LITERATER

LITERATER is focused on essay grading automation to help drive improvements in literacy, addressing stagnation in literacy across United States by focusing state-of-the-art Natural Language Processing techniques to support automatic grading of passage-dependent essays. Using publicly available essay sets as well as constructing artificial essays from other natural language inference and comprehension datasets, we move beyond BERT and the quadratic bottleneck for attention sequence length through encoding sentence embeddings. Having already received stakeholder feedback from educators in Tennessee, we intend to have the tool for limited release to Tennessee teachers in January of 2021.