: We have integrated nine event extraction systems in the U-Compare framework, making them intercompatible and interoperable with other U-Compare components. The U-Compare event meta-service provides various meta-level features for comparison and ensemble of multiple event extraction systems. Experimental results show that the performance improvements achieved by the ensemble are significant. CONCLUSIONS: While individual event extraction systems themselves provide useful features for bio text mining, the U-Compare meta-service is expected to improve the accessibility to the individual systems, and to enable meta-level uses over multiple event extraction systems such as comparison and ensemble.

In our approach to event extraction, dependency graphs constitute the fundamental data structure for knowledge capture. Two types of trimming operations pave the way to more effective relation extraction. First, we simplify the syntactic representation structures resulting from parsing by pruning informationally irrelevant lexical material from dependency graphs. Second, we enrich informationally relevant lexical material in the simplified dependency graphs with additional semantic meta data at several layers of conceptual granularity. These two aggregation operations on linguistic representation structures are intended to avoid overfitting of machine learning-based classifiers which we use for event extraction (besides manually curated dictionaries). Given this methodological framework, the corresponding JReX system developed by the JulieLab Team from Friedrich-Schiller-Universität Jena (Germany) scored on 2nd rank among 24 competing teams for Task 1 in the “BioNLP’09 Shared Task on Event Extraction,” with 45.8% recall, 47.5% precision and 46.7% F1-score on all 3,182 events. In more recent experiments, based on slight modifications of JReX and using the same data sets, we were able to achieve 45.9% recall, 57.7% precision, and 51.1% F1-score.

In lots of natural language processing tasks, the classes to be dealt with often occur heavily imbalanced in the underlying data set and classifiers trained on such skewed data tend to exhibit poor performance for low-frequency classes. We introduce and compare different approaches to reduce class imbalance by design within the context of active learning (AL). Our goal is to compile more balanced data sets up front during annotation time when AL is used as a strategy to acquire training material. We situate our approach in the context of named entity recognition. Our experiments reveal that we can indeed reduce class imbalance and increase the performance of classifiers on minority classes while preserving a good overall performance in terms of macro F-score.

Faced with the challenges to design an easy-to-use, immediately comprehensible and powerful expert-user interface to search very large document collections in the life sciences, we developed several system prototypes. Their main features were faceting of the domain vocabulary for browsing and searching, flexible search-state-dependent drilling of the terminological hierarchy, dynamic query term auto-completions, and highlighting of matched terms (including synonyms and spelling variants). Under lab conditions we then evaluated these features in several task-based scenarios using camera recordings, thinking-aloud protocols and questionnaires. The results reveal that faceting and highlighting were very well received, while auto-completions seemed less important or were misconceptualized as spelling aids.

MOTIVATION: The high level of polymorphism associated with the major histocompatibility complex (MHC) poses a challenge to organizing associated bioinformatic data, particularly in the area of hematopoietic stem cell transplantation. Thus, this area of research has great potential to profit from the ongoing development of biomedical ontologies, which offer structure and definition to MHC-data related communication and portability issues. RESULTS: We introduce the design considerations, methodological foundations and implementational issues underlying MaHCO, an ontology which represents the alleles and encoded molecules of the major histocompatibility complex. Importantly for human immunogenetics, it includes a detailed level of human leukocyte antigen (HLA) classification. We then present an ontology browser, search interfaces for immunogenetic fact and document retrieval, and the specification of an annotation language for semantic metadata, based on MaHCO. These use cases are intended to demonstrate the utility of ontology-driven bioinformatics in the field of immunogenetics. AVAILABILITY AND IMPLEMENTATION: The MaHCO Ontology is available via the BioPortal: http://www.bioontology.org/tools/portal/bioportal.html, and at: http://purl.org/stemnet/.

MOTIVATION: The recognition and normalization of textual mentions of gene and protein names is both particularly important and challenging. Its importance lies in the fact that they constitute the crucial conceptual entities in biomedicine. Their recognition and normalization remains a challenging task because of widespread gene name ambiguities within species, across species, with common English words and with medical sublanguage terms. RESULTS: We present GeNo, a highly competitive system for gene name normalization, which obtains an F-measure performance of 86.4% (precision: 87.8%, recall: 85.0%) on the BioCreAtIvE-II test set, thus being on a par with the best system on that task. Our system tackles the complex gene normalization problem by employing a carefully crafted suite of symbolic and statistical methods, and by fully relying on publicly available software and data resources, including extensive background knowledge based on semantic profiling. A major goal of our work is to present GeNo's architecture in a lucid and perspicuous way to pave the way to full reproducibility of our results. AVAILABILITY: GeNo, including its underlying resources, will be available from www.julielab.de. It is also currently deployed in the Semedico search engine at www.semedico.org.

