Dieser Webauftritt wurde archiviert und wird nicht mehr aktualisiert.

Bei technischen Fragen kontaktieren Sie bitte das Webteam. Es gilt die Datenschutzerklärung der TU Darmstadt.

Advanced tutorials

The Language and Knowledge Engineering (LKE) doctoral program occasionally hosted advanced tutorials on certain topics of general interest for LKE researchers. The goal of the advanced tutorials is to:

  • collectively learn basic knowledge and best practices in areas that are important in everyone's daily work and require substantial effort to master;
  • practice teaching and presentation skills.

The Language and Knowledge Engineering tutorials were organized on a case-by-case basis, when a topic of collective interest is found. LKE hosted the following tutorials.

Previous sessions


  • JoBimText: A framework for distributional semantics.
  • Tutorial on Motives.
  • Preference Learning / Reinforcement Learning.


  • Semi-supervised learning for semantic text processing- an introductory overview: Machine Learning methods that are able to learn from both labeled and unlabeled data are commonly subsumed under the term semi-supervised learning. Semi-supervised learning is a very important approach for many practical applications in text processing, like relation extraction and sentiment analysis.
    The first part of the tutorial reviews the most important methods for semi-supervised learning in the context of semantic text processing. Topics include bootstrapping methods, self-learning, distant supervision and their usage for various NLP tasks (e.g. sentiment analysis).
    The second part of the tutorial will be organized as a mini-workshop: participants discuss in small groups, which semi-supervised methods would be best suited for their particular tasks and how they could be applied. Each group will then summarize the outcome of their discussion briefly in a 3-minute slot, leading to a plenary discussion and identification of topics relevant for further study.  This tutorial was developed and presented by Dr. Judith Eckle-Kohler and Silvana Hartmann.

    • Please contact Dr. Judith Eckle-Kohler or Silvana Hartmann for furter information.

  • Introduction to Deep Learning for Natural Language Processing:  Deep Learning has been successfully applied to a wide range of NLP tasks, including various tagging tasks, sentiment analysis, paraphrasing, question-answering, machine translation, reasoning over knowledge bases and many more. All methods have in common that they use unsupervised pre-training, which in many cases makes external hand-designed resources or time-intensive feature engineering obsolete. Instead, the unsupervised pre-training learns by itself good features and representations for the desired task.
    In this tutorial we cover the basics of deep learning and its application to NLP. We start with an introduction to deep learning and deep neural networks, followed by introducing the concept of language models. Language models cover the semantic and morphosyntactic regularities for natural language and are useful for many tasks. Then we go on with different common neural network architectures used in NLP. Besides covering the theoretical aspects of deep learning, we will also present useful tools and frameworks for deep learning. These tools are handy for many classification tasks and can often be used out of the box.  This tutorial was developed and presented by Nils Reimers, Carsten Schnober, Dr. Judith Eckle-Kohler and Silvana Hartmann.

  • Shortest Path Algorithms and Algorithmic Efficiency Tutorial:  Finding a shortest (fastest, cheapest ...) path between node A and B in a network is a classical algorithmic problem and of practical relevance in numerous settings.  In particular, it is a prototypical showcase for various aspects of algorithmics in general.  Taking the shortest path problem and Dijkstra's algorithm as representatives of other algorithmic problems, we will study unforeseen applications, speed-up techniques, and aspects of software engineering.  This tutorial was developed and presented by Prof. Karsten Weihe.

  • Multilabel Classification and Mulan Tutorial:  In learning from multilabel data, the objective is to learn a mapping between an object and a set of non-exclusive labels, in contrast to (single-label) multiclass or binary classification, where it is only allowed to associate exactly one class to each object.  This research field has become very popular in recent times.  This is mainly due to the inherently multilabel nature of many data and that this data has become actually available to the research community.  Classical scenarios are keyword tagging of texts, images, videos and other media.  In this tutorial, we gave an introduction into multilabel classification.  We showed typical and non-typical applications, the differences and commonalities to other scenarios such as multiclass and binary classification, and we showed up the important issues, challenges and pitfalls in multilabel classification and how to solve them and which algorithms and approach exist for this.  We closed the tutorial with a short introduction into the usage of one of the main frameworks for handling multilabel data, namely MULAN.  This tutorial was developed and presented by Prof. Johannes Fürnkranz and Dr. Eneldo Loza Mencía.


