Labeling the Semantic Roles of Commas

Authors: Naveen Arivazhagan, Christos Christodoulopoulos, Dan Roth

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental This paper proposes a set of relations commas participate in, expanding on previous work in this area, and develops a new dataset annotated with this set of labels. We identify features that are important to achieve a good performance on comma labeling and then develop a machine learning method that achieves high accuracy on identifying comma relations, improving over previous work. Finally, we build a simple model to learn these new relations and outperform previous systems.
Researcher Affiliation Academia Department of Computer Science University of Illinois at Urbana-Champaign {arivazh2, christod, danr}@illinois.edu
Pseudocode No The paper describes the methods and features used for the comma classifier, but it does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statement about making the source code for their methodology publicly available, nor does it include a link to a code repository.
Open Datasets Yes We build on the corpus previously annotated by (Srikumar et al. 2008) by refining existing relations and adding new ones. The dataset is available at: http://cogcomp.cs.illinois.edu/page/publication_view/780
Dataset Splits Yes In all experiments, 5-fold cross-validation is used to evaluate the performance of the learned classifier specified in the previous subsection.
Hardware Specification No The paper does not explicitly describe the hardware specifications (e.g., specific CPU, GPU models, or memory) used to run the experiments.
Software Dependencies No The paper mentions several tools used, such as 'Illinois POS tagger', 'Illinois Shallow Parser', 'Charniak parser', 'Illinois Named Entity Tagger', and 'LBJava', but it does not provide specific version numbers for these software dependencies.
Experiment Setup Yes We build this classifier by training on a combination of the corpus and comma-syntax-pattern annotations we produced, using a Sparse Averaged Perceptron (Jackson and Craven 1996) with LBJava (Rizzolo and Roth 2010) trained over 160 rounds (learning rate = 0.024, thickness = 3.9).