The field of artificial intelligence in education is moving towards enhancing the reasoning capabilities of large language models (LLMs) and improving virtual learning environments. Researchers are exploring innovative methods to fine-tune LLMs, such as learning from errors and using reinforcement learning, to improve their performance in tasks like automatic math correction. Additionally, there is a growing focus on promoting equity in team-based learning and understanding the factors that influence interpersonal trust among students in virtual learning environments. Noteworthy papers in this area include:
- LEMMA, which proposes a method to enhance LLMs' reasoning ability by learning from errors, and
- Teaching LLMs for Step-Level Automatic Math Correction via Reinforcement Learning, which introduces a reinforcement learning-based method for step-level automatic math correction.