United Kingdom

Off-policy evaluation (OPE) is the problem of evaluating new policies using historical data obtained from a different policy. In the recent OPE context, most studies have focused on single-player cases, and not on multi-player cases. In this study, we propose OPE estimators constructed by the doubly robust and double reinforcement learning estimators in two-player zero-sum Markov games. The proposed estimators project exploitability that is often used as a metric for determining how close a policy profile (i.e., a tuple of policies) is to a Nash equilibrium in two-player zero-sum games. We prove the exploitability estimation error bounds for the proposed estimators. We then propose the methods to find the best candidate policy profile by selecting the policy profile that minimizes the estimated exploitability from a given policy profile class. We prove the regret bounds of the policy profiles selected by our methods. Finally, we demonstrate the effectiveness and performance of the proposed estimators through experiments.

AAMAS 2021

Off-Policy Exploitability-Evaluation in Two-Player Zero-Sum Markov Games

2021 Annual Conference of the North American Chapter of the Association for Computational Linguistics.

**Whova App** 
Stay in touch with your fellow conference attendees via the [Whova App](https://whova.com/portal/webapp/nacon_202106/)

**Conference Structure**
https://2021.naacl.org/blog/conference-structure/

**Walkthrough video of how to NAACL 2021** 

Please take a moment to view this video explaining how to navigate the platform, attend sessions network with other attendees. 


<figure class="video_container">
  <iframe src="https://screencast-o-matic.com/watch/crhwbGVh3vx?v=6&ff=1&title=0&controls=1" width=640  height=350  frameborder="0" allowfullscreen="true"> </iframe>
</figure>

<br>
<br>

NAACL 2021

technical paper

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 

Mechanism design has traditionally assumed that the set of participants are fixed and known to the mechanism (the market owner) in advance. However, in practice, the market owner can only directly reach a small number of participants (her neighbours). Hence the owner often needs costly promotions to recruit more participants in order to get desirable outcomes such as social welfare or revenue maximization. In this paper, we propose to incentivize existing participants to invite their neighbours to attract more participants. However, they would not invite each other if they are competitors. We discuss how to utilize the conflict of interest between the participants to incentivize them to invite each other to form larger markets. We will highlight the early solutions and open the floor for discussing the fundamental open questions in the settings of auctions, coalitional games, matching and voting.

Mechanism Design Powered by Social Interactions (Blue Sky Ideas Track)

When trains are finished with their transportation tasks during the day, they are moved to a shunting yard where they are routed, parked, cleaned, subject to regular maintenance checks and repaired during the night. The resulting Train Unit Shunting and Servicing problem motivates advanced research in planning and scheduling in general since it integrates several known individually hard problems while incorporating many real-life details. We developed an event-based simulator called TORS (Dutch acronym for Train Shunting and Servicing Simulator), that provides the user with a state and all feasible actions. After an action is picked, TORS calculates the result and the process repeats.   This simulator facilitates research into a realistic application of multi-agent path finding.

TORS: A Train Unit Shunting and Servicing Simulator

A self-governed society must have decision rules by which group decisions are made, and these rules are often codified in a written constitution. One of the defining features of a constitution is its degree of entrenchment, or how hard it is to change by amendment. If it is too easy to make amendments, then the constitution can change too frequently, leading to chaos. On the other hand, if it is too hard to make amendments, then this can also be destabilizing, as voters may begin to see the rules as less legitimate, or even seek to overturn the status quo in a revolt. As norms, priorities, and circumstances change over time and over generations a constitution must be able to adapt. Our work considers a stylized model of constitutions that use reality-aware supermajority rules to make decisions. We propose principles for designing amendment procedures for changing decision rules in these constitutions and propose a novel procedure based on these principles.

How to Amend a Constitution? Model, Axioms, and Supermajority Rules

Evolutionary Game Theory is an application of game theory to evolving populations of organisms. Of recent interest are EGT models situated on structured populations or spatial evolutionary games. Due to the complexity added by introducing a population structure, model analysis is usually performed through agent-based Monte-Carlo simulations. However, it can be difficult to obtain desired quantities of interest from these simulations due to stochastic effects. We define a framework for modeling spatial evolutionary games using Dynamic Bayesian Networks that capture the underlying stochastic process. The resulting Dynamic Bayesian Networks can be queried for quantities of interest by performing exact inference on the network. We then propose a method for producing approximations of the spatial evolutionary game through the truncation of the corresponding DBN, taking advantage of the high symmetry of the model. This method generalizes mean-field and pair approximations in the literature for spatial evolutionary games. Furthermore, we show empirical results demonstrating the capability of the method to obtain much better accuracy than pair approximation with respect to stochastic simulations.

