Formation control of multiple Unmanned Aerial Vehicles (UAVs) is vital for practical applications. This paper tackles the task of behavior-based UAV formation while avoiding static and dynamic obstacles during directed flight. We present a two-stage reinforcement learning (RL) training pipeline to tackle the challenge of multi-objective optimization, large exploration spaces, and the sim-to-real gap. The first stage searches in a simplified scenario for a linear utility function that balances all task objectives simultaneously, whereas the second stage applies the utility function in complex scenarios, utilizing curriculum learning to navigate large exploration spaces. Additionally, we apply an attention-based observation encoder to enhance formation maintenance and manage varying obstacle quantity. Experiments in simulation and real world demonstrate that our method outperforms planning-based and RL-based baselines regarding collision-free rate and formation maintenance in scenarios with static, dynamic, and mixed obstacles.

A longstanding goal of artificial general intelligence is highly capable generalists that can learn from diverse experiences and generalize to unseen tasks. The language and vision communities have seen remarkable progress toward this trend by scaling up transformer-based models trained on massive datasets, while reinforcement learning (RL) agents still suffer from poor generalization capacity under such paradigms. To tackle this challenge, we propose Meta Decision Transformer (Meta-DT), which leverages the sequential modeling ability of the transformer architecture and robust task representation learning via world model disentanglement to achieve efficient generalization in offline meta-RL. We pretrain a context-aware world model to learn a compact task representation, and inject it as a contextual condition to the causal transformer to guide task-oriented sequence generation. Then, we subtly utilize history trajectories generated by the meta-policy as a self-guided prompt to exploit the architectural inductive bias. We select the trajectory segment that yields the largest prediction error on the pretrained world model to construct the prompt, aiming to encode task-specific information complementary to the world model maximally. Notably, the proposed framework eliminates the requirement of any expert demonstration or domain knowledge at test time. Experimental results on MuJoCo and Meta-World benchmarks across various dataset types show that Meta-DT exhibits superior few and zero-shot generalization capacity compared to strong baselines while being more practical with fewer prerequisites. Our code is available at https://github.com/NJU-RL/Meta-DT.

Understanding species dynamics in heterogeneous environments is essential for ecosystem studies. Traditional models assumed homogeneous habitats, but recent approaches include spatial and temporal variability, highlighting species migration. We adopt starvation-driven diffusion (SDD) models as nonlinear diffusion to describe species dispersal based on local resource conditions, showing advantages for species survival. However, accurate prediction remains challenging due to model simplifications. This study uses multi-agent reinforcement learning (MARL) with deep Q-networks (DQN) to simulate single species and predator-prey interactions, incorporating SDD-type rewards. Our simulations reveal evolutionary dispersal strategies, providing insights into species dispersal mechanisms and validating traditional mathematical models.

This work presents the experiments and solution outline for our teams winning submission in the Learn To Race Autonomous Racing Virtual Challenge 2022 hosted by AIcrowd. The objective of the Learn-to-Race competition is to push the boundary of autonomous technology, with a focus on achieving the safety benefits of autonomous driving. In the description the competition is framed as a reinforcement learning (RL) challenge. We focused our initial efforts on implementation of Soft Actor Critic (SAC) variants. Our goal was to learn non-trivial control of the race car exclusively from visual and geometric features, directly mapping pixels to control actions. We made suitable modifications to the default reward policy aiming to promote smooth steering and acceleration control. The framework for the competition provided real time simulation, meaning a single episode (learning experience) is measured in minutes. Instead of pursuing parallelisation of episodes we opted to explore a more traditional approach in which the visual perception was processed (via learned operators) and fed into rule-based controllers. Such a system, while not as academically "attractive" as a pixels-to-actions approach, results in a system that requires less training, is more explainable, generalises better and is easily tuned and ultimately out-performed all other agents in the competition by a large margin.

In complex real-world tasks such as robotic manipulation and autonomous driving, collecting expert demonstrations is often more straightforward than specifying precise learning objectives and task descriptions. Learning from expert data can be achieved through behavioral cloning or by learning a reward function, i.e., inverse reinforcement learning. The latter allows for training with additional data outside the training distribution, guided by the inferred reward function. We propose a novel approach to construct compact and transparent reward models from automatically selected state features. These inferred rewards have an explicit form and enable the learning of policies that closely match expert behavior by training standard reinforcement learning algorithms from scratch. We validate our method's performance in various robotic environments with continuous and high-dimensional state spaces. Webpage: \url{https://sites.google.com/view/transparent-reward}.

