Cluster of Excellence CoLi DFKI Computer Science MPI

Forum Entwicklung und Anwendung von Sprach-Technologien

FEAST is a series of lunch time talks devoted to all aspects of automatic language and speech processing. The series is open to everybody at the university and affiliated institutions such as DFKI and MPI. It is intended as a forum where researchers can informally present their work and get feedback from colleagues outside their immediate circle of collaborators. It will also provide an overview of the various lines of language and speech research carried out across the campus. We welcome reports on work in progress and hope that the series will foster inter-group discussion and lead to new collaborations.

Next meetings

Building A.22, room 2.01
Dr. Kilu von Prince
Annotating tense and modality in Oceanic
Our knowledge about tense and modality in the languages of the world is primarily based on languages that have been described as tense-prominent. Many Oceanic languages have long been described as mood-prominent, but until recently there was not enough empirical data to understand precisely how meaningful this label is or what it entails. In the MelaTAMP project, I am investigating seven small corpora from Oceanic languages of Melanesia that have been created during various documentation projects to gain more insight into their tense and modality features. In tagging different modal contexts, we decided to depart from the tradition of using modal flavors such as deontic and epistemic as the main values for our tags, since they do not appear to drive differential marking of modal meanings in these languages. Instead, we use the three-way distinction between actual, possible and counterfactual contexts as our main modal categories, in combination with tense, aspect and polarity features. While this annotation scheme has been designed for Oceanic languages, we believe that it may also be a practical and informative tool for the annotation of unrelated languages and would welcome any feedback.

Building A2 2, Department of Language Science and Technology, Room 213
Dr. Annemarie Verkerk
Current and future work in phylogenetic approaches to language history and diversity

Human languages are characterized by great amount of linguistic diversity but also by a great amount of interdependence through genealogical relations and geographic affinities. In the past, linguistic typology has attempted to avoid the confound introduced by these interdependencies by selective sampling, taking only a single language from a set of related languages or area. Phylogenetic methods offer another possibility to deal with non-independence of languages. Phylogenetic tree sets, which are built through computational analysis of cross-linguistic datasets, can be used to account for the genealogical relations found in a sample of related languages when doing typology.
In this talk, I give an overview of my recent work using these methods. Two recent projects deal with building phylogenetic tree sets. The first is an exploration of a method that can estimate two different tree structures at the same time on existing datasets from four linguistic families: Austronesian, Chinese, Indo-European, and Japonic. The second is a new phylogenetic analysis of the Dravidian language family.
Three other projects are typological in nature. The first is an analysis of Bantu noun class systems, showing which nominal affixes can be reconstructed on a Proto-Bantu level and which are most likely to be lost over time. The analysis demonstrates a clear influence of semantics on the loss and maintenance of morphology. At the other end of the world, I am involved with a project on numerals in the Pacific. Across the Oceanic languages, many have shifted from the ancestral decimal system to various types of quinary systems. We investigate whether this has been caused by contact with Papuan languages. Lastly, I am working on directional change in existential negation, attempting to find out if phylogenetic comparative methods find the same cyclical pathways that linguists find.
These projects range from being recently submitted for publication to representing the first steps in getting interpretable results, so any feedback is very welcome.

Google calendar representation

We now update a Google calendar with current talks, as well as show titles in the table below.

For those of you who want a provider-independent solution, at least in user-interface terms, we have iCal (ICS) and XML links for you for import into iCalendar, Lightning, etc.

