This series features presentations and discussions of current issues in linguistic modeling and its interfaces. This includes linguistic modeling in computational linguistics, language acquisition research, Intelligent Computer-Assisted Language Learning, and education, as well as theoretical linguistic research with a focus on the interfaces of syntax and information structure. It is open to anyone interested in this interdisciplinary enterprise.

• When: Fr 10ct-12
• Where: Room 1.13 (Blochbau, Wilhelmstr. 19)
• Mailing list for related announcements: http://mailman.sfs.uni-tuebingen.de/cgi-bin/mailman/listinfo/icall

A list of talks in previous semesters can be found here: Winter 13/14, Summer 13, Winter 12/13, Summer 12, Winter 11/12, Summer 11, Summer 10, Winter 09/10, Summer 09

Sessions

• April 25 Sowmya Vajjala and Detmar Meurers:

  Assessing the relative reading level of sentence pairs for text simplification
  (to be presented at EACL 2014, cf. http://purl.org/dm/papers/Vajjala.Meurers-14-eacl.html)

  While the automatic analysis of the readability of texts has a long history, the use of readability assessment for text simplification has received only little attention so far. In this paper, we explore readability models for identifying differences in the reading levels of simplified and unsimplified versions of sentences.

  Our experiments show that a relative ranking is preferable to an absolute binary one and that the accuracy of identifying relative simplification depends on the initial reading level of the unsimplified version. The approach is particularly successful in classifying the relative reading level of harder sentences.

  In terms of practical relevance, the approach promises to be useful for identifying particularly relevant targets for simplification and to evaluate simplifications given specific readability constraints.  
   
  

  Exploring Measures of \Readability” for Spoken Language:
  Analyzing linguistic features of subtitles to identify age-specific TV programs
  (to be presented at PITR held at EACL 2014, cf. http://purl.org/dm/papers/Vajjala.Meurers-14-pitr.html)

  We investigate whether measures of readability can be used to identify age-specific TV programs. Based on a corpus of BBC TV subtitles, we employ a range of linguistic readability features motivated by Second Language Acquisition and Psycholinguistics research.

  Our hypothesis that such readability features can successfully distinguish between spoken language targeting different age groups is fully confirmed. The classifiers we trained on the basis of these readability features achieve a classification accuracy of 95.9%. Investigating several feature subsets, we show that the authentic material targeting specific age groups exhibits a broad range of linguistics and psycholinguistic characteristics that are indicative of the complexity of the language used.  
   
  

• May 2: Detmar Meurers Unterscheidet sich die Sprache von deutschen Muttersprachlern, chinesischen, weissrussischen, und schwedisen Deutschlernern hinsichtlich der Art der sprachlichen Komplexifizierung? Erste Analysen der Kobalt-DaF Daten  
   
  
• May 9: no meeting (MERLIN Meeting in Dresden)  
   
  
• May 16, 10st!: B8 Workshop “Modifikatoren in Korpora des Deutschen” von Hagen Hirschmann im SFB 833 (Nauklerstrasse 35, Room 3.08)

  Abstract: Der Workshop richtet sich an Studierende, Forschende und Lehrende in linguistischen Fachbereichen. Es wird praktisch vermittelt, welche korpuslinguistische Methoden genutzt werden können, um verschiedene Modifikatoren (z. B. semantische, syntaktische oder formspezifische Adverbialklassen) innerhalb vorhandener Korpusdaten gefunden und ggf. statistisch ausgewertet werden können. Die TeilnehmerInnen können hierbei selber Korpussuchen durchführen. In vielen Fällen ist eine Weiterverarbeitung von Korpusdaten nötig, um auf gewünschte Kategorien zuverlässig zugreifen zu können. Auch hier werden effiziente Methoden (Annotationsguidelines und -werkzeuge) angesprochen und je nach zeitlichen Möglichkeiten vom gesamten Kurs erprobt. Die Teilnahme erfordert keine Vorkenntnisse. Alle Teilnehmer/innen sollten ihr Laptop mitbringen!  
   