Pricing algorithms have demonstrated the capability to learn tacit collusion that is largely unaddressed by current regulations. Their increasing use in markets, including oligopolistic industries with a history of collusion, calls for closer examination by competition authorities. In this paper, we extend the study of tacit collusion in learning algorithms from basic pricing games to more complex markets characterized by perishable goods with fixed supply and sell-by dates, such as airline tickets, perishables, and hotel rooms. We formalize collusion within this framework and introduce a metric based on price levels under both the competitive (Nash) equilibrium and collusive (monopolistic) optimum. Since no analytical expressions for these price levels exist, we propose an efficient computational approach to derive them. Through experiments, we demonstrate that deep reinforcement learning agents can learn to collude in this more complex domain. Additionally, we analyze the underlying mechanisms and structures of the collusive strategies these agents adopt.

A 2021 paper in Nature by Mirhoseini et al.30 about the use of reinforcement learning (RL) in the physical design of silicon chips raised eyebrows, drew critical media coverage, and stirred up controversy due to poorly documented claims. The paper, authored by Google researchers, withheld critical methodological steps, and most inputs needed to reproduce its results. Our meta-analysis shows how two separate evaluations filled in the gaps and demonstrated that Google RL lags behind human chip designers, a well-known algorithm (simulated annealing), and generally available commercial software, while also being slower. Crosschecked data indicates that the integrity of the Nature paper is substantially undermined, owing to errors in conduct, analysis, and reporting. Before publishing, Google rebuffed internal allegations of fraud which still stand.

Effective Traffic Signal Control (TSC) is fundamental to urban traffic management, responsible for guiding the movement of vehicles through intersections by controlling traffic lights. The primary goals of TSC are to minimize traffic congestion, enhance traffic flow, and improve safety for both vehicles and pedestrians. Poor TSC optimization leads to increased congestion, fuel consumption, and pollution. Longer wait times at signals lead to increased fuel consumption, which not only exacerbates environmental issues through higher emissions but also results in economic losses due to delays. Moreover, inefficient TSC negatively impacts the quality of life in urban areas, contributing to increased noise and air pollution.

Reinforcement learning has achieved promising results on robotic control tasks but struggles to leverage information effectively from multiple sensory modalities that differ in many characteristics. Recent works construct auxiliary losses based on reconstruction or mutual information to extract joint representations from multiple sensory inputs to improve the sample efficiency and performance of reinforcement learning algorithms. However, the representations learned by these methods could capture information irrelevant to learning a policy and may degrade the performance. We argue that compressing information in the learned joint representations about raw multimodal observations is helpful, and propose a multimodal information bottleneck model to learn task-relevant joint representations from egocentric images and proprioception. Our model compresses and retains the predictive information in multimodal observations for learning a compressed joint representation, which fuses complementary information from visual and proprioceptive feedback and meanwhile filters out task-irrelevant information in raw multimodal observations. We propose to minimize the upper bound of our multimodal information bottleneck objective for computationally tractable optimization. Experimental evaluations on several challenging locomotion tasks with egocentric images and proprioception show that our method achieves better sample efficiency and zero-shot robustness to unseen white noise than leading baselines. We also empirically demonstrate that leveraging information from egocentric images and proprioception is more helpful for learning policies on locomotion tasks than solely using one single modality.

Model-free reinforcement learning (RL) is inherently a reactive method, operating under the assumption that it starts with no prior knowledge of the system and entirely depends on trial-and-error for learning. This approach faces several challenges, such as poor sample efficiency, generalization, and the need for well-designed reward functions to guide learning effectively. On the other hand, controllers based on complete system dynamics do not require data. This paper addresses the intermediate situation where there is not enough model information for complete controller design, but there is enough to suggest that a model-free approach is not the best approach either. By carefully decoupling known and unknown information about the system dynamics, we obtain an embedded controller guided by our partial model and thus improve the learning efficiency of an RL-enhanced approach. A modular design allows us to deploy mainstream RL algorithms to refine the policy. Simulation results show that our method significantly improves sample efficiency compared with standard RL methods on continuous control tasks, and also offers enhanced performance over traditional control approaches. Experiments on a real ground vehicle also validate the performance of our method, including generalization and robustness.