Previous and upcoming talks

02-06-2017Dr. Kilu von PrinceAnnotating tense and modality in Oceanic
29-05-2017Dr. Annemarie Verkerk Current and future work in phylogenetic approaches to language history and diversity
08-05-2017Dr. Richard Eckart de CastilhoTowards an Infrastructure for the Distributed Exploration and Annotation of Large Corpora and Knowledge Bases
03-05-2017Denis PapernoOn typological closeness and semantic vector spaces
06-03-2017Jet Hoek Inter-coder agreement for annotation tasks: AC1 as an alternative to Kappa
13-02-2017Frances YungPredicting Discourse Markedness as Rational Speech Acts
13-12-2016Prof. Dr. Augustin SpeyerNew Directions@Saarland: Information Density and Information Structure as factors for information outsourcing
12-12-2016Prof. Dr. Remus GergelNew Directions@Saarland: From decompositional elements to presuppositions
05-12-2016Prof. Dr. Alexander KollerNew Directions@Saarland: Reference in interaction
24-11-2016Prof. Dr. Axel MecklingerNew Directions@Saarland: Unitization encoding and schema knowledge can reduce information density during learning
15-11-2016Viviana CotikNegation detection in clinical reports written in German
09-11-2016Prof. Dr. Joerg Hoffmann"New Directions@Saarland": Artificial Intelligence Planning and its Application to Search Problems in NLP
30-09-2016Mirjana SekickiCosts and Gains of Using Visual Context for Referent Prediction
29-09-2016Alejandra Quiros-RamirezBody talk: crowdshaping realistic 3D avatars with words
21-09-2016Sameer PradhanAbstract Meaning Representation (AMR) -- What is IT? How well can current algorithms reproduce IT?
15-09-2016Carina SilbererGrounding Language in the Visual World
18-08-2016Dat Quoc NguyenModeling topics and knowledge bases with embeddings
01-07-2016Katja HäuserThe Neuro-Cognitive Bases of Idiom Processing
29-06-2016Jesús CalvilloConnectionist Semantic Systematicity in Language Production
27-06-2016Aine ItoEvaluating the role of prediction during comprehension: How robust are the prediction effects?
01-06-2016Moreno CocoPerformance in a Collaborative Search Task: The Role of Feedback and Alignment
11-05-2016Michael RothImproving computational models of semantics using discourse
03-03-2016Rajakrishnan RajkumarInvestigating Locality Effects and Surprisal in Written English Syntactic Choice Phenomena
16-12-2015Catharine OertelTowards building a Conversational Speech Synthesiser
09-12-2015Eran RavehAutomated Generation of Customizable Recall Questions
07-12-2015Tal Linzen How might entropy affect comprehension difficulty?
23-11-2015Daniel McDonaldShifting discourse-semantics of risk in US newspapers, 1987–2014
12-11-2015Matti VarjokallioSpeech recognition of morphologically rich languages
04-11-2015Lisa BeinbornPredicting the difficulty of exercises for language learners
21-10-2015Tomas VelozToward a Quantum Theory of Cognition: History, Development and Perspectives
20-10-2015Nathan SchneiderDealing with off-the-charts polysemy in English prepositions: representation, annotation, and disambiguation
23-09-2015Sigrid KlerkeComparing representations of readers’ eye movements for natural language processing.
10-09-2015Kleo-Isidora MavridouLinking discourse modes and situation entity types in a cross-linguistic corpus study
28-07-2015Marten van SchijndelControlling for confounds in online measures of sentence complexity
22-07-2015Ani NenkovaComputational analysis of writing style and text quality, with practical applications to automatic summarization
21-07-2015Te RutherfordNeural Networks for Discourse Parsing
16-07-2015Annemarie FriedrichDiscourse-sensitive Automatic Identification of Generic Expressions
14-07-2015Dave HowcroftGermanic adjectival morphology and information theory
02-07-2015Marten van SchijndelIdentifying EEG Measures of Memory and Attentional Load in Language Processing
17-06-2015Artuur LeeuwenbergExploring Pattern Structures of Syntactic Trees for Relation
03-06-2015Ottokar TilkNeural networks for semantic role-filler prediction
07-05-2015Jonathan PoitzUsing Shallow Syntactic Features to Measure Influences of L1 and Proficiency Level in EFL Writings
14-04-2015Hannah KermesInformation Density and Scientific Literacy in English - Preliminary Analyses Using Language Modeling