Approximating Spatial Evolutionary Games using Bayesian Networks

In this work we present a novel succinct representation for large partially observed cooperative games. The proposed representation exploits estimates over marginal contributions to form compact rules representing collaboration patterns with uncertain value. Specifically, given an initial set of MC-nets rules that use prior beliefs over values instead of the actual ones, we propose two types of merging that lead to a new set of even more compact rules.

A Succinct Representation Scheme for Cooperative Games under Uncertainty

This paper studies cohorting in public transit systems and its usefulness in mitigating disease transmission. The Mumbai suburban railway system is used as a case study.

Cohorting to Isolate Asymptomatic Spreaders: An Agent-Based Simulation Study on the Mumbai Suburban Railway

Multi-agent reinforcement learning (MARL) has been increasingly used in a wide range of safety-critical applications, which require guaranteed safety (e.g., no unsafe states are ever visited) during the learning process.Unfortunately, current MARL methods do not have safety guarantees. Therefore, we present two shielding approaches for safe MARL. In centralized shielding, we synthesize a single shield to monitor all agents' joint actions and correct any unsafe action if necessary.In factored shielding, we synthesize multiple shields based on a factorization of the joint state space observed by all agents; the set of shields monitors agents concurrently and each shield is only responsible for a subset of agents at each step.Experimental results show that both approaches can guarantee the safety of agents during learning without compromising the quality of learned policies; moreover, factored shielding is more scalable in the number of agents than centralized shielding.

Safe Multi-Agent Reinforcement Learning via Shielding

Goal recognition is the problem of determining an agent's intent by observing its actions. In the context of AI research, the problem is tackled for two quite different purposes: to determine an agent's most probable goal or, for human-aware planning including planned  or strategic  deception, to determine an observer's most likely belief about that goal. Making no distinction, contemporary models tend to assume an infallible observer, deceived only while it has limited access to information or if the environment itself is only partially observable. Focusing on the second purpose, we propose an extended framework that incorporates formal definitions of confirmation bias, selective attention and memory decay. In contrast to pre-existing models, our approach combines explicit consideration of prior probabilities with a principled representation of observer confidence and distinguishes between potential observations  i.e., every observable event within the observer's frame of reference  and recalled observations which we model as a function of attention and memory. We show that when these factors are taken into consideration, false beliefs may arise and can be made to persist, even in a fully observable environment  thus providing a perceptual model readily incorporated into the "thinking" of an adversarial agent for the purpose of strategic deception.

Extended Goal Recognition: A Planning-Based Model for Strategic Deception

Recent work on promoting cooperation in multi-agent learning has resulted in many methods which successfully promote cooperation at the cost of becoming more vulnerable to exploitation by malicious actors. We show that this is an unavoidable trade-off and propose an objective which balances these concerns, promoting both safety and long-term cooperation. Moreover, the trade-off between safety and cooperation is not severe, and you can receive exponentially large returns through cooperation from a small amount of risk. We study both an exact solution method and propose a method for training policies that targets this objective, Accumulating Risk Capital Through Investing in Cooperation (ARCTIC), and evaluate them in iterated Prisoner's Dilemma and Stag Hunt.

Accumulating Risk Capital Through Investing in Cooperation

Prediction markets are powerful tools to elicit and aggregate beliefs from strategic agents. However, in current prediction markets, agents may exhaust the social welfare by competing to be the first to update the market. We initiate the study of the trade-off between how quickly information is aggregated by the market, and how much this information costs. We design markets to aggregate timely information from strategic agents to maximize social welfare. To this end, the market must incentivize agents to invest the correct amount of effort to acquire information: quickly enough to be useful, but not faster (and more expensively) than necessary. The market also must ensure that agents report their information truthfully and on time.  We consider two settings: in the first, information is only valuable before a deadline; in the second, the value of information decreases as time passes.  We use both theorems and simulations to demonstrate the mechanisms.

Off-Policy Exploitability-Evaluation in Two-Player Zero-Sum Markov Games

Downloads

Next from AAMAS 2021

Mechanism Design Powered by Social Interactions (Blue Sky Ideas Track)

Similar lecture

Modeling Content and Context with Deep Relational Learning

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Off-Policy Exploitability-Evaluation in Two-Player Zero-Sum Markov Games

.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

Mechanism Design Powered by Social Interactions (Blue Sky Ideas Track)

Similar lecture

Modeling Content and Context with Deep Relational Learning

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads