Decision Transformer (DT), as one of the representative Reinforcement Learning via Supervised Learning (RvS) methods, has achieved strong performance in offline learning tasks by leveraging the powerful Transformer architecture for sequential decision-making. However, in adversarial environments, these methods can be non-robust, since the return is dependent on the strategies of both the decision-maker and adversary. Training a probabilistic model conditioned on observed return to predict action can fail to generalize, as the trajectories that achieve a return in the dataset might have done so due to a suboptimal behavior adversary. To address this, we propose a worst-case-aware RvS algorithm, the Adversarially Robust Decision Transformer (ARDT), which learns and conditions the policy on in-sample minimax returns-to-go. ARDT aligns the target return with the worst-case return learned through minimax expectile regression, thereby enhancing robustness against powerful test-time adversaries. In experiments conducted on sequential games with full data coverage, ARDT can generate a maximin (Nash Equilibrium) strategy, the solution with the largest adversarial robustness. In large-scale sequential games and continuous adversarial RL environments with partial data coverage, ARDT demonstrates significantly superior robustness to powerful test-time adversaries and attains higher worst-case returns compared to contemporary DT methods.

Dissociative Identity Disorder (DID) is characterized by the presence of at least two distinct identities in the same individual. This paper describes a co-design process with a person living with DID. We first aimed to uncover the main challenges experienced by the co-designer as well as design opportunities for novel technologies. We then engaged in a prototyping stage to design a wearable display (WhoDID) to facilitate in-person social interactions. The prototype aims to be used as a necklace and enable the user to make their fronting personality visible to others. Thus, facilitating social encounters or sudden changes of identity. We reflect on the design features of WhoDID in the broader context of supporting people with DID. Moreover, we provide insights on co-designing with someone with multiple (sometimes conflicting) personalities regarding requirement elicitation, decision-making, prototyping, and ethics. To our knowledge, we report the first design process with a DID user within the ASSETS and CHI communities. We aim to encourage other assistive technology researchers to design with DID users.