Training Data for Aspect-Controlled Neural Argument Generation 

Thumbnail
Count of file(s): 4
generated_arguments.7z (Arguments generated with the language model) (1.029MB) 
argument_aspect_detection_v1.0.7z (Argument Aspect Detection Corpus (v1.0)) (416.8KB) 
reddit_training_data_1.1.7z (Training data for CTRL_reddit) (240.2MB) 
cc_training_data_1.1.7z (Training data for CTRL_cc) (804.0MB) 
Date
Author
Type
Dataset
Metadata
Show full item record
Export
Text
BibTex
DataCite XML
JSON

Description

This collection holds (1) the argument aspect detection corpus, (2) the data to reproduce the training of the language models described in the paper, and (3) the arguments generated with both language models. The argument aspect detection corpus includes 5,032 arguments over eight controversial topics, annotated with aspects by crowdworkers. DISCLAIMER: All material provided in the archives reddit_training_data.7z and cc_training_data.7z may be used for research purposes only. For all other material in this collection (namely argument_aspect_detection_v1.0.7z and generated_arguments.7z), this restriction does not apply. The user acknowledges and agrees that the data is provided on an “as­-is” basis and that the licensor makes no representations or warranties of any kind.
 
 

Subject

argument generation;aspect detection;controlled argument generation

URI

Related third party funded projects

DFG | GU798/25-1 | Offenes Argument-Min
PTJ | 03VP02540 | ArgumenText

Related Resources

Collections

The following license files are associated with this item:

    In Copyright
    Except where otherwise noted, this item's license is described as In Copyright