Artificial Intelligence in Education

Teachers often conduct surveys in order to collect data from a predefined group of students gain insights into topics interest. When analyzing with open-ended textual responses, it is extremely time-consuming, labor-intensive, and difficult manually process all the responses an insightful comprehensive report. In analysis step, traditionally, teacher has read each decide on how them extract information. Even though possible only using certain keywords, such approach would be limited since not fails account for embedded contexts but also cannot detect polysemous words or phrases semantics that are expressible single words. this work, we present novel end-to-end context-aware framework extracts, aggregates, abbreviates semantic patterns open-response survey data. Our relies pre-trained natural language model encode vectors. The encoded vectors then get clustered either optimally tuned number groups set pre-specified titles. former case, clusters further analyzed representative keywords summary sentences serve as labels clusters. our framework, designated clusters, finally provide wordclouds demonstrate semantically prominent within group. Honoring user privacy, have successfully built on-device implementation suitable real-time mobile devices tested synthetic dataset. reduces costs at-scale by automating extracting most information pieces