14-04-2015Francesca DeloguScript-based surprisal: evidence from event-related potentials
13-04-2015Zofia MaliszThe relationship between information rate and speech rate in several European languages
13-04-2015Vera DembergOn the information conveyed by discourse connectives
09-04-2015Ekaterina KravtchenkoThe processing of predictable events in a script context
17-02-2015Ottokar TilkNeural networks for language modeling
02-02-2015Ines RehbeinProcessing urban youth language
30-01-2015Anna KazantsevaHierarchical Topical Segmentation Using Affinity Propagation
21-01-2015Benjamin RothKnowledge base completion using compositional vector space models
09-01-2015Motaz SaadMining Documents and Opinions in Cross-lingual Context
03-12-2014Annelen BrunnerAutomatic recognition of speech, thought and writing representation (ST&WR) in German narrative texts
12-11-2014Gerold HintzSemantic parsing using Lojban
19-08-2014Annemarie FriedrichAnnotation and automatic classification of situation entity types
12-08-2014Michael Fell and Caroline SporlederLyrics-based Analysis and Classification of Music
04-08-2014Sébastien Le MaguerSynthesize French using the statistical speech synthesis system HTS -- a complete study
03-07-2014Rachel RudingerLearning Scripts from Thematically Related Texts
17-06-2014Ashutosh Modi, Ivan TitovInducing Neural Models of Script Knowledge
17-06-2014Fatemeh Torabi Asr, Jonathan Sonntag, Yulia Grishna, Manfred StedeConceptual and Practical Steps in Event Coreference Analysis of Large-scale Data
11-06-2014Guy Emerson, Liling Tan, Susanne Fertmann, Alexis Palmer, and Michaela RegneriSeedLing: Building and using a seed corpus for the Human Language Project
11-06-2014Marc Schulder, Eduard HovyMetaphor detection through term relevance
21-05-2014Michaela Regneri, Rui Wang, and Manfred PinkalAligning Predicate-Argument Structures for Paraphrase Fragment Extraction
21-05-2014Andrea Horbach, Alexis Palmer and Magdalena WolskaFinding a Tradeoff between Accuracy and Rater's Workload in Grading Clustered Short Answers
20-05-2014Marcos ZampieriVarClass: An Open Source Language Identification Tool for Language Varieties
20-05-2014Annemarie Friedrich, Marina Valeeva and Alexis PalmerLQVSumm: A corpus of linguistic quality violations in multi-document summarization
07-05-2014Anjana Vakil and Alexis PalmerCross-language mapping for small-vocabulary ASR in under-resourced languages: investigating the impact of source language choice
28-03-2014Jordan Boyd-GraberBesting the quiz master: crowdsourcing incremental classification games
27-03-2014Alessandra ZarconeEvent knowledge and models of logical metonymy interpretation
15-01-2014Yannis KonstasJoint models for concept-to-text generation
24-10-2013Vera DembergMeasuring linguistically-induced cognitive load during driving using the ConTRe task
11-09-2013Hans-Ulrich KriegerAn Efficient Implementation of Equivalence Relations in OWL via Rule and Query Rewriting
19-06-2013Abdelhadi SoudiStandard Arabic (SA)-to-Moroccan Sign Language (MSL) machine translation: results and challenges
13-06-2013Kevin DuhLearning with multiple objectives
06-03-2013Ndapandula NakasholeA taxonomy of relational patterns and its use for semantic typing of emerging entities
13-02-2013Fabrizio NunnariThe ATLAS Project: Automatic translation of digital media for the hearing-impaired
16-01-2013Aline VillavicencioMultiword Expressions in language technology and language acquisition
12-12-2012Katrin MenzelCohesive ellipsis in English and German and their equivalents in a translation corpus - a quantitative contrastive analysis
23-11-2012Kate?ina VeselovskáAutomatic polarity classification systems for Czech
21-11-2012Ulrich SchäferCreating a cow from a hamburger: Extracting text and logical document structure from scientific papers in PDF format
27-09-2012Heriberto CuayáhuitlA Conversational Humanoid Robot with Multiple Coordinated Behaviours
12-09-2012Mihael ArcanExperiments in Term TranslationSlides
04-09-2012Fatemeh Torabi AsrTendency for Causality in Implicit Discourse Relations
04-07-2012Talk 1: Vera Demberg, Asad Sayeed, Philip Gorinski and Nikolaos Engonopoulos
Talk 2: Lea Frermann and Francis Bond
Talk 1: Syntactic Surprisal Affects Spoken Word Duration in Conversational Contexts

Talk 2: Cross-lingual Parse Disambiguation based on Semantic Correspondence
09-05-2012Diana McCarthyCompositionality modeling and non-compositionality detection with distributional semanticsSlides
08-02-2012Ulrich Schaefer, Christian Spurk et al.The 'final' ACL Anthology Searchbench and details about the upcoming ACL-2012 Workshop and its Contributed Task
18-01-2012Nikolina Koleva Ontology-based iterative detection of characters and their recognition in folktales
07-12-2011Mihaela Vela (DFKI) Extraction of Ontology Schema Components from Financial News
16-11-2011Martin Popel (Charles University, Prague) TectoMT - machine translation systemSlides
11-05-2011Angeliki Lazaridou (Coli)ELS: A Word-Level Method for Entity-Level Sentiment Analysis (Dry-run for 'Web Intelligence, Mining and Semantics 2011')
24-11-2010Alexis Heloir (DFKI) and Andrew Gargett (MMCI)Framing Signs
25-10-2010Minwoo JeongMulti-document Topic Segmentation
18-10-2010David Mare?ek, Charles University, PragueDependency tree projection across parallel textsSlides
11-08-2010Thierry Declerck, Antonia Scheidel and Piroska LendvaiProppian Content Descriptors in an Augmented Annotation Schema for Fairy Tales (LaTeCH dry-run)Slides
04-08-2010Andreas Schwarte, Christopher Haccius, Sebastian Steenbuck and Sven SteudterUsability Enhancement by Mining, Processing and Visualizing Data from the Federal German Archive. (LaTeCH dry-run)
28-07-2010Andrei Beliankou, Uni TrierTowards a German frame based QA System Slides
21-07-2010Yevgeni Berzak, Michal Richter and Carsten Ehrler, CoLi/CS/BioInfSimilarity-Based Navigation in Visualized Collections of Historical Documents (LaTeCH dry run)Slides
07-07-2010Grzegorz Chrupala and Afra AlishaiOnline Entropy-based Model of Lexical Category AcquisitionSlides
02-07-2010Michaela Regneri; Ivan Titov; Konstantina GaroufiACL dry runsSlides
26-05-2010Alexandre Klementiev, Johns Hopkins UniversityUnsupervised Prediction Aggregation
12-05-2010Bart Cramer, COLIConstraining robust constructions for broad-coverage parsing with precision grammarsSlides
28-04-2010Linlin Li, MMCIUsing Gaussian Mixture Models to Detect Figurative Language in ContextSlides
21-04-2010Saeedeh Momtazi, LSVLanguage Model-based Sentence Retrieval for Question Answering: a Trained Triggering Approach
17-02-2010Ivan Titov, CoEExploiting Weak Supervision in NLP tasks: Application to Sentiment Summarization
16-12-2009Berry Claus, CoLi'For one hour', 'One year later' et al.: Melting frozen time in narrative text comprehensionSlides
09-12-2009Djamé Seddah, Alpage (Inria) & Université Paris-Sorbonne (Paris 4)Cross parser evaluation and impact of annotation schemesSlides
25-11-2009Igor Farkaš, Comenius University, BratislavaInvestigating systematicity in the linear RAAM neural networkSlides
18-11-2009Cécile GrivazHuman judgements on causation in French textsSlides
04-11-2009Antske Fokkens, CoLiThe LinGO Grammar Matrix Customization SystemSlides
28-10-2009Arnab Goshal, LSVWeb-Derived PronunciationsSlides
21-10-2009Sebastian Germesin, DFKIAgreement Detection in Multiparty ConversationSlides
