The field of active learning and materials design is moving towards more efficient and robust methods for optimizing performance and reducing computational costs. Researchers are exploring the use of large language models, granular-ball structures, and surrogate-based active learning to improve the accuracy and effectiveness of active learning strategies. Additionally, there is a growing interest in developing methods for estimating the Bayes error rate and optimizing functional materials design. Noteworthy papers include: No Free Lunch in Active Learning: LLM Embedding Quality Dictates Query Strategy Success, which establishes a benchmark for evaluating the influence of LLM embedding quality on query strategies in deep active learning. GAdaBoost: An Efficient and Robust AdaBoost Algorithm Based on Granular-Ball Structure, which proposes a novel two-stage framework for enhancing efficiency and robustness under noisy conditions. Optimization of Functional Materials Design with Optimal Initial Data in Surrogate-Based Active Learning, which investigates the optimal initial data sizes required for efficient convergence across various design space sizes.