  

• May 23 Yulia Krivanek and Johannes Widmann (Transline GmbH)
  Current developments in the translation industry – workflows and computational linguistics

  Abstract: This talk will present an overview on the state of the art in the translation of technical documents. Johannes Widmann and Yulia Krivanek, two practicing professionals in the industry, will present 4 myths about the translation industry and juxtapose them to the (dire) reality. We will start by discussing translation workflows that aim at automation. Secondly, we will point out issues where we believe that computational linguistics could positively influence productivity. There will be plenty of time for discussion and questions.  
   
  

• May 30 Ramon Ziai (on joint work of the ETS team during his time there):
  Is getting the right answer just about choosing the right words? The role of syntactically-informed features in short answer scoring

  Abstract: Developments in the educational landscape have spurred greater interest in the problem of automatically scoring short answer questions. A recent shared task on this topic revealed a fundamental divide in the modeling approaches that have been applied to this problem, with the best-performing systems split between those that employ a knowledge engineering approach and those that almost solely leverage lexical information (as opposed to higher-level syntactic information) in assigning a score to a given response. This paper aims to introduce the NLP community to the largest corpus currently available for short-answer scoring, provide an overview of methods used in the shared task using this data, and explore the extent to which more syntactically-informed features can contribute to the short answer scoring task in a way that avoids the question-specific manual effort of the knowledge engineering approach.

  Joint work by Derrick Higgins, Chris Brew, Michael Heilman, Ramon Ziai, Lei Chen, Aoife Cahill, Michael Flor, Nitin Madnani, Joel Tetreault, Dan Blanchard, Diane Napolitano, Chong Min Lee and John Blackmore, cf. http://arxiv.org/abs/1403.0801  
   
  

• June 6 Heike Cardoso, Yulia Svetashova
  Visualization of Verbs and their Dependency Structure in English for Academic Purpose Corpora

  Abstract: In order to improve the students’ proficiency level in academic English, EAP practitioners provide training in a distinctive use of certain verbs (reporting / ergative / modal verbs, etc.). The verb usage is fully described by its argument frames (i.e. by specifying the number and type of arguments which the verb needs to be used in a well-formed construction (Allen, 2009)).

  The proposed visualisation tool takes a text corpus as input, which is further processed by the Stanford CoreNLP pos-tagger and dependency parser. The results are stored in a complex data structure which allows both verb and argument queries (form- or lemma-based). In the next step, we use results from the previous processing steps to visualise verb dependencies.

  Our approach goes beyond concordance tools with their basic frequency lists and key word in context functionality. EAP students can find and explore the verb dependencies, their relation types, linguistic properties, and contexts of use. The questions that can be addressed using the presented visualisation include (but are not limited to) the following: What are the typical contexts for a particular verb? Does it require any phrasal verb particle? What prepositions are appropriate before a noun verb dependency? Can it take a clausal complement? Is it followed by a gerund or an infinitive form of a verb? What verb tense is more frequent in a given corpus? The tool is designed to facilitate verb comparison across corpora (specifically, a corpus of “expert” writing and a learner corpus (Nesi, 2008) can be compared).

  (to be presented in “EAP and Corpora”, BALEAP PIM, Coventry University)  
   
  

• June 27 Katrin Wisniewski (TU Dresden)
  Connecting the CEFR level system to learner language (slides of talk)

  Abstract: In spite of the widespread distribution of the Common European Framework of Reference (CEFR) and its level system, almost nothing is known about its power in describing empirical learner language. Up to now, validation has focused exclusively on the reliability with which human raters apply the CEFR scales for assessing learner texts. This talk will present an approach for empirical scale validation that is based on the analysis of learner language and CEFR scale operationalization.  
   
