Some highlights from AAMAS 2018 and ICML 2018: We’ve presented a number of papers: Learning with Opponent-Learning Awareness @ AAMAS. Evaluating Generalization in Multiagent Systems using Agent-Interaction Graphs (short) @ AAMAS. Learning Policy Representations in Multiagent Systems @ ICML. DiCE: The Infinitely Differentiable Monte-Carlo Estimator @ ICML. Code for LOLA and DiCE has been released. Check it out on GitHub.