The field of reinforcement learning is moving towards addressing non-stationarity and improving robustness in dynamic environments. Researchers are exploring new frameworks and algorithms that can adapt to changing conditions, such as non-stationary Markov decision processes and transportable causal bandits. These advancements have the potential to improve the performance of reinforcement learning agents in real-world applications. Noteworthy papers include: The Non-stationary and Varying-discounting Markov Decision Processes paper, which introduces a novel framework for reinforcement learning that accommodates non-stationarity and allows for flexible optimal policy shaping. The On Transportability for Structural Causal Bandits paper, which investigates the use of transportability to enhance learning in causal bandit settings. The Computing Strategic Responses to Non-Linear Classifiers paper, which presents a novel method for computing best responses in strategic classification settings with non-linear classifiers.