Automatic scoring and feedback tools have become critical components of online learning proliferation. These tools range from multiple-choice questions to grading essays using machine learning (ML). Learning environments such as massive open online courses (MOOCs) would not be possible without them. The usage of this mechanism has brought many exciting areas of study, from the design of questions to the ML grading tools’ precision and accuracy. This paper analyzes the findings of 125 studies published in journals and proceedings between 2016 and 2020 on the usages of automatic scoring and feedback as a learning tool. This analysis gives an overview of the trends, challenges, and open questions in this research area. The results indicate that automatic scoring and feedback have many advantages. The most important benefits include enabling scaling the number of students without adding a proportional number of instructors, improving the student experience by reducing the time between submission grading and feedback, and removing bias in scoring. On the other hand, these technologies have some drawbacks. The main problem is creating a disincentive to develop innovative answers that do not match the expected one or have not been considered when preparing the problem. Another drawback is potentially training the student to answer the question instead of learning the concepts. With this, given the existence of a correct answer, such an answer could be leaked to the internet, making it easier for students to avoid solving the problem. Overall, each of these drawbacks presents an opportunity to look at ways to improve technologies to use these tools to provide a better learning experience to students.

Evaluating web traffic on a web server is highly critical for web service providers since, without a proper demand forecast, customers could have lengthy waiting times and abandon that website. However, this is a challenging task since it requires making reliable predictions based on the arbitrary nature of human behavior. We introduce an architecture that collects source data and in a supervised way performs the forecasting of the time series of the page views. Based on the Wikipedia page views dataset proposed in a competition by Kaggle in 2017, we created an updated version of it for the years 2018–2020. This dataset is processed and the features and hidden patterns in data are obtained for later designing an advanced version of a recurrent neural network called Long Short-Term Memory. This AI model is distributed training, according to the paradigm called data parallelism and using the Downpour training strategy. Predictions made for the seven dominant languages in the dataset are accurate with loss function and measurement error in reasonable ranges. Despite the fact that the analyzed time series have fairly bad patterns of seasonality and trend, the predictions have been quite good, evidencing that an analysis of the hidden patterns and the features extraction before the design of the AI model enhances the model accuracy. In addition, the improvement of the accuracy of the model with the distributed training is remarkable. Since the task of predicting web traffic in as precise quantities as possible requires large datasets, we designed a forecasting system to be accurate despite having limited data in the dataset. We tested the proposed model on the new Wikipedia page views dataset we created and obtained a highly accurate prediction; actually, the mean absolute error of predictions regarding the original one on average is below 30. This represents a significant step forward in the field of time series prediction for web traffic forecasting.