  

• July 11 Thomas François (CENTAL, Université catholique de Louvain)
  Challenges for specializing readability formulas: a case study for administrative texts (slides of talk)

  Abstract: Readability aims at automatically assessing the difficulty of texts for a given population, using some of the linguistic characteristics of the texts. The classic attempts to do so (Flesch, 1984 ; Dale and Chall, 1948) have been widely used for educational or communicative purposes in the U.S. However, their limitations have been stressed as soon as the end of the 70’s (Kintsch and Vipond, 1979), which eventually led to the investigation of new research avenues, relying on computational linguistics as well as machine learning techniques to improve traditional approaches (Collins-Thompson and Callan, 2005 ; Schwarm and Ostendorf, 2005). However, although the efficiency of this approach has been widely acknowledged, a recent report (Nelson et al., 2012) has stressed that such models might not behave better than traditional formulas. This might indicate the existence of a performance ceiling in the field that may stem from the generic nature of the models.

  In this presentation, we first discuss the need for more specialized models, as regards the population considered as well as the genre of the texts. We will report preliminary data in favor of such specialization based on our own research on the readability of French as a foreign language (FFL). We will then discuss the pragmatic challenges of this type of specialization, in particular the issue of obtaining data annotated in terms of difficulty for the target population. Finally, we will report investigation as regards possible answers to these challenges, among which the development a crowd-sourcing platform to get annotated data for readability. Our approach will be illustrated through a actual case study on the readability of administrative texts for French.  
   
  

• July 25 Willy Cheung (Universität Tübingen, Neural and Behavioral Sciences Program)
  Improving Alingment-Based Content Assessment: Contextualizing Alignments using Transformation-Based Error-Driven Learning

  Abstract: Content Assessment is currently a prominent research area in Computational Linguistics and Natural Language Processing, and remains a challenging and interesting topic. In this talk I will present the application of Transformation-Based Error-Driven learning to improve alignment-based content assessment. Specifically, I focus on improving the quality of linguistically abstract alignments by contextualizing them using dependency relations constraints. I will explain the intuition and motivation behind the technique, present some results, and discuss current issues and future directions.  
   
  

• August 15 Ramon Ziai (SFB 833, Universität Tübingen)
  Focus Annotation in Reading Comprehension Data Abstract: When characterizing the information structure of sentences, the so-called focus identifies the part of a sentence addressing the current question under discussion in the discourse. While this notion is precisely defined in formal semantics and potentially very useful in theoretical and practical terms, it has turned out to be difficult to reliably annotate focus in corpus data.

  We present a new focus annotation effort designed to overcome this problem. On the one hand, it is based on a task-based corpus providing more explicit context. The annotation study is based on the CREG corpus (Ott et al., 2012), which consists of answers to explicitly given reading comprehension questions. On the other hand, we operationalize focus annotation as an incremental process including several substeps which provide guidance, such as explicit answer typing.

  We evaluate the focus annotation both intrinsically by calculating agreement between annotators and extrinsically by showing that the focus information substantially improves the automatic meaning assessment of answers in the CoMiC system (Meurers et al., 2011).

  (Joint work with Detmar Meurers to be presented at the ACL Linguistic Annotation Workshop, August 23-24, 2014, Dublin, Ireland)  
   
  

• August 21 Serhiy Bykh (Seminar für Sprachwissenschaft, Universität Tübingen)
  Exploring Syntactic Features for Native Language Identification: A Variationist Perspective on Feature Encoding and Ensemble Optimization Abstract: In this talk, we systematically explore lexicalized and non-lexicalized local syntactic features for the task of Native Language Identification (NLI). We investigate different types of feature representations in single- and cross-corpus settings, including two representations inspired by a variationist perspective on the choices made in the linguistic system. To combine the different models, we use a probabilities-based ensemble classifier and propose a technique to optimize and tune it. Combining the best performing syntactic features with four types of n-grams outperforms the best approach of the NLI Shared Task 2013.

  (Joint work with Detmar Meurers to be presented at COLING 2014, August 23-29, 2014, Dublin, Ireland)  
   
  

_________________________________________________________________________________

Last updated: August 18, 2014