English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We address the problem of predicting the correctness of the student'sresponse on the next exam question based on their previous interactions in thecourse of their learning and evaluation process. We model the studentperformance as a dynamic problem and compare the two major classes of dynamicneural architectures for its solution, namely the finite-memory Time DelayNeural Networks (TDNN) and the potentially infinite-memory Recurrent NeuralNetworks (RNN). Since the next response is a function of the knowledge state ofthe student and this, in turn, is a function of their previous responses andthe skills associated with the previous questions, we propose a two-partnetwork architecture. The first part employs a dynamic neural network (eitherTDNN or RNN) to trace the student knowledge state. The second part applies ontop of the dynamic part and it is a multi-layer feed-forward network whichcompletes the classification task of predicting the student response based onour estimate of the student knowledge state. Both input skills and previousresponses are encoded using different embeddings. Regarding the skillembeddings we tried two different initialization schemes using (a) randomvectors and (b) pretrained vectors matching the textual descriptions of theskills. Our experiments show that the performance of the RNN approach is bettercompared to the TDNN approach in all datasets that we have used. Also, we showthat our RNN architecture outperforms the state-of-the-art models in four outof five datasets. It is worth noting that the TDNN approach also outperformsthe state of the art models in four out of five datasets, although it isslightly worse than our proposed RNN approach. Finally, contrary to ourexpectations, we find that the initialization of skill embeddings usingpretrained vectors offers practically no advantage over random initialization.

Student Performance Prediction Using Dynamic Neural Models