A Neural Approach to Automated Essay Scoring A Neural Approach to Automated Essay Scoring
Paper summary The authors construct a neural network for automated essay scoring. https://i.imgur.com/XTWGpmy.png Convolution window of 3 is passed over the text, which is used as input to an LSTM. The output of the LSTM is averaged over all timesteps and then a single value in the range of [0,1] is predicted as a scaled-down score for the essay. They evaluate by measuring quadratic weighted Kappa on the Kaggle essay scoring dataset.

Summary by Marek Rei 3 years ago
Your comment:

ShortScience.org allows researchers to publish paper summaries that are voted on and ranked!

Sponsored by: and