14-10-2009Ingrid Zukerman, Monash University, MelbourneTowards the Interpretation of Utterance Sequences in a Dialogue SystemSlides
30-09-2009Sebastian Germesin, DFKIAutomatic Speech Disfluency Detection and Correction in Multi-party InteractionSlides
23-09-2009Isabel Segura Bedmar and César de Pablo Sánchez, UC3MExtracting drug-drug interactions from pharmacological textsSlides
16-09-2009Michael Feld, DFKIAutomatic Recognition of Speaker Age in an Inter-cultural ContextSlides
09-09-2009Ines Rehbein, CoLiAssessing the bene?ts of partial automatic pre-labeling for frame-semantic annotationSlides
22-07-2009Afra Alishahi, CoLi, and Grzegorz Chrupa?a, LSVLexical Category Acquisition as an Incremental ProcessSlides
08-07-2009Linlin Li, CoLiClassifier Combination for Contextual Idiom Detection Without Labelled DataSlides
01-07-2009Roser Morante, CNTSProcessing the scope of negation and hedge cues in biomedical textsSlides
24-06-2009Matt Lease, LSVBeyond keywords: finding information more accurately and easily using natural language Slides
03-06-2009Grzegorz Chrupa?a, LSVHarnessing Raw Text for Learning Named EntitiesSlides
27-05-2009Alexis Palmer, CoEComputational linguistics and language documentation: Evaluating automation strategies for documenting endangered languagesSlides
06-05-2009Eva Lasarcyk, CoLiMachine vs. Human: A Cross-Discipline Study on Synthetic Speaker Age RecognitionSlides
22-04-2009Michael Wiegand, SLSFeature Design for Polarity ClassificationSlides
25-03-2009Hagen Fürstenau, CoLi, Pierre Lison, DFKI, Caroline Sporleder, CoEEACL dry-run talks
18-02-2009Maria Staudte, CoLiVisual attention in human-robot interactionSlides
11-02-2009Valia Kordoni & Yi Zhang, DFKI + CoLiAutomated Acquisition of Linguistic Knowledge for Robust Multilingual Grammar DevelopmentSlides Slides
04-02-2009Georgiana Dinu, CoLiInference rules for textual entailmentSlides
28-01-2009Thierry Declerck, DFKITowards Cross-Media Information ExtractionSlides
21-01-2009Mihaela Vela, DFKIConcept and Relation Extraction in the Finance DomainSlides
17-12-2008Afra Alishahi, CoLiA probabilistic model of cross-situational word learning from noisy and ambiguous dataSlides
10-12-2008Christian Müller, DFKISpeaker Classification: Supervector Approach and Detection TaskSlides
03-12-2008Berry Claus, CoLiEmbodied Language Processing: Desiderative Sentence Mood and Approach/Avoidance ActionsSlides
26-11-2008Hagen Fürstenau, CoLiSemi-Supervised Semantic Role LabelingSlides
14-01-2008Fabian Suchanek, MPIYAGO: A Core of Semantic KnowledgeSlides
07-01-2008Michael Wirth, CoLiMeasuring frame relatednessSlides

Mailing list

To receive notifications about future talks and events you may want to consider subscribing to our mailing list Of course, we respect your privacy and will not hand out your email address to third parties. You can unsubscribe from the list at any time.
Note: When signing up please use your UdS / DFKI / MPI email address to help us to identify legitimate subscriptions and avoid spam.


Interested in giving a talk yourself? If so, or in case you have any questions, suggestions or praise, please contact the organizers.

Former team members: