Miguel Ballesteros

 

See my website for more information: Website

 

Miguel Ballesteros is a postdoc at Universitat Pompeu Fabra.

Miguel's main research interest relies on applying new methods for automatic feature selection in order to speed up and facilitate the process of optimizing a data-driven dependency parser. Besides this, he is carrying out several experiments by modifying the work-flow and behaviour of current transition-based and graph-based dependency parsing algorithms with the intention of enhancing the performance and speeding up the training process.

News

  • April 2012. Release of MaltOptimizer: Miguel and Joakim Nivre have released MaltOptimizer, a tool for MaltParser optimization and automatic feature selection. They demonstrated the system at EACL 2012 and Miguel presented it at LREC 2012.

Contact

    Research Interests

    • Dependency Parsing.
    • Data-Driven Dependency Parsing.
    • Machine Learning for Natural Language Processing.
    • Algorithms for natural language processing.
    • Processing negation and modality.
    • Simplified Texts.
    • Design principles and annotation schemes for treebanks.

    Publications

    Projects

    Scientific Software

    Activities