Interactive Intelligence group has big success at AAMAS 2021

We are happy to announce that in this years iteration of AAMAS, the flagship conference on (multi)agent systems, our Interactive Intelligence group had a big success!

In the Blue Sky Ideas Track we co-authored the paper ‘Responsibility Research for Trustworthy Autonomous Systems’, which explores research gaps for upcoming social, technological, ethical and legal challenges.
In the same track our exploration of multiagent systems for non-stationary machine learning titled ‘Environment Shift Games: Are Multiple Agents the Solution, and not the Problem, to Non-Stationarity?’ got a Special Mention coupled to a $500 travel grant!

Similarly, our paper ‘Difference Rewards Policy Gradients’, presented as an extended abstract, won the best paper award in the Adaptive and Learning Agents workshop!

In the main track we (co)-authored three papers, making contributions to diverse topics by ‘Axies: Identifying and Evaluating Context-Specific Values’, providing ‘Loss Bounds for Approximate Influence-Based Abstraction’ and extending Cross-Entropy methods through ‘Learning Complex Policy Distribution with CEM Guided Adversarial Hypernetwork’.

Our paper ‘A Collaborative Platform for Identifying Context-Specific Values’, published in the demonstration track, provides a collaborative web platform that eases the process of identifying context-specific values with the help of AI techniques.

All in all we think that the conference was a success, albeit not being live. We made good use of the gathertown venue and had many nice discussion, in particular in the informal sessions!