Learning physical dynamics in a series of non-stationary environments is a challenging but essential task for model-based reinforcement learning (MBRL) with visual inputs. It requires the agent to consistently adapt to novel tasks without forgetting previous knowledge. In this paper, we present a new continual learning approach for visual dynamics modeling and explore its efficacy in visual control and forecasting. The key assumption is that an ideal world model can provide a non-forgetting environment simulator, which enables the agent to optimize the policy in a multi-task learning manner based on the imagined trajectories from the world model. To this end, we first propose the mixture world model that learns task-specific dynamics priors with a mixture of Gaussians, and then introduce a new training strategy to overcome catastrophic forgetting, which we call predictive experience replay. Finally, we extend these methods to continual RL and further address the value estimation problems with the exploratory-conservative behavior learning approach. Our model remarkably outperforms the naive combinations of existing continual learning and visual RL algorithms on DeepMind Control and Meta-World benchmarks with continual visual control tasks. It is also shown to effectively alleviate the forgetting of spatiotemporal dynamics in video prediction datasets with evolving domains.

We propose a reinforcement-learning algorithm to tackle the challenge of reconstructing phylogenetic trees. The search for the tree that best describes the data is algorithmically challenging, thus, all current algorithms for phylogeny reconstruction use various heuristics to make it feasible. In this study, we demonstrate that reinforcement-learning can be used to learn an optimal search strategy, thus providing a novel paradigm for predicting the maximum-likelihood tree. Our proposed method does not require likelihood calculation with every step, nor is it limited to greedy uphill moves in the likelihood space. We demonstrate the use of the developed deep-Q-learning agent on a set of unseen empirical data, namely, on unseen environments defined by nucleotide alignments of up to 20 sequences. Our results show that the likelihood scores of the inferred phylogenies are similar to those obtained from widely-used software. It thus establishes a proof-of-concept that it is beneficial to optimize a sequence of moves in the search-space, rather than optimizing the progress made in every single move only. This suggests that a reinforcement-learning based method provides a promising direction for phylogenetic reconstruction.

Robust Markov decision processes (MDPs) aim to handle changing or partially known system dynamics. To solve them, one typically resorts to robust optimization methods. However, this significantly increases computational complexity and limits scalability in both learning and planning. On the other hand, regularized MDPs show more stability in policy learning without impairing time complexity. Yet, they generally do not encompass uncertainty in the model dynamics. In this work, we aim to learn robust MDPs using regularization. We first show that regularized MDPs are a particular instance of robust MDPs with uncertain reward. We thus establish that policy iteration on reward-robust MDPs can have the same time complexity as on regularized MDPs. We further extend this relationship to MDPs with uncertain transitions: this leads to a regularization term with an additional dependence on the value function. We then generalize regularized MDPs to twice regularized MDPs ($\text{R}^2$ MDPs), i.e., MDPs with $\textit{both}$ value and policy regularization. The corresponding Bellman operators enable us to derive planning and learning schemes with convergence and generalization guarantees, thus reducing robustness to regularization. We numerically show this two-fold advantage on tabular and physical domains, highlighting the fact that $\text{R}^2$ preserves its efficacy in continuous environments.

Current approaches to multi-agent cooperation rely heavily on centralized mechanisms or explicit communication protocols to ensure convergence. This paper studies the problem of distributed multi-agent learning without resorting to centralized components or explicit communication. It examines the use of distribution matching to facilitate the coordination of independent agents. In the proposed scheme, each agent independently minimizes the distribution mismatch to the corresponding component of a target visitation distribution. The theoretical analysis shows that under certain conditions, each agent minimizing its individual distribution mismatch allows the convergence to the joint policy that generated the target distribution. Further, if the target distribution is from a joint policy that optimizes a cooperative task, the optimal policy for a combination of this task reward and the distribution matching reward is the same joint policy. This insight is used to formulate a practical algorithm (DM$^2$), in which each individual agent matches a target distribution derived from concurrently sampled trajectories from a joint expert policy. Experimental validation on the StarCraft domain shows that combining (1) a task reward, and (2) a distribution matching reward for expert demonstrations for the same task, allows agents to outperform a naive distributed baseline. Additional experiments probe the conditions under which expert demonstrations need to be sampled to obtain the learning benefits.

Reinforcement learning (RL) has become a promising approach to developing controllers for quadrupedal robots. Conventionally, an RL design for locomotion follows a position-based paradigm, wherein an RL policy outputs target joint positions at a low frequency that are then tracked by a high-frequency proportional-derivative (PD) controller to produce joint torques. In contrast, for the model-based control of quadrupedal locomotion, there has been a paradigm shift from position-based control to torque-based control. In light of the recent advances in model-based control, we explore an alternative to the position-based RL paradigm, by introducing a torque-based RL framework, where an RL policy directly predicts joint torques at a high frequency, thus circumventing the use of a PD controller. The proposed learning torque control framework is validated with extensive experiments, in which a quadruped is capable of traversing various terrain and resisting external disturbances while following user-specified commands. Furthermore, compared to learning position control, learning torque control demonstrates the potential to achieve a higher reward and is more robust to significant external disturbances. To our knowledge, this is the first sim-to-real attempt for end-to-end learning torque control of quadrupedal locomotion.