BACKGROUND: An adequate and expressive ontological representation of biological organisms and their parts requires formal reasoning mechanisms for their relations of physical aggregation and containment. RESULTS: We demonstrate that the proposed formalism allows to deal consistently with "role propagation along non-taxonomic hierarchies", a problem which had repeatedly been identified as an intricate reasoning problem in biomedical ontologies. CONCLUSION: The proposed approach seems to be suitable for the redesign of compositional hierarchies in (bio)medical terminology systems which are embedded into the framework of the OBO (Open Biological Ontologies) Relation Ontology and are using knowledge representation languages developed by the Semantic Web community.

In the past years, a number of lexical association measures have been studied to help extract new scientific terminology or general-language collocations. The implicit assumption of this research was that newly designed term measures involving more sophisticated statistical criteria would outperform simple counts of co-occurrence frequencies. We here explicitly test this assumption. By way of four qualitative criteria, we show that purely statistics-based measures reveal virtually no difference compared with frequency of occurrence counts, while linguistically more informed metrics do reveal such a marked difference.

Researchers working in the life sciences domain in the past years have witnessed an enormous growth of literature—for the whole field as well as for their highly specialized areas of expertise. Only small portions of the biomedical knowledge are accessible in a structured way, i.e. through formatted databases. These few pieces of textually encoded knowledge that have gone into databases are, by default, manually extracted from documents and manually inserted into databases after careful curation efforts by highly skilled domain experts. Still, the vast majority of biomedical knowledge captured in texts is not at disposal when biomedical databases are queried. Life scientists have realized this loss of possibly highly relevant information and devised various forms of support. The weakest one is provided by information retrieval (IR) systems [for a life-science-centred survey; cf. Hersh (2002)]. Given a user-formulated query the terms from this query are appropriately matched with the terms occurring in documents from a large collection (e.g. the, currently, 14 million abstracts from Medline). Documents matching the query (up to a specified degree) are returned to the user for closer inspection and, possibly, ranked by some relevance-based sorting criterion (e.g. closeness of match). Information extraction (IE) provides a more powerful alternative that has mainly been developed in other areas different from molecular biology by the natural language processing community. IE aims at directly extracting relevant information from natural language documents [usually original text snippets, sentences, relevant phrases or even quasi-logical propositions, such as predicate–argument structures—for a general survey, cf. Gaizauskas and Wilks (1998) and for a life-science-centred view, cf. Blaschke et al. (2002) and Hoffmann et al. (2005)]. Unlike the output of IR systems, which only list relevant documents, IE systems provide immediate access to relevant information pieces via pre-specified information templates. This is achieved, however, at the price of supplying rather sophisticated language processing methodologies [e.g. taggers, chunkers, light semantic interpreters and information extraction rules; cf. for a survey, Hahn and Wermter (2006)], domain-specific developments and resources (e.g. databases and ontologies) and machine learning methodologies usually lack in IR systems. The evaluation of the degree of achievements from a biomedical perspective is an issue of active research. The IR stream is currently mainly investigated in the TREC (Text Retrieval Conference) Genomics track (Author Webpage), whereas there are several challenge evaluation platforms for IE that deal with often complementary problems from a biological perspective, the most important, currently, being the BioCreAtIvE (Critical Assessment of Information Extraction systems in Biology) contest (Author Webpage) [surveyed in Hirschman et al. (2005); see also Blaschke et al. (2005)]. Both forms of activities, IR as well as IE, are often labelled as text mining but miss a major extra requirement, namely the knowledge discovery perspective usually attributed to text mining procedures as well (Hearst, 1999). In particular, this relates to the identification and elimination of redundant knowledge as well as the recognition of (user-new?, expert-new? and community-new?) novel information. This value-adding, summarizing and selective aspect of text mining could be particularly helpful in taming the flood of literature for biomedical researchers, and will certainly be the focus of new developments in the years to come. The challenge evaluations, however, have already revealed some of the most pressing research problems for text analysis in the biomedical domain. In particular, biomedical terminology is extremely hard to deal with, in part because of the poor introduction of standards. It starts from identifying biological terms in a document (terms have a complex internal structure and are often composed of multiple, up to four or five, words), and leads to determining their conceptual type (e.g. genes, proteins and cell lines) and the way they are relationally linked (e.g. in terms of taxonomies or partonomies that in biology are often related with the organization of protein families). Further on, concrete factual biomedical knowledge (sometimes called relation mining) is also hard to extract from documents (e.g. ‘protein X inhibits protein Y’). This step is