UniER Benchmark Unifies Item-Level and Path-Level Exercise Recommendation

Researchers have unveiled UniER (Unified Benchmark for Exercise Recommendation), an all-encompassing evaluation framework that integrates two previously distinct approaches in personalized exercise recommendation: Item-Level Exercise Recommendation (ILER) and Path-Level Exercise Recommendation (PLER). While ILER focuses on optimizing immediate single-step transitions, PLER aims to create cohesive learning paths for enhanced cumulative benefits. Although both methods strive to align educational resources with individual knowledge acquisition, differing evaluation structures have complicated unified benchmarking and fair assessments. UniER fills this void by offering a shared platform for evaluating both methodologies. A significant advancement is the Weighted Cognitive Gain (WCG) metric, which assesses algorithmic performance across paradigms. This framework allows for direct comparisons between ILER and PLER, promoting advancements in personalized education. The findings are published in a paper on arXiv (2605.16750).