Despite impressive progress, today's AI models are very inefficient learners, taking huge amounts of time and data to solve problems humans pick up almost instantaneously. A new approach could drastically speed things up by getting AI to read instruction manuals before attempting a challenge. One of the most promising approaches to creating AI that can solve a diverse range of problems is reinforcement learning, which involves setting a goal and rewarding the AI for taking actions that work towards that goal. This is the approach behind most of the major breakthroughs in game-playing AI, such as DeepMind's AlphaGo. As powerful as the technique is, it essentially relies on trial and error to find an effective strategy.

This work focuses on the persistent monitoring problem, where a set of targets moving based on an unknown model must be monitored by an autonomous mobile robot with a limited sensing range. To keep each target's position estimate as accurate as possible, the robot needs to adaptively plan its path to (re-)visit all the targets and update its belief from measurements collected along the way. In doing so, the main challenge is to strike a balance between exploitation, i.e., re-visiting previously-located targets, and exploration, i.e., finding new targets or re-acquiring lost ones. Encouraged by recent advances in deep reinforcement learning, we introduce an attention-based neural solution to the persistent monitoring problem, where the agent can learn the inter-dependencies between targets, i.e., their spatial and temporal correlations, conditioned on past measurements. This endows the agent with the ability to determine which target, time, and location to attend to across multiple scales, which we show also helps relax the usual limitations of a finite target set. We experimentally demonstrate that our method outperforms other baselines in terms of number of targets visits and average estimation error in complex environments. Finally, we implement and validate our model in a drone-based simulation experiment to monitor mobile ground targets in a high-fidelity simulator.

Multi-input multi-out and non-orthogonal multiple access (MIMO-NOMA) internet-of-things (IoT) systems can improve channel capacity and spectrum efficiency distinctly to support the real-time applications. Age of information (AoI) is an important metric for real-time application, but there is no literature have minimized AoI of the MIMO-NOMA IoT system, which motivates us to conduct this work. In MIMO-NOMA IoT system, the base station (BS) determines the sample collection requirements and allocates the transmission power for each IoT device. Each device determines whether to sample data according to the sample collection requirements and adopts the allocated power to transmit the sampled data to the BS over MIMO-NOMA channel. Afterwards, the BS employs successive interference cancelation (SIC) technique to decode the signal of the data transmitted by each device. The sample collection requirements and power allocation would affect AoI and energy consumption of the system. It is critical to determine the optimal policy including sample collection requirements and power allocation to minimize the AoI and energy consumption of MIMO-NOMA IoT system, where the transmission rate is not a constant in the SIC process and the noise is stochastic in the MIMO-NOMA channel. In this paper, we propose the optimal power allocation to minimize the AoI and energy consumption of MIMO- NOMA IoT system based on deep reinforcement learning (DRL). Extensive simulations are carried out to demonstrate the superiority of the optimal power allocation.

The spread of online misinformation threatens public health, democracy, and the broader society. While professional fact-checkers form the first line of defense by fact-checking popular false claims, they do not engage directly in conversations with misinformation spreaders. On the other hand, non-expert ordinary users act as eyes-on-the-ground who proactively counter misinformation -- recent research has shown that 96% counter-misinformation responses are made by ordinary users. However, research also found that 2/3 times, these responses are rude and lack evidence. This work seeks to create a counter-misinformation response generation model to empower users to effectively correct misinformation. This objective is challenging due to the absence of datasets containing ground-truth of ideal counter-misinformation responses, and the lack of models that can generate responses backed by communication theories. In this work, we create two novel datasets of misinformation and counter-misinformation response pairs from in-the-wild social media and crowdsourcing from college-educated students. We annotate the collected data to distinguish poor from ideal responses that are factual, polite, and refute misinformation. We propose MisinfoCorrect, a reinforcement learning-based framework that learns to generate counter-misinformation responses for an input misinformation post. The model rewards the generator to increase the politeness, factuality, and refutation attitude while retaining text fluency and relevancy. Quantitative and qualitative evaluation shows that our model outperforms several baselines by generating high-quality counter-responses. This work illustrates the promise of generative text models for social good -- here, to help create a safe and reliable information ecosystem. The code and data is accessible on https://github.com/claws-lab/MisinfoCorrect.

While combining imitation learning (IL) and reinforcement learning (RL) is a promising way to address poor sample efficiency in autonomous behavior acquisition, methods that do so typically assume that the requisite behavior demonstrations are provided by an expert that behaves optimally with respect to a task reward. If, however, suboptimal demonstrations are provided, a fundamental challenge appears in that the demonstration-matching objective of IL conflicts with the return-maximization objective of RL. This paper introduces D-Shape, a new method for combining IL and RL that uses ideas from reward shaping and goal-conditioned RL to resolve the above conflict. D-Shape allows learning from suboptimal demonstrations while retaining the ability to find the optimal policy with respect to the task reward. We experimentally validate D-Shape in sparse-reward gridworld domains, showing that it both improves over RL in terms of sample efficiency and converges consistently to the optimal policy in the presence of suboptimal demonstrations.