Concept

Results (Using deep reinforcement learning for personalizing review sessions on e-learning platforms with spaced repetition)

The results of the study on using deep reinforcement learning for personalizing review sessions with spaced repetition explore five key areas: (1) the relation between rewards and thresholds, (2) the performance of the deep reinforcement learning (DRL) agent when the number of items is varied, (3) a comparison of the performance of the TRPO and TNPG algorithms, (4) a comparison between likelihood-based and average sum of outcomes-based reward functions, and (5) the performance of TRPO with reward shaping.

0

1

Updated 2026-07-04

Tags

Data Science