United Kingdom

Centralized Training for Decentralized Execution, where agents are trained offline using centralized information but execute in a decentralized manner online, has gained popularity in the multi-agent reinforcement learning community. In particular, actor-critic methods with a centralized critic and decentralized actors are a common instance of this idea. However, the implications of using a centralized critic in this context are not fully discussed and understood even though it is the standard choice of many algorithms. We therefore formally analyze centralized and decentralized critic approaches, providing a deeper understanding of the implications of critic choice. Because our theory makes unrealistic assumptions, we also empirically compare the centralized and decentralized critic methods over a wide set of environments to validate our theories and to provide practical advice. We show that there exist misconceptions regarding centralized critics in the current literature and show that the centralized critic design is not strictly beneficial, but rather both centralized and decentralized critics have different pros and cons that should be taken into account by algorithm designers.

AAMAS 2021

Contrasting Centralized and Decentralized Critics in Multi-Agent Reinforcement Learning

AAMAS is the largest and most influential conference in the area of agents and multiagent systems, bringing together researchers and practitioners in all areas of agent technology and providing and internationally renowned high-profile forum for publishing and finding out about the latest developments in the field. AAMAS is the flagship conference of the non-profit International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS).  

![](https://assets.underline.io/uploads/markdown_image/1/image/ce4604bd54468a006164d16e080204fb.jpg)

Please click on the button below to register for this event.

If you are registered and see this page please check the account email you are logged in with. It must be the same as for the registration.

If you still have trouble accessing the contact please contact support at aamas2021@underline.io 

technical paper

When solving a complex task, humans will spontaneously form teams and to complete different parts of the whole task, respectively. Meanwhile, the cooperation between teammates will improve efficiency. However, for current cooperative MARL methods, the cooperation team is constructed through either heuristics or end-to-end blackbox optimization. In order to improve the efficiency of cooperation and exploration, we propose a structured diversification emergence MARL framework named 
 based on reinforced organization control and hierarchical consensus learning.  
 first learns an adaptive grouping policy through the organization control module, which is established by independent multi-agent reinforcement learning.  Further, the hierarchical consensus module based on the hierarchical intentions with consensus constraint is introduced after team formation.  Simultaneously, utilizing the hierarchical consensus module and a self-supervised intrinsic reward enhanced decision module, the proposed cooperative MARL algorithm 
 can output the final diversified multi-agent cooperative policy.

Structured Diversification Emergence via Reinforced Organization Control and Hierachical Consensus Learning

Value-based methods of multi-agent reinforcement learning (MARL), especially the value decomposition methods, have been demonstrated on a range of challenging cooperative tasks. However, current methods pay little attention to the interaction between agents, which is essential to teamwork in games or real life. This limits the efficiency of value-based MARL algorithms in the two aspects: collaborative exploration and value function estimation.  In this paper, we propose a novel cooperative MARL algorithm named as interactive actor-critic (IAC), which models the interaction of agents from the perspectives of policy and value function.  On the policy side, a multi-agent joint stochastic policy is introduced by adopting a collaborative exploration module, which is trained by maximizing the entropy-regularized expected return. On the value side, we use the shared attention mechanism to estimate the value function of each agent, which takes the impact of the teammates into consideration. At the implementation level, we extend the value decomposition methods to continuous control tasks and evaluate IAC on benchmark tasks including classic control and multi-agent particle environments. Experimental results indicate that our method outperforms the state-of-the-art approaches and achieves better performance in terms of cooperation.

Modeling the Interaction between Agents in Cooperative Multi-Agent Reinforcement Learning

Training a multi-agent reinforcement learning (MARL) algorithm is more challenging than training a single-agent reinforcement learning algorithm, because the result of a multi-agent task strongly depends on the complex interactions among agents and their interactions with a stochastic and dynamic environment. We propose an algorithm that boosts MARL training using the biased action information of other agents based on a friend-or-foe concept. For a cooperative and competitive environment, there are generally two groups of agents: cooperative-agents and competitive-agents. In the proposed algorithm, each agent updates its value function using its own action and the biased action information of other agents in the two groups. The biased joint action of cooperative agents is computed as the sum of their actual joint action and the imaginary cooperative joint action, by assuming all the cooperative agents jointly maximize the target agent's value function. The biased joint action of competitive agents can be computed similarly. Each agent then updates its own value function using the biased action information, resulting in a biased value function and corresponding biased policy. Subsequently, the biased policy of each agent is inevitably subjected to recommend an action to cooperate and compete with other agents, thereby introducing more active interactions among agents and enhancing the MARL policy learning. We empirically demonstrate that our algorithm outperforms existing algorithms in various mixed cooperative-competitive environments. Furthermore, the introduced biases gradually decrease as the training proceeds and the correction based on the imaginary assumption vanishes.

Cooperative and Competitive Biases for Multi-Agent Reinforcement Learning

Wind farms are a crucial driver toward the generation of ecological and renewable energy. Due to their rapid increase in capacity, contemporary wind farms need to adhere to strict constraints on power output to ensure stability of the electricity grid. Specifically, a wind farm controller is required to match the farm's power production with a power demand imposed by the grid operator. This is a non-trivial optimization problem, as complex dependencies exist between the wind turbines.  State-of-the-art wind farm control typically relies on physics-based heuristics that fail to capture the full load spectrum that defines a turbine's health status. When this is not taken into account, the long-term viability of the farm's turbines is put at risk.  Given the complex dependencies that determine a turbine's lifetime, learning a flexible and optimal control strategy requires a data-driven approach. However, as wind farms are large-scale multi-agent systems, optimizing control strategies over the full joint action space is intractable. We propose a new learning method for wind farm control that leverages the sparse wind farm structure to factorize the optimization problem. Using a Bayesian approach, based on multi-agent Thompson sampling, we explore the factored joint action space for configurations that match the demand, while considering the lifetime of turbines. We apply our method to a grid-like wind farm layout, and evaluate configurations using a state-of-the-art wind flow simulator. Our results are competitive with a physics-based heuristic approach in terms of demand error, while, contrary to the heuristic, our method prolongs the lifetime of high-risk turbines.

