Jose A. Arjona-Medina

Julius Ott, Lorenzo Servadei, Jose Arjona-Medina et al.

Data selection is essential for any data-based optimization technique, such as Reinforcement Learning. State-of-the-art sampling strategies for the experience replay buffer improve the performance of the Reinforcement Learning agent. However, they do not incorporate uncertainty in the Q-Value estimation. Consequently, they cannot adapt the sampling strategies, including exploration and exploitation of transitions, to the complexity of the task. To address this, this paper proposes a new sampling strategy that leverages the exploration-exploitation trade-off. This is enabled by the uncertainty estimation of the Q-Value function, which guides the sampling to explore more significant transitions and, thus, learn a more efficient policy. Experiments on classical control environments demonstrate stable results across various environments. They show that the proposed method outperforms state-of-the-art sampling strategies for dense rewards w.r.t. convergence and peak performance by 26% on average.

Vihang P. Patil, Markus Hofmarcher, Marius-Constantin Dinu et al.

Reinforcement learning algorithms require many samples when solving complex hierarchical tasks with sparse and delayed rewards. For such complex tasks, the recently proposed RUDDER uses reward redistribution to leverage steps in the Q-function that are associated with accomplishing sub-tasks. However, often only few episodes with high rewards are available as demonstrations since current exploration strategies cannot discover them in reasonable time. In this work, we introduce Align-RUDDER, which utilizes a profile model for reward redistribution that is obtained from multiple sequence alignment of demonstrations. Consequently, Align-RUDDER employs reward redistribution effectively and, thereby, drastically improves learning on few demonstrations. Align-RUDDER outperforms competitors on complex artificial tasks with delayed rewards and few demonstrations. On the Minecraft ObtainDiamond task, Align-RUDDER is able to mine a diamond, though not frequently. Code is available at https://github.com/ml-jku/align-rudder. YouTube: https://youtu.be/HO-_8ZUl-UY

Jose A. Arjona-Medina, Michael Gillhofer, Michael Widrich et al.

We propose RUDDER, a novel reinforcement learning approach for delayed rewards in finite Markov decision processes (MDPs). In MDPs the Q-values are equal to the expected immediate reward plus the expected future rewards. The latter are related to bias problems in temporal difference (TD) learning and to high variance problems in Monte Carlo (MC) learning. Both problems are even more severe when rewards are delayed. RUDDER aims at making the expected future rewards zero, which simplifies Q-value estimation to computing the mean of the immediate reward. We propose the following two new concepts to push the expected future rewards toward zero. (i) Reward redistribution that leads to return-equivalent decision processes with the same optimal policies and, when optimal, zero expected future rewards. (ii) Return decomposition via contribution analysis which transforms the reinforcement learning task into a regression task at which deep learning excels. On artificial tasks with delayed rewards, RUDDER is significantly faster than MC and exponentially faster than Monte Carlo Tree Search (MCTS), TD(λ), and reward shaping approaches. At Atari games, RUDDER on top of a Proximal Policy Optimization (PPO) baseline improves the scores, which is most prominent at games with delayed rewards. Source code is available at \url{https://github.com/ml-jku/rudder} and demonstration videos at \url{https://goo.gl/EQerZV}.

Jose Arjona-Medina, Ramil Nugmanov

Despite the rapid and significant advancements in deep learning for Quantitative Structure-Activity Relationship (QSAR) models, the challenge of learning robust molecular representations that effectively generalize in real-world scenarios to novel compounds remains an elusive and unresolved task. This study examines how atom-level pretraining with quantum mechanics (QM) data can mitigate violations of assumptions regarding the distributional similarity between training and test data and therefore improve performance and generalization in downstream tasks. In the public dataset Therapeutics Data Commons (TDC), we show how pretraining on atom-level QM improves performance overall and makes the activation of the features distributes more Gaussian-like which results in a representation that is more robust to distribution shifts. To the best of our knowledge, this is the first time that hidden state molecular representations are analyzed to compare the effects of molecule-level and atom-level pretraining on QM data.

Markus Holzleitner, Lukas Gruber, José Arjona-Medina et al.

We prove under commonly used assumptions the convergence of actor-critic reinforcement learning algorithms, which simultaneously learn a policy function, the actor, and a value function, the critic. Both functions can be deep neural networks of arbitrary complexity. Our framework allows showing convergence of the well known Proximal Policy Optimization (PPO) and of the recently introduced RUDDER. For the convergence proof we employ recently introduced techniques from the two time-scale stochastic approximation theory. Our results are valid for actor-critic methods that use episodic samples and that have a policy that becomes more greedy during learning. Previous convergence proofs assume linear function approximation, cannot treat episodic examples, or do not consider that policies become greedy. The latter is relevant since optimal policies are typically deterministic.