Takashi Kuremoto, Takaomi Hirata, Masanao Obayashi, Shingo Mabu and Kunikazu Kobayashi (April 3rd 2019). Training Deep Neural Networks with Reinforcement Learning for Time Series Forecasting, Time Series Analysis - Data, Methods, and Applications, Chun-Kit Ngan, IntechOpen, DOI: 10.5772/intechopen.85457. 

Masakazu Sato, Kaori Koga, Tomoyuki Fujii and Yutaka Osuga (June 27th 2018). Can Reinforcement Learning Be Applied to Surgery?, Artificial Intelligence - Emerging Trends and Applications, Marco Antonio Aceves-Fernandez, IntechOpen, DOI: 10.5772/intechopen.76146.

Mengchun Xie (March 18th 2019). Improvement of Cooperative Action for Multi-Agent System by Rewards Distribution, Assistive and Rehabilitation Engineering, Yves Rybarczyk, IntechOpen, DOI: 10.5772/intechopen.85109. 

New Zealand

This video presentation presents the main work we have done for investigating of Multi-agent Reinforcement Learning for Multi-traffic Light Control.

AAMAS 2020

Integrating independent and centralized multi-agent  reinforcement learning for traffic signal network optimization

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 

AAMAS 2021

technical paper

AAMAS is the leading scientific conference for research in autonomous agents and multi-agent systems. The AAMAS conference series was initiated in 2002 as the merging of three respected scientific meetings: the International Conference on Multi-Agent Systems (ICMAS), the International Workshop on Agent Theories, Architectures, and Languages (ATAL), and the International Conference on Autonomous Agents (AA). The aim of the joint conference is to provide a single, high-profile, internationally-respected archival forum for scientific research in the theory and practice of autonomous agents and multi-agent systems.

Browse keynotes, discussions, panels and over 300 presentations.


AAMAS is the leading scientific conference for research in autonomous agents and multi-agent systems. The AAMAS conference series was initiated in 2002 as the merging of three respected scientific meetings: the International Conference on Multi-Agent Systems (ICMAS), the International Workshop on Agent Theories, Architectures, and Languages (ATAL), and the International Conference on Autonomous Agents (AA). 

Agents that are tailored to appear and behave as members of a particular culture are more acceptable by and persuasive to members of that culture than agents that are not tailored. We report a study that systematically unpacks two tailoring components—appearance and argumentation—for virtual exercise coaches designed for the Indian and American cultures. Indian participants who interacted with an agent whose argumentation was tailored to their culture were significantly more satisfied with the agent irrespective of the agent’s appearance.

Argumentation is More Important than Appearance for Designing Culturally Tailored Virtual Agents

Restricted Domains of Dichotomous Preferences with Possibly Incomplete Information

BitcoinF: Achieving Fairness for Bitcoin in Transaction-Free-Only Model

Games are an increasingly useful tool for training and testing learning algorithms. Recent examples include GANs, AlphaZero and the AlphaStar league. However, multi-agent learning can be extremely difficult to predict and control. Learning dynamics can yield chaotic behavior even in simple games. In this paper, we present basic mechanism design tools for constructing games with predictable and controllable dynamics. We present a robust framework for dimensionality reduction arguments in large network games.

Robust Self-Organization in Games: Symmetries, Conservation Laws and Dimensionality Reduction

The open-source Unity platform, where agents can be trained using hierarchical or non-hierarchical reinforcement learning, supports the use of games and simulations as environments for multiple- agent interactions. In this demonstration, we present hierarchical and non-hierarchical multi-agent interactions based on Unity reinforcement learning, specifically, hierarchical reinforcement learning that sets different levels of agent’s observations to achieve the goal. We created four multi-agent scenarios in the Unity environment, namely, Crawler, Tennis, Banana Collector, and Soccer, to test the interaction performances of hierarchical and non- hierarchical reinforcement learning. The simulation-interaction performances show that hierarchical reinforcement learning can be applied to multi-agent environments and can compete with agents trained via non-hierarchical reinforcement learning. 


Peer Reviewing in Participatory Guarantee Systems

Axiomatic approaches are an appealing method for designing fair algorithms, as they provide formal structure for reasoning about and rationalizing individual decisions. However, to make these algorithms useful in practice, their axioms must appropriately capture social norms. We explore this tension between fairness axioms and socially acceptable decisions in the context of cooperative game theory. We use two crowdsourced experiments to study people's impartial reward divisions in cooperative games, focusing on games that systematically vary the values of the single-player coalitions. Our results show that people select rewards that are remarkably consistent, but place much more emphasis on the single-player coalitions than the Shapley value does. Further, their reward divisions violate both the null player and additivity axioms, but support weaker axioms. We argue for a more general methodology of testing axioms against experimental data, retaining some of the conceptual simplicity of the axiomatic approach while still using people’s opinions to drive the design of fair algorithms.

