Multi-agent reinforcement learning (MARL) has found various applications in the field of transportation and simulating [50, 1], stock price analyzing and trading [32, 31], wireless communication networks [12, 11, 13], and learning behaviors in social dilemmas [33, 28, 34]. MARL, however, becomes intractable due to the complex interactions among agents as the number of agents increases. A recent tractable approach is a mean-field approach by considering MARL in the regime with a large number of homogeneous agents under weak interactions [20]. According to the number of agents and learning goals, there are three subtle types of mean-field theories for MARL. The first one is called mean-field MARL (MF-MARL), which refers to the empirical average of the states or actions of a finite population. For example, [52] proposes to approximate interactions within the population of agents by averaging the actions of the overall population or neighboring agents.

We shall have a hard look at ethics and try to extract insights in the form of abstract properties that might become tools. We want to connect ethics to games, talk about the performance of ethics, introduce curiosity into the interplay between competing and coordinating in well-performing ethics, and offer a view of possible developments that could unify increasing aggregates of entities. All this is under a long shadow cast by computational complexity that is quite negative about games. This analysis is the first step toward finding modeling aspects that might be used in AI ethics for integrating modern AI systems into human society.

On an assigned graph, the problem of Multi-Agent Pathfinding (MAPF) consists in finding paths for multiple agents, avoiding collisions. Finding the minimum-length solution is known to be NP-hard, and computation times grows exponentially with the number of agents. However, in industrial applications, it is important to find feasible, suboptimal solutions, in a time that grows polynomially with the number of agents. Such algorithms exist for undirected and biconnected directed graphs. Our main contribution is to generalize these algorithms to the more general case of strongly connected directed graphs. In particular, given a MAPF problem with at least two holes, we present an algorithm that checks the problem feasibility in linear time with respect to the number of nodes, and provides a feasible solution in polynomial time.

We show that in a cooperative $N$-agent network, one can design locally executable policies for the agents such that the resulting discounted sum of average rewards (value) well approximates the optimal value computed over all (including non-local) policies. Specifically, we prove that, if $|\mathcal{X}|, |\mathcal{U}|$ denote the size of state, and action spaces of individual agents, then for sufficiently small discount factor, the approximation error is given by $\mathcal{O}(e)$ where $e\triangleq \frac{1}{\sqrt{N}}\left[\sqrt{|\mathcal{X}|}+\sqrt{|\mathcal{U}|}\right]$. Moreover, in a special case where the reward and state transition functions are independent of the action distribution of the population, the error improves to $\mathcal{O}(e)$ where $e\triangleq \frac{1}{\sqrt{N}}\sqrt{|\mathcal{X}|}$. Finally, we also devise an algorithm to explicitly construct a local policy. With the help of our approximation results, we further establish that the constructed local policy is within $\mathcal{O}(\max\{e,\epsilon\})$ distance of the optimal policy, and the sample complexity to achieve such a local policy is $\mathcal{O}(\epsilon^{-3})$, for any $\epsilon>0$.

Lately, with the wide adoption of Deep Learning techniques for compact representations of value functions and policies in model-free problems [16, 21, 34], the field of Multi-Agent Reinforcement Learning (MARL) has seen an explosion in the applications of such algorithms to solve real-world problems [19]. However, this has naturally led to a trend where both the amount of data handled in such data driven approaches and the complexity of the targeted problems grow exponentially. In a MARL setting where communication between agents is required, this may inevitably lead to restrictive requirements in the frequency and reliability of the communication to and from each agents (as it was already pointed out in [23]). The effect of asynchronous communication in dynamic programming problems was studied already in [2]. In particular, one of the first examples of how communication affects learning and policy performance in MARL is found in [31], where the author investigates the impact of agents sharing different combinations of state variable subsets or Q values.

Training multiple agents to perform safe and cooperative control in the complex scenarios of autonomous driving has been a challenge. For a small fleet of cars moving together, this paper proposes Lepus, a new approach to training multiple agents. Lepus adopts a pure cooperative manner for training multiple agents, featured with the shared parameters of policy networks and the shared reward function of multiple agents. In particular, Lepus pre-trains the policy networks via an adversarial process, improving its collaborative decision-making capability and further the stability of car driving. Moreover, for alleviating the problem of sparse rewards, Lepus learns an approximate reward function from expert trajectories by combining a random network and a distillation network. We conduct extensive experiments on the MADRaS simulation platform. The experimental results show that multiple agents trained by Lepus can avoid collisions as many as possible while driving simultaneously and outperform the other four methods, that is, DDPG-FDE, PSDDPG, MADDPG, and MAGAIL(DDPG) in terms of stability.

The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJACI-ECAI 2022) took place from 23-29 July, in Vienna. The title of her talk was "Engineering sociality and collaboration in AI systems". Robots are widely used in industrial settings, but what happens when they enter our everyday world, and, specifically, social situations? Ana believes that social robots, chatbots and social agents have the potential to change the way we interact with technology. She envisages a hybrid society where humans and AI systems work in tandem.

The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJACI-ECAI 2022) took place from 23-29 July, in Vienna. In this post, we continue our round-up of the invited talks, summarising the presentation by Ana Paiva, University of Lisbon and INESC-ID. The title of her talk was "Engineering sociality and collaboration in AI systems". Robots are widely used in industrial settings, but what happens when they enter our everyday world, and, specifically, social situations? Ana believes that social robots, chatbots and social agents have the potential to change the way we interact with technology.

In Multi-Agent Reinforcement Learning, communication is critical to encourage cooperation among agents. Communication in realistic wireless networks can be highly unreliable due to network conditions varying with agents' mobility, and stochasticity in the transmission process. We propose a framework to learn practical communication strategies by addressing three fundamental questions: (1) When: Agents learn the timing of communication based on not only message importance but also wireless channel conditions. (2) What: Agents augment message contents with wireless network measurements to better select the game and communication actions. (3) How: Agents use a novel neural message encoder to preserve all information from received messages, regardless of the number and order of messages. Simulating standard benchmarks under realistic wireless network settings, we show significant improvements in game performance, convergence speed and communication efficiency compared with state-of-the-art.

We study cooperative online learning in stochastic and adversarial Markov decision process (MDP). That is, in each episode, $m$ agents interact with an MDP simultaneously and share information in order to minimize their individual regret. We consider environments with two types of randomness: \emph{fresh} -- where each agent's trajectory is sampled i.i.d, and \emph{non-fresh} -- where the realization is shared by all agents (but each agent's trajectory is also affected by its own actions). More precisely, with non-fresh randomness the realization of every cost and transition is fixed at the start of each episode, and agents that take the same action in the same state at the same time observe the same cost and next state. We thoroughly analyze all relevant settings, highlight the challenges and differences between the models, and prove nearly-matching regret lower and upper bounds. To our knowledge, we are the first to consider cooperative reinforcement learning (RL) with either non-fresh randomness or in adversarial MDPs.