Grounded action transformation for sim-to-real reinforcement learning

Reinforcement learning in simulation is a promising alternative to the prohibitive sample cost of reinforcement learning in the physical world. Unfortunately, policies learned in simulation often perform worse than hand-coded policies when applied on the target, physical system. Grounded simulation learning (gsl) is a general framework that promises to address this issue by altering the simulator to better match the real world (Farchy et al. 2013 in Proceedings of the 12th international conference on autonomous agents and multiagent systems (AAMAS)). This article introduces a new algorithm for gsl—Grounded Action Transformation (GAT)—and applies it to learning control policies for a humanoid robot. We evaluate our algorithm in controlled experiments where we show it to allow policies learned in simulation to transfer to the real world. We then apply our algorithm to learning a fast bipedal walk on a humanoid robot and demonstrate a 43.27% improvement in forward walk velocity compared to a state-of-the art hand-coded walk. This striking empirical success notwithstanding, further empirical analysis shows that gat may struggle when the real world has stochastic state transitions. To address this limitation we generalize gat to the stochasticgat (sgat) algorithm and empirically show that sgat leads to successful real world transfer in situations where gat may fail to find a good policy. Our results contribute to a deeper understanding of grounded simulation learning and demonstrate its effectiveness for applying reinforcement learning to learn robot control policies entirely in simulation.

Evan Mawarire’s #ThisFlag as Tactical Lyric: The Role of Digital Speech in Imagining a Networked Zimbabwean Nation

Evan Mawarire’s poetic video “This Flag,” first posted on Facebook on April 20, 2016, mobilized an international protest movement against then-president of Zimbabwe Robert Mugabe between April and September of 2016. In the video, Mawarire built on the poetics of anti-colonial resistance and nationalization to create a rallying cry. The piece’s remediation through the hashtag channel #ThisFlag created rhetorical links between digital organizing and grounded action. This literary perspective on contemporary discussions of social media and collective identity formation shows how the poetic elements of the video enabled Mawarire’s claim to spread and motivate a grounded movement.

Studying Human-Centered IT Innovation Using a Grounded Action Learning Approach

This paper describes how two research methodologies, grounded theory and action learning, were combined to produce a rigorous yet creative and flexible method for field study of a recent IT-based innovation, virtual teams. Essentially, an action learning program was used to train facilitators of virtual teams and generate research data while grounded theory techniques were used to analyze and interpret the data. This paper shows how this combined method can be used to develop local and practical theory for complex, human-centered areas of information technology. The implications of this grounded action learning approach for practice and research in IS will be discussed.