Testing Axioms Against Human Reward Divisions in Cooperative Games

Recent years have seen various designs of strategyproof mechanisms in the facility location game and the obnoxious facility game, by considering the facility as a point. In this paper, we extend that point to be an interval and study a novel activity scheduling game to schedule an activity in the time domain [0, 1] based on all agents’ time reports. The activity lasts for a time period of d with 0 <=d<=1, and each agent i wants his private time 𝑡𝑖 to be within the activity duration [y,y+d] or at least as close as possible. Thus his cost is the time difference between his time t_i and the activity duration [y, y+d]. The social cost is the summation of all agents’ costs. Our objective is to choose the activity starting time 𝑦 so that the mechanisms are strategyproof (truthful) and efficient. We design a mechanism outputting an optimal solution and prove it is group strategyproof. For minimizing the maximum cost, we also design a strategyproof mechanism with approximation ratio 2. In the obnoxious activity scheduling game, each agent prefers his conflict time t_i to be far away from the activity duration [y,y+d]. We respectively design deterministic and randomized group strategyproof mechanisms with provable approximation ratios and also show the lower bounds. Besides, for extension, we consider the cost/utility as the characteristic function and find group strategyproof mechanisms for minimizing the social cost and maximizing the social utility.

Strategyproof Mechanisms for Activity Scheduling

From skipped exercise classes to last-minute cancellation of dentist appointments, underutilization of reserved resources abounds. Likely reasons include uncertainty about the future, further exacerbated by present bias. In this paper, we unite resource allocation and commitment devices through the design of contingent payment mechanisms, and propose the two-bid penalty-bidding mechanism. This extends an earlier mechanism proposed by Ma et al. (2019), assigning the resources based on willingness to accept a no-show penalty, while also allowing each participant to increase her own penalty in order to counter present bias. We establish a simple dominant strategy equilibrium, regardless of an agent's level of present bias or degree of "sophistication". Via simulations, we show that the proposed mechanism substantially improves utilization and achieves higher welfare and better equity in comparison with mechanisms used in practice and mechanisms that optimize welfare in the absence of present bias.

Penalty Bidding Mechanisms for Allocating Resources and Overcoming Present Bias

Many websites that recommend various services use crowdsourcing to collect reviews and rankings. These rankings, usually concerning a subset of all the offered alternatives, are then aggregated. Motivated by such scenarios, we axiomatise a family of positional scoring rules for profiles of possibly incomplete individual preferences. Many opportunities arise for the agents to manipulate the outcome in this setting. They may lie in order to obtain a better result by: (i) switching the order of a ranked pair of alternatives, (ii) omitting some of their truthful preferences, or (iii) reporting more preferences than the ones they truthfully hold. After formalising these new concepts, we characterise all positional scoring rules that are immune to manipulation.

Strategic Manipulation with Incomplete Preferences: Possibilities and Impossibilities for Positional Scoring Rules

Vaccination is a standard public health intervention for controlling the spread of epidemics. However, the supply of vaccines is typically limited, and therefore, their deployment needs to be optimized. Further, vaccines are produced over time, so the strategies have to be temporal. We study the problem \prob{} of designing vaccination strategies, within available budget constraints, to minimize the spread of an outbreak. This is a challenging stochastic optimization problem. We design a bicriteria approximation algorithm, which combines a linear programming based rounding, along with the sample average approximation technique. Our approach also provides the empirical approximation factor for the problem instance, relative to the optimum. We find that the approximation factor is significantly better than the worst case bound, and, in practice, is a small constant factor. Further, our method shows significantly better performance than all prior heuristics for this problem. With additional pruning techniques, we are able to scale our algorithm to networks with millions of edges.

Integrating independent and centralized multi-agent reinforcement learning for traffic signal network optimization

Downloads

Next from AAMAS 2020

Argumentation is More Important than Appearance for Designing Culturally Tailored Virtual Agents

Similar lecture

Mitigating Negative Side Effects via Environment Shaping

Integrating independent and centralized multi-agent reinforcement learning for traffic signal network optimization

.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 2020

Argumentation is More Important than Appearance for Designing Culturally Tailored Virtual Agents

Similar lecture

Mitigating Negative Side Effects via Environment Shaping

Downloads