A code and dataset for the paper "Learning Representations for Soft Skill Matching" (the code includes PyTorch implementations of CNN, LSTM models with and without concatenated soft skill vectors).
Please find Hierarchical Attention Network implementation here: https://github.com/EdGENetworks/anuvada
Paper can be downloaded from Springer https://link.springer.com/chapter/10.1007/978-3-030-11027-7_15 or Arxiv https://arxiv.org/pdf/1807.07741.pdf
Due to the fact that the original code was improved further, the results coming from current implementation are different than in the paper. Early stopping was modified and dataset was slightly processed.
In case you want to download a large resource of soft skills similar to what used in the paper, please download publicly available list from here: https://datorium.gesis.org/xmlui/handle/10.7802/1707
We have as well new journal paper "Responsible team players wanted: an analysis of soft skill requirements in job advertisements" which provides the details of creating a resource of soft skills as well as contains a deeper analysis of soft skills on earnings: https://epjdatascience.springeropen.com/articles/10.1140/epjds/s13688-019-0190-z