  • Structured Prediction Models Tutorial:  Learning mappings between arbitrary structured input and output variables is a fundamental problem in machine learning. It covers many natural learning tasks and challenges the standard model of learning a mapping from independently drawn instances to a small set of labels. Potential applications include classification with a class taxonomy, named entity recognition, and natural language parsing. The tutorial reviewed large-margin classifiers and generalizes the ideas to structured prediction problems using graphical models.  This tutorial was developed and presented by Prof. Ulf Brefeld.

  • Scrum Tutorial: This tutorial was for the development and exchange of ideas for student supervision using Scrum.  There was a short brainstorming session and a brief introduction of Scrum concepts. Next, there was a Scrum project in small groups and a discussion of which Scrum aspects can be used for student supervision or  research projects. This tutorial was developed and presented by Nico Erbs.

    • Please contact Nico Erbs for further information.

  • DKPro Big Data  Tutorial:  In this tutorial, Hans-Peter Zorn gave an introduction on large scale natural language processing using DKPro Big Data und using the UKP Hadoop Cluster.  The tutorial covered: (1) Very compressed introduction to MapReduce and Hadoop, (2) Guide how to use DKPro Big Data to execute UIMA pipelines on the cluster, and (3) A summary on methods for aggregating and analyzing the results. 

    • Please contact Hans-Peter Zorn for further information.

  • WebAnno: A Flexible, Web-based and Visually Supported System for Distributed Annotations: This tutorial covers the use of the WebAnno tool for annotation, annotation analysis, and coordination with a crowdsourcing platform.  The tutorial was organized by Seid Muhie Yimam.

  • How to Design Effective Visualizations for Natural Language Processing: The tutorial covers the fundamentals of visualization and provides examples for advanced visualization techniques with a focus on Visual Document Analysis. The tutorial was developed by Dr. Daniela Oelke and held at the GSCL 2013 conference.

    • Please contact Dr. Daniela Oelke for further information.

  • The Practitioner's Cookbook for Linked Lexical Resources: This tutorial covers the fundamental concepts of linked lexical resources and gives a practical hands-on introduction to the particular Opens internal link in current windowlinked lexical resource UBY. The tutorial was developed by Prof. Iryna Gurevych and Dr. Judith Eckle-Kohler. The organizers held the tutorial at two international conferences, RANLP 2013 and GSCL 2013.

  • DKPro-TC tutorial: This tutorial gave a high-level overview of the DKPro Text Classification framework. It included a hands-on session for interested developers and users with the help of two simple use cases. The tutorial was organized by Johannes Daxenberger.

  • Two-Day Invited Tutorial: Watson and the Deep QA architecture: This tutorial covered the Watson Deep QA architecture. Open domain Question Answering (QA) is a long-standing research problem. Recently, IBM took on this challenge in the context of Jeopardy!, a well-known TV quiz show.  The development of a system able to compete with grand champions in the Jeopardy! challenge led to the design of the DeepQA architecture and the implementation of Watson. The tutorial was given by Dr. Alfio Gliozzo and Prof. Chris Biemann.

    • Please contact Dr. Alfio Gliozzo or Prof. Chris Biemann for further information.



  • Inter Coder Agreement analysis: This tutorial gave a detailed introduction to studying and interpreting inter coder agreement in annotation projects. The tutorial was held by Christian Meyer.


Prof. Iryna Gurevych

WeRC Coordinator

+49 6151 16 25290

Foto von Iryna Gurevych
A A A | Drucken Drucken | Impressum Impressum | Sitemap Sitemap | Suche Suche | Kontakt Kontakt | Webseitenanalyse: Mehr Informationen
zum Seitenanfangzum Seitenanfang