Scalable Optimization for Wind Farm Control using Coordination Graphs

A key problem for Belief-Desire-Intention (BDI) agents is <i>intention progression, i.e., which plans should be selected and how the execution of these plans should be interleaved so as to achieve the agent's goals. Monte-Carlo Tree Search (MCTS) has been shown to be a promising approach to the intention progression problem, out-performing other approaches in the literature. However, MCTS relies on runtime simulation of possible interleavings of the plans in each intention, which may be computationally costly. In this paper, we introduce the notion of quantitative summary information</i> which can be used to estimate the likelihood of conflicts between an agent's intentions. We show how offline simulation can be used to precompute quantitative summary information prior to execution of the agent's program, and how the precomputed summary information can be used at runtime to guide the expansion of the MCTS search tree and avoid unnecessary runtime simulation. We compare the performance of our approach with standard MCTS in a range of scenarios of increasing difficulty. The results suggest our approach can significantly improve the efficiency of MCTS in terms of the number of runtime simulations performed.

Intention Progression using Quantitative Summary Information

Embodied conversational agents (ECAs) benefit from non-verbal behavior for natural and efficient interaction with users. Gesticulation -- hand and arm movements accompanying speech -- is an essential part of non-verbal behavior. Gesture generation models have been developed for several decades: starting with rule-based and ending with mainly data-driven methods. To date, recent end-to-end gesture generation methods have not been evaluated in a real-time interaction with users. We present a proof-of-concept framework, which is intended to facilitate evaluation of modern gesture generation models in interaction.

A Framework for Integrating Gesture Generation Models into Interactive Conversational Agents

The introduction of driver-less technologies can improve on-demand transport (ODT) systems and help make passenger transportation and logistics more efficient. Here, we aim to provide a generic model of the online ODT with autonomous vehicles problem and a multi-agent model specific to resource allocation and scheduling in vehicle fleets. Our model considers autonomous vehicles that communicate via peer-to-peer radio channels to meet passenger requirements and satisfy trip requests in an online ODT system. We experiment this model with several allocation mechanisms (mathematical programming, greedy heuristic, distributed constraint optimization, and auctions) and compare their performance on synthetic scenarios on a real-world city road network.

A Generic Multi-Agent Model for Resource Allocation Strategies in Online On-Demand Transport with Autonomous Vehicles

Automatic gesture generation from speech generally relies on implicit modelling of the nondeterministic speech-gesture relationship and can result in averaged motion lacking defined form. Here, we propose a database-driven approach of selecting gestures based on specific motion characteristics that have been shown to be associated with the speech audio. We extend previous work that identified expressive parameters of gesture motion that can both be predicted from speech and are perceptually important for a good speech-gesture match, such as gesture velocity and finger extension. A perceptual study was performed to evaluate the appropriateness of the gestures selected with our method. We compare our method with two baseline selection methods. The first respects timing, the desired onset and duration of a gesture, but does not match gesture form in other ways. The second baseline additionally disregards the original gesture timing for selecting gestures. The gesture sequences from our method were rated as a significantly better match to the speech than gestures selected by either baseline method.

It's A Match! Gesture Generation Using Expressive Parameter Matching

Multi-Agent Path Finding (MAPF) is the challenging problem of computing collision-free paths for a cooperative team of moving agents. Algorithms for solving MAPF can be categorized on a spectrum. At one end are (bounded-sub)optimal algorithms that can find high-quality solutions for small problems. At the other end are unbounded-suboptimal algorithms (including prioritized and rule-based algorithms) that can solve very large practical problems but usually find low-quality solutions. In this paper, we consider a third approach that combines both advantages: anytime algorithms that quickly find an initial solution, including for large problems, and that subsequently improve the solution to near-optimal as time progresses. To improve the solution, we replan subsets of agents using Large Neighborhood Search, a popular meta-heuristic often applied in combinatorial optimization. Empirically, we compare our algorithm MAPF-LNS to the state-of-the-art anytime MAPF algorithm anytime BCBS and report significant gains in scalability, runtime to the first solution, and speed of improving solutions.

Anytime Multi-Agent Path Finding via Large Neighborhood Search

Relational agents who personalise their message to individual users need to learn about the user first. In the context of an agent who personalises the inclusion of 10 relational cues in its dialogue, we explore training the agent based on first time user’s responses to a single example of each cue prior to the session with the virtual advisor. We designed a between subjects study with three groups: Empathic (all relational cues); Neutral (no relational cues) and Adaptive (only helpful cues included). We found that in the Adaptive group, students received what they found helpful more often than in the other two groups. Analysis of the discrepancy between what user’s found helpful and what they received was least in the Adaptive group and greatest in the Neutral group.

Contrasting Centralized and Decentralized Critics in Multi-Agent Reinforcement Learning

Downloads

Next from AAMAS 2021

Structured Diversification Emergence via Reinforced Organization Control and Hierachical Consensus Learning

Similar lecture

Learning Node-Selection Strategies in Bounded-Suboptimal Conflict-Based Search for Multi-Agent Path Finding

Contrasting Centralized and Decentralized Critics in Multi-Agent Reinforcement Learning

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAMAS 2021

Structured Diversification Emergence via Reinforced Organization Control and Hierachical Consensus Learning

Similar lecture

Learning Node-Selection Strategies in Bounded-Suboptimal Conflict-Based Search for Multi-Agent Path Finding

Downloads