crucial for any sort of automated functional annotation in biological databases. Once this kind of knowledge has been successfully captured on a large scale making thousands of these propositions available, another severe follow-up problem arises, namely how to communicate this mass of information in a concise, comprehensible and, finally, useful way to the researcher in the laboratory. For this purpose, text mining systems have simply borrowed visualization techniques that were originally developed for numerical data mining. However, symbolic abstraction mechanisms leading, e.g. to the automatic generation of pathway diagrams from this huge dataset are still an area that requires further developments. The above-mentioned research problems have motivated the creation of a Network of Excellence—‘Semantic Interoperability and Data Mining in Biomedicine’ (Semantic Mining, Author Webpage)—which has been funded by the European Community since 2004 under the FP6 Programme ‘Integrating and Strengthening the European Research Area’. The NoE has initiated a series of conferences dedicated to these particular challenges of data mining and text mining in the life sciences [for a life-science-centred survey, cf. Ananiadou and McNaught (2006)]. The first of these symposia was held under the title ‘Semantic Mining in Biomedicine’ (SMBM) in Hinxton (Cambridgeshire, UK) from April 10–13, 2005 organized by Stefan Schulz, Freiburg University Hospital, and Dietrich Rebholz-Schuhmann, EBI-EMBL, Hinxton (see Author Webpage). A specific feature of SMBM meetings is their focus on content-oriented methodologies and semantic resources—either controlled vocabularies, terminologies and formal domain ontologies, or conceptually as well as propositionally annotated corpora—in order to improve text-based biomedical knowledge management, e.g. through document classification, text or fact retrieval, information extraction, or (real) text mining. Also methodologies being discussed should look at applications to real-world problems in molecular biology and biomedicine [for a review of systems currently operational in this domain, see Krallinger and Valencia (2005)]. We had the honour of chairing the programme committee that comprised 21 scientists who evaluated the 28 submissions and selected 12 papers for their presentation in the conference. Four outstanding papers were selected for publication in Bioinformatics, after additional extensive reviews and revisions. Seven full papers plus the abstracts of these selected presentations appeared in the proceedings of the conference [Hahn and Valencia (2005)]. The selected papers cover research performed under the following headings: (1) entity identification—identification of gene names, (2) text classification classification—assignment of sentences to known Gene Ontology (GO) and Medical Subject Headings (MeSH) classes, (3) identification of relations in text—extracting phosphorylation and gene control networks; and (4) identification of new concepts—proposing new GO categories and their corresponding associated genes. Automatic Term List Generation for Entity Tagging by Ted Sandler, Andrew I. Schein and Lyle H. Ungar from the University of Pennsylvania. The basic problem of term characterization is tackled here with an unsupervised approach based on clustering terms (gene names) using additional context information. The clustering approach is related to the distributional clustering technique published previously and the context information provided include neighbouring and syntactic relations. The basic sources of information were sentences from the Biocreative gene tagging challenge and a set of two million Medline abstracts. The results are significantly better than those obtained with standard taggers based on dictionaries of genes. Interestingly enough, the results are still far from matching those obtained in other domains such as newswire information, most probably owing to the additional complexity of biological nomenclature. Automatic Assignment of Biomedical Categories: Toward a Generic Approach by Patrick Ruch from the University Hospitals of Geneva. Describes new results on the automatical assignment of biomedical categories with a system that is designed to be largely data-independent. The system includes a pattern-based identification and vector space retrieval engine, and uses both stems and linguistically motivated information, and it is applied to the classification of sentences in MeSH and GO classes. The results are compared with those obtained in the related BioCreative task. Extraction of Regulatory Gene/Protein Networks from Medline by Jasmin Saric, Lars Juhl Jensen, Rossitza Ouzounova, Isabel Rojas and Peer Bork, from EML Research and EMBL both in Heidelberg. The authors address the problem of extracting two key types of biological relations, which are the regulators of protein function by phosphorylation and the control of gene expression. Their rule-based String-IE system uses organism-specific lexicons that are incorporated in the training of a part-of-speech tagger that uses the GENIA corpus as background information. In practice, the system is able to extract 3319 phosphorylations or gene expression relations, with a sustained level of accuracy across different organisms. Automatic Extension of GO with Flexible Identification of Candidate Terms by Jin-Bok Lee, Jung-jae Kim and Jong C. Park from KAIST in Daejeon, Korea. The authors tackle the problem of identifying new GO concepts using existing GO concepts and their relations in text. The proposed new terms are compared with those created by human experts in subsequent releases of GO. This type of approaches can be useful for speeding up the process of annotation, and for increasing the number of categories in which GO concepts can be divided when they have a large number of genes assigned.

Most technical and scientific terms are comprised of complex, multi-word noun phrases but certainly not all noun phrases are technical or scientific terms. The distinction of specific terminology from common non-specific noun phrases can be based on the observation that terms reveal a much lesser degree of distributional variation than non-specific noun phrases. We formalize the limited paradigmatic modifiability of terms and, subsequently, test the corresponding algorithm on bigram, trigram and quadgram noun phrases extracted from a 104-million-word biomedical text corpus. Using an already existing and community-wide curated biomedical terminology as an evaluation gold standard, we show that our algorithm significantly outperforms standard term identification measures and, therefore, qualifies as a high-performant building block for any terminology identification system. We also provide empirical evidence that the superiority of our approach, beyond a 10-million-word threshold, is essentially domain- and corpus-size-independent.

We introduce a light-weight interlingua for a cross-language document retrieval system in the medical domain. It is composed of equivalence classes of semantically primitive, language-specific subwords which are clustered by interlingual and intralingual synonymy. Each subword cluster represents a basic conceptual entity of the language-independent interlingua. Documents, as well as queries, are mapped to this interlingua level on which retrieval operations are performed. Evaluation experiments reveal that this interlingua-based retrieval model outperforms a direct translation approach.

The bottleneck for dictionary-based cross-language information retrieval is the lack of comprehensive dictionaries, in particular for many different languages. We here introduce a methodology by which multilingual dictionaries (for Spanish and Swedish) emerge automatically from simple seed lexicons. These seed lexicons are automatically generated, by cognate mapping, from (previously manually constructed) Portuguese and German as well as English sources. Lexical and semantic hypotheses are then validated and new ones iteratively generated by making use of co-occurrence patterns of hypothesized translation synonyms in parallel corpora. We evaluate these newly derived dictionaries on a large medical document collection within a cross-language retrieval setting.

We describe an ontology engineering methodology by which conceptual knowledge is extracted from an informal medical thesaurus (UMLS) and automatically converted into a formal description logics system (LOOM). Our approach consists of four steps: concept definitions are automatically generated from the UMLS, Integrity checking of taxonomic and partonomic hierarchies is performed by LOOM's terminological classifier, cycles and inconsistencies are eliminated, as well as incremental refinement of the evolving knowledge base is performed by a domain expert. We report an experiments with a very large knowledge base composed of 164,000 concepts and 76,000 relations.

We here deal with mereotopological properties of parts and associated wholes, locations and empty spaces (holes), with particular reference to biological structures. Our considerations lead to a basic ontology which contains 'solid object', 'hole' and 'boundary' as mutually disjoint primitives. Formally, we embed the relations 'part-of' and 'location-of' into a parsimonious description logic (ALC) and emulate partonomic and spatial reasoning involving these relations by terminological subsumption. In contrast to common conceptualizations, we do not distinguish between solids and the regions they occupy, as well as we allow solids to have holes as proper parts. In order to support these modeling decisions, we discuss various concrete examples from human anatomy.

We introduce a methodology for automating the maintenance and growth of domain-specific concept taxonomies and grammatical class hierarchies simultaneously, based on knowledge capture from natural language texts. The assimilation process is centered around the linguistic and conceptual `quality' of various forms of evidence underlying the generation, assessment and on-going refinement of lexical and concept hypotheses. On the basis of the strength of evidence, hypotheses are ranked according to plausibility, and the most reasonable ones are selected for assimilation into the given lexical class hierarchy and domain ontology.

We specify an algorithm that builds up a hierarchy of referential discourse segments from local centering data. The spatial extension and nesting of these discourse segments constrain the reachability of potential antecedents of an anaphoric expression beyond the local level of adjacent center pairs. Thus, the centering model is scaled up to the level of the global referential structure of discourse. An empirical evaluation of the algorithm is supplied.

Article Knowledge mining from textual sources Share on Authors: Udo Hahn Computational Linguistics Lab - Text Knowledge Engineering Group, Freiburg University, Werthmannplatz 1, D-79085 Freiburg, Germany Computational Linguistics Lab - Text Knowledge Engineering Group, Freiburg University, Werthmannplatz 1, D-79085 Freiburg, GermanyView Profile , Klemens Schnattinger Computational Linguistics Lab - Text Knowledge Engineering Group, Freiburg University, Werthmannplatz 1, D-79085 Freiburg, Germany Computational Linguistics Lab - Text Knowledge Engineering Group, Freiburg University, Werthmannplatz 1, D-79085 Freiburg, GermanyView Profile Authors Info & Claims CIKM '97: Proceedings of the sixth international conference on Information and knowledge managementJanuary 1997 Pages 83–90https://doi.org/10.1145/266714.266865Online:01 January 1997Publication History 5citation572DownloadsMetricsTotal Citations5Total Downloads572Last 12 Months10Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

Based on empirical evidence from a free word order language (German) we propose a fundamental revision of the principles guiding the ordering of discourse entities in the forward-looking centers within the centering model. We claim that grammatical role criteria should be replaced by indicators of the functional information structure of the utterances, i.e., the distinction between context-bound and unbound discourse elements. This claim is backed up by an empirical evaluation of functional centering.

Development assistance for interactive database applications (DAIDA) is an experimental environment for the knowledge-assisted development and maintenance of database-intensive information systems from object-oriented requirements and specifications. Within the DAIDA framework, an approach to integrate different tasks encountered in software projects via a conceptual modeling strategy has been developed. Emphasis is put on integrating the semantics of the software development domain with aspects of group work, on social strategies to negotiate problems by argumentation, and on assigning responsibilities for task fulfillment by way of contracting. The implementation of a prototype is demonstrated with a sample session.>

This panel will employ two different interpretations of the phrase “growing up” to address areas of common interest between hypertext and information retrieval researchers. First, the panelists will question whether or not hypertext is “growing up” as a scientific discipline; They will discuss characteristics that separate hypertext research from other related disciplines. Second, the panelists will discuss the problems encountered when a hypertext system “grows up” in size and complexity; They will discuss the very real problems expected when representing and integrating large knowledge bases, accommodating multiple users, and distributing single logical hypertexts across multiple physical sites.

Semantically driven natural language parsers have found wide-spread application as a text processing methodology for knowledge-based information retrieval systems. It is argued that this parsing technique particularly corresponds to the requirements inherent to large-scale text analysis. Unfortunately, this approach suffers from several shortcomings which demand a thorough reformulation of its paradigm. Incorporating principles from conceptual analysis and word expert parsing in a model of lexically distributed text parsing, the focus of the modifications proposed in this article, is on a clean declarative separation of linguistio and other knowledge representation levels, abstraction mechanisms leading to a small collection of specification primitives for the parser, and an attempt to incorporate linguistic generalizations and modularization principles into the design of a semantic text grammar. A sample parse illustrates the operation and linguistic coverage of a lexically distributed text parser based on these theoretical considerations with respect to the semantic analysis of noun groups, simple assertional sentences, nominal anaphora, and textual ellipsis.

In this paper prototype versions of two word experts for text analysis are dealt with which demonstrate that word experts are a feasible tool for parsing texts on the level of text cohesion as well as text coherence. The analysis is based on two major knowledge sources: context information is modelled in terms of a frame knowledge base, while the co-text keeps record of the linear sequencing of text analysis. The result of text parsing consists of a text graph reflecting the thematic organization of topics in a text.