The field of machine learning is moving towards developing more robust and reliable methods for data valuation and model training. Recent research has focused on improving the efficiency and accuracy of algorithms for distinct element estimation, Byzantine robust aggregation, and data valuation. New parameterizations and protocols have been introduced to reduce communication complexity and improve the performance of these algorithms. Additionally, there is a growing interest in developing methods for explainable and transparent data valuation, such as Data Valuation Cards and counterfactual explanations of Shapley value. Notable papers in this area include: A Practical and Secure Byzantine Robust Aggregator, which presents a quasi-linear time algorithm for robust aggregation with near-optimal bias bounds. KAIROS: Scalable Model-Agnostic Data Valuation, which introduces a scalable framework for model-agnostic data valuation with a closed-form solution for influence scores.