TTLab – Text Technology Lab

The TTLab (Text Technology Lab), headed by Prof. Alexander Mehler, is part of the Department of Computer Science and Mathematics (Fachbereich Informatik und Mathematik) at the Goethe Universität in Frankfurt. It investigates formal, algorithmic models to deepen our understanding of information processing in the humanities. We examine diachronic, time-dependent as well as synchronic aspects of processing linguistic and non-linguistic, multimodal signs. The Lab works across several disciplines to bridge between computer science on the one hand and corpus-based research in the humanities on the other. To this end, we develop information models and algorithms for the analysis of texts, images, and other objects relevant to research in the humanities.

News

  • New publications published in the special issue New Review of Hypermedia and Multimedia

    by

    The following publications have been published in the special issue New Review of Hypermedia and Multimedia:

    Viki LibraRy: Collaborative Hypertext Browsing and Navigation in Virtual Reality

    Kevin Bönisch, Alexander Mehler, Shaduan Babbili, Yannick Heinrich, Philipp Stephan and Giuseppe Abrami. 2024. Viki LibraRy: Collaborative Hypertext Browsing and Navigation in Virtual Reality. New Review of Hypermedia and Multimedia, 0(0):1–31.
    BibTeX
    @article{Boenisch:et:al:2024:b,
      author    = {B\"{o}nisch, Kevin and Mehler, Alexander and Babbili, Shaduan
                   and Heinrich, Yannick and Stephan, Philipp and Abrami, Giuseppe},
      abstract  = {We present Viki LibraRy, a dynamically built library in virtual
                   reality (VR) designed to visualize hypertext systems, with an
                   emphasis on collaborative interaction and spatial immersion. Viki
                   LibraRy goes beyond traditional methods of text distribution by
                   providing a platform where users can share, process, and engage
                   with textual information. It operates at the interface of VR,
                   collaborative learning and spatial data processing to make reading
                   tangible and memorable in a spatially mediated way. The article
                   describes the building blocks of Viki LibraRy, its underlying
                   architecture, and several use cases. It evaluates Viki LibraRy
                   in comparison to a conventional web interface for text retrieval
                   and reading. The article shows that Viki LibraRy provides users
                   with spatial references for structuring their recall, so that
                   they can better remember consulted texts and their meta-information
                   (e.g. in terms of subject areas and content categories)},
      title     = {{Viki LibraRy: Collaborative Hypertext Browsing and Navigation
                   in Virtual Reality}},
      journal   = {New Review of Hypermedia and Multimedia},
      volume    = {0},
      number    = {0},
      pages     = {1--31},
      year      = {2024},
      publisher = {Taylor \& Francis},
      doi       = {10.1080/13614568.2024.2383581},
      url       = {https://doi.org/10.1080/13614568.2024.2383581},
      eprint    = {https://doi.org/10.1080/13614568.2024.2383581}
    }

    Geo-spatial Hypertext in Virtual Reality: Mapping and Navigating Global News Event Spaces

    Patrick Schrottenbacher, Alexander Mehler, Theresa Berg, Jasper Hustedt, Julian Gagel, Timo Lüttig and Giuseppe Abrami. 2024. Geo-spatial hypertext in virtual reality: mapping and navigating global news event spaces. New Review of Hypermedia and Multimedia, 0(0):1–30.
    BibTeX
    @article{Schrottenbacher:et:al:2024,
      author    = {Schrottenbacher, Patrick and Mehler, Alexander and Berg, Theresa
                   and Hustedt, Jasper and Gagel, Julian and Lüttig, Timo and Abrami, Giuseppe},
      title     = {Geo-spatial hypertext in virtual reality: mapping and navigating
                   global news event spaces},
      journal   = {New Review of Hypermedia and Multimedia},
      volume    = {0},
      number    = {0},
      pages     = {1--30},
      year      = {2024},
      publisher = {Taylor \& Francis},
      doi       = {10.1080/13614568.2024.2383601},
      url       = {https://doi.org/10.1080/13614568.2024.2383601},
      eprint    = {https://doi.org/10.1080/13614568.2024.2383601},
      abstract  = {Every day, a myriad of events take place that are documented and
                   shared online through news articles from a variety of sources.
                   As a result, as users navigate the Web, the volume of data can
                   lead to information overload, making it difficult to find specific
                   details about an event. We present News in Time and Space (NiTS)
                   to address this issue: NiTS is a fully immersive system integrated
                   into Va.Si.Li-Lab that organises textual information in a geospatial
                   hypertext system in virtual reality. With NiTS, users can visualise,
                   filter and interact with information currently based on GDELT
                   on a virtual globe providing document networks to analyse global
                   events and trends. The article describes NiTS, its event semantics
                   and architecture. It evaluates NiTS in comparison to a classic
                   search engine website, extended by NiTSs information filtering
                   capabilities to make it comparable. Our comparison with this website
                   technology, which is directly linked to the user's usage habits,
                   shows that NiTS enables comparable information exploration even
                   if the users have little or no experience with VR. That is, we
                   observe an equivalent search result behaviour, but with the advantage
                   that VR allows users to get their results with a higher level
                   of usability without distracting them from their tasks. Through
                   its integration with Va.Si.Li-Lab, a simulation-based learning
                   environment, NiTS can be used in simulations of learning processes
                   aimed at studying critical online reasoning, where Va.Si.Li-Lab
                   guarantees that this can be done in relation to individual or
                   groups of learners.}
    }

  • Two new papers at SemDial 2024 — TrentoLogue

    by

    The Semantics and Pragmatics of Dialogue, September 11th – 12th, 2024

    On gesture semantics:

    Andy Lücking, Alexander Mehler and Alexander Henlein. 2024. The Linguistic Interpretation of Non-emblematic Gestures Must be agreed in Dialogue: Combining Perceptual Classifiers and Grounding/Clarification Mechanisms. Proceedings of the 28th Workshop on The Semantics and Pragmatics of Dialogue.
    BibTeX
    @inproceedings{Luecking:Mehler:Henlein:2024-classifier,
      title     = {The Linguistic Interpretation of Non-emblematic Gestures Must
                   be agreed in Dialogue: Combining Perceptual Classifiers and Grounding/Clarification
                   Mechanisms},
      author    = {Lücking, Andy and Mehler, Alexander and Henlein, Alexander},
      year      = {2024},
      booktitle = {Proceedings of the 28th Workshop on The Semantics and Pragmatics of Dialogue},
      series    = {SemDial'24 -- TrentoLogue},
      location  = {Università di Trento, Palazzo Piomarta, Rovereto},
      url       = {https://www.semdial.org/anthology/papers/Z/Z24/Z24-4031/},
      pdf       = {http://semdial.org/anthology/Z24-Lucking_semdial_0031.pdf}
    }

    On brain-based semantics:

    Jonathan Ginzburg, Chris Eliasmith and Andy Lücking. 2024. Swann's name: Towards a Dialogical Brain Semantics. Proceedings of the 28th Workshop on The Semantics and Pragmatics of Dialogue.
    BibTeX
    @inproceedings{Ginzburg:Eliasmith:Luecking:2024-swann,
      title     = {Swann's name: {Towards} a Dialogical Brain Semantics},
      author    = {Ginzburg, Jonathan and Eliasmith, Chris and Lücking, Andy},
      year      = {2024},
      booktitle = {Proceedings of the 28th Workshop on The Semantics and Pragmatics of Dialogue},
      series    = {SemDial'24 -- TrentoLogue},
      location  = {Università di Trento, Palazzo Piomarta, Rovereto},
      url       = {https://www.semdial.org/anthology/papers/Z/Z24/Z24-3007/},
      pdf       = {http://semdial.org/anthology/Z24-Ginzburg_semdial_0007.pdf}
    }
  • New Publication Accepted for the 2nd Workshop on Legal Information Retrieval meets AI (LIRAI24)

    by

    Our paper, “Finding Needles in Emb(a)dding Haystacks: Legal Document Retrieval via Bagging and SVR Ensembles,” has been accepted to the 2nd Workshop on Legal Information Retrieval Meets AI. In this work, we present an approach that leverages embedding spaces, bootstrap aggregation, and SVR ensembles to retrieve legal passages efficiently, demonstrating improved recall compared to baseline methods (0.849 > 0.803 | 0.829):

    Kevin Bönisch and Alexander Mehler. 2024. Finding Needles in Emb(a)dding Haystacks: Legal Document Retrieval via Bagging and SVR Ensembles. Proceedings of the 2nd Legal Information Retrieval meets Artificial Intelligence Workshop LIRAI 2024. accepted.
    BibTeX
    @inproceedings{Boenisch:Mehler:2024,
      title     = {Finding Needles in Emb(a)dding Haystacks: Legal Document Retrieval
                   via Bagging and SVR Ensembles},
      author    = {B\"{o}nisch, Kevin and Mehler, Alexander},
      year      = {2024},
      booktitle = {Proceedings of the 2nd Legal Information Retrieval meets Artificial
                   Intelligence Workshop LIRAI 2024},
      location  = {Poznan, Poland},
      publisher = {CEUR-WS.org},
      address   = {Aachen, Germany},
      series    = {CEUR Workshop Proceedings},
      note      = {accepted},
      abstract  = {We introduce a retrieval approach leveraging Support Vector Regression
                   (SVR) ensembles, bootstrap aggregation (bagging), and embedding
                   spaces on the German Dataset for Legal Information Retrieval (GerDaLIR).
                   By conceptualizing the retrieval task in terms of multiple binary
                   needle-in-a-haystack subtasks, we show improved recall over the
                   baselines (0.849 > 0.803 | 0.829) using our voting ensemble, suggesting
                   promising initial results, without training or fine-tuning any
                   deep learning models. Our approach holds potential for further
                   enhancement, particularly through refining the encoding models
                   and optimizing hyperparameters.},
      keywords  = {legal information retrieval, support vector regression, word embeddings, bagging ensemble}
    }

Sign up to our mailing list to receive news updates.