United Kingdom

Equilibrium computation in markets usually considers settings where player valuation functions are known. We consider the setting where player valuations are unknown; using a PAC learning-theoretic framework, we analyze some classes of common valuation functions, and provide algorithms which output direct PAC equilibrium allocations, not estimates based on attempting to learn valuation functions. Since there exist trivial PAC market outcomes with an unbounded worst-case efficiency loss, we lower-bound the efficiency of our algorithms. While the efficiency loss under general distributions is rather high, we show that in some cases (e.g., unit-demand valuations), it is possible to find a PAC market equilibrium with significantly better utility.

AAMAS 2021

The Price is (Probably) Right: Learning Market Equilibria from Samples

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

Expressing attack-defence trees (ADTrees) in a multi-agent setting allows for studying a new aspect of security scenarios, namely how the number of agents and their task assignment impact the performance of attacking and defending strategies executed by agent coalitions. Our tool ADT2AMAS allows for transforming ADTrees into extended asynchronous multi-agent systems and computing an optimal schedule with the minimal number of agents. ADT2AMAS is integrated within the graphical verification platform CosyVerif, but can also be run standalone.

ADT2AMAS: Managing Agents in Attack-Defence Scenarios

In cooperative human decision-making, agreements are often not total; a partial degree of agreement is sufficient to commit to a decision and move on, as long as one is somewhat confident that the involved parties are likely to stand by their commitment in the future, given no drastic unexpected changes. In this work, we introduce models that allow autonomous agents to reach such agreements, using abstract argumentation as the underlying model.

Toward Consistent Agreement Approximation in Abstract Argumentation and Beyond

In this paper, centralized and decentralized laws for the coverage problem under connectivity constraints are proposed. We formulated the problem as a continuous optimization task with the inequality constraint that algebraic connectivity must not be less than a small positive number, and solved the problem based on the active set method. This formulation can cause agents to be trapped in undesired local minima; to address this, we added the squared distance between the centroid of the coverage area and that of the agent system to the objective function of the coverage control. To derive the decentralized law, we employed an average consensus estimator that determines the algebraic connectivity and system centroid. When the algebraic connectivity is greater than the threshold, the proposed control laws allow agents to advance along the direction of steepest descent of the objective function. Once the connectivity is equal to the threshold, agents maintain the connectivity while decreasing the objective function by moving along the projection vector of the steepest descent in the direction in which the connectivity is constant. Simulation results confirmed the effectiveness of our proposed laws.

Coverage Control under Connectivity Constraints

Designing autonomous agents that act collaboratively to achieve a common set of goals has been a major goal for Artificial Intelligence research for many years. Collaborative Privacy-Preserving Planning (cppp) is a multi-agent planning task in which agents need to achieve a common set of goals without revealing certain private information [2]. In particular, in cppp an individual agent may have a set of private facts and actions that it does not share with the other agents. cppp has important motivating examples, such as planning for organizations that outsource some of their tasks, and has recently received considerable attention from the academic community.

Partial Disclosure of Private Dependencies in Privacy Preserving Planning

Business models rooted in shared economy, electrification, and automation are transforming urban mobility. Accounting for how these transformations interact is crucial if synergies are to be exploited. In this paper, we focus on how a cost-effective charging infrastructure for e-mobility can support the emergence of shared, autonomous mobility. This study addresses the problem of siting and sizing of charging stations for a fleet of shared autonomous electric vehicles (SAEVs). We develop a hybrid simulation-optimization model to find locations and numbers of chargers needed to serve charging demands. Our agent-based model provides an enhanced representation of SAEV operations allowing for smart charging and vehicle cruising when parking/charging is not available. Also, we model charging station placement as full covering optimization and solve the location-allocation problem simultaneously. Finally, we employ real-world trip data from ShareNow in Berlin to evaluate our approach for realistic demand patterