The Conference on Computational Natural Language Learning features a shared task, in which participants train and test their learning systems on the same data sets. In 2007, as in 2006, the shared task has been devoted to dependency parsing, this year with both a multilingual track and a domain adaptation track. In thispaper, we definethe tasksof the different tracks and describe how the data sets were created from existing treebanks for ten languages. In addition, we characterize the different approaches of the participating systems, report the test results, and provide a first analysis of these results.

/pdf/the-conll-2007-shared-task-on-dependency-parsing-3erv3sel00.pdf

The CoNLL 2007 Shared Task on Dependency Parsing

This paper presents a part-of-speech tagger which is specifically tuned for biomedical text We have built the tagger with maximum entropy modeling and a state-of-the-art tagging algorithm The tagger was trained on a corpus containing newspaper articles and biomedical documents so that it would work well on various types of biomedical text Experimental results on the Wall Street Journal corpus, the GENIA corpus, and the PennBioIE corpus revealed that adding training data from a different domain does not hurt the performance of a tagger, and our tagger exhibits very good precision (97% to 98%) on all these corpora We also evaluated the robustness of the tagger using recent MEDLINE articles

Developing a robust part-of-speech tagger for biomedical text

Special Report: NCBI disease corpus: A resource for disease name recognition and concept normalization

Motivation Text mining has become an important tool for biomedical research. The most fundamental text-mining task is the recognition of biomedical named entities (NER), such as genes, chemicals and diseases. Current NER methods rely on pre-defined features which try to capture the specific surface properties of entity types, properties of the typical local context, background knowledge, and linguistic information. State-of-the-art tools are entity-specific, as dictionaries and empirically optimal feature sets differ between entity types, which makes their development costly. Furthermore, features are often optimized for a specific gold standard corpus, which makes extrapolation of quality measures difficult. Results We show that a completely generic method based on deep learning and statistical word embeddings [called long short-term memory network-conditional random field (LSTM-CRF)] outperforms state-of-the-art entity-specific NER tools, and often by a large margin. To this end, we compared the performance of LSTM-CRF on 33 data sets covering five different entity classes with that of best-of-class NER tools and an entity-agnostic CRF implementation. On average, F1-score of LSTM-CRF is 5% above that of the baselines, mostly due to a sharp increase in recall. Availability and implementation The source code for LSTM-CRF is available at https://github.com/glample/tagger and the links to the corpora are available at https://corposaurus.github.io/corpora/ . Contact habibima@informatik.hu-berlin.de.

/pdf/deep-learning-with-word-embeddings-improves-biomedical-named-4ertt2v5or.pdf

Deep learning with word embeddings improves biomedical named entity recognition.

Dependency-based methods for syntactic parsing have become increasingly popular in natural language processing in recent years. This book gives a thorough introduction to the methods that are most widely used today. After an introduction to dependency grammar and dependency parsing, followed by a formal characterization of the dependency parsing problem, the book surveys the three major classes of parsing models that are in current use: transition-based, graph-based, and grammar-based models. It continues with a chapter on evaluation and one on the comparison of different methods, and it closes with a few words on current trends and future prospects of dependency parsing. The book presupposes a knowledge of basic concepts in linguistics and computer science, as well as some knowledge of parsing methods for constituency-based representations. Table of Contents: Introduction / Dependency Parsing / Transition-Based Parsing / Graph-Based Parsing / Grammar-Based Parsing / Evaluation / Comparison / Final Thoughts

Dependency Parsing

We describe an approach to two areas of biomedical information extraction, drug development and cancer genomics. We have developed a framework which includes corpus annotation integrated at multiple levels: a Treebank containing syntactic structure, a Propbank containing predicate-argument structure, and annotation of entities and relations among the entities. Crucial to this approach is the proper characterization of entities as relation components, which allows the integration of the entity annotation with the syntactic structure while retaining the capacity to annotate and extract more complex events. We are training statistical taggers using this annotation for such extraction as well as using them for improving the annotation process.

/pdf/integrated-annotation-for-biomedical-information-extraction-379asx0jgt.pdf

Integrated Annotation for Biomedical Information Extraction

A complex relation is any n-ary relation in which some of the arguments may be be unspecified. We present here a simple two-stage method for extracting complex relations between named entities in text. The first stage creates a graph from pairs of entities that are likely to be related, and the second stage scores maximal cliques in that graph as potential complex relation instances. We evaluate the new method against a standard baseline for extracting genomic variation relations from biomedical text.

/pdf/simple-algorithms-for-complex-relation-extraction-with-19vgjp26rb.pdf

Simple Algorithms for Complex Relation Extraction with Applications to Biomedical IE

Summary: MTag is an application for identifying and extracting clinical descriptions of malignancy presented in text. The application uses the machine learning technique Conditional Random Fields and incorporates domain-specific features. MTag was tested with 1,010 training and 432 evaluation documents pertaining to cancer genomics. Our experiments resulted in 0.85 precision, 0.82 recall, and 0.83 F-measure on the evaluation set. Availability: The software is available at http://bioie.ldc.upenn.edu/index.jsp Contact: yajin@mail.med.upenn.edu

Scott Winters

Papers

Integrated Annotation for Biomedical Information Extraction

Simple Algorithms for Complex Relation Extraction with Applications to Biomedical IE

Identifying and extracting malignancy types in cancer literature