It is widely known that policies trained using reinforcement learning (RL) to solve simulated robotics problems (MuJoCo) are extremely brittle and unstable, i.e. your solution will most likely break down after perturbing a bit (e.g. poking the robot) or transferring it to a similar task. It is often impossible to provide any safety guarantees for constraint satisfaction or an interpretation of how the trained policies work.
Question we address: How to develop physics-informed reinforcement learning algorithms that guarantee safety and interpretability ?
So-called Neural ODEs recently generated quite a hype in deep learning. The main idea is to formulate artificial neural networks as continuous and parametrized dynamical systems, trained using gradient descent like in deep learning. This has biological motivation, as clearly the neurons in the brain are continuous. One of the applications is to replace the deep segment of Resnet blocks with Neural ODE block, where many questions arise about the properties of the learned dynamical system.
Question we address: Study properties of dynamical systems arising from training neural ODE architectures?
Unity ML-Agents is one of the major open-source frameworks for reinforcement learning research. In our research, we have been using the newest environment, the Unity Dodgeball. We generalize the Fictitious Co-Play algorithm from Collaborating with Humans without Human Data (NeurIPS 2021) for multi-player & multi-team settings and evaluate it within the Dodgeball environment. For the purpose of performing Human-AI experiments, we built a hand-crafted FPP computer game mod of the Dodgeball environment, utilizing the Unity 3D game engine capabilities. More details can be found here.
Question we address: Can we train smart RL agents capable of collaborating with human players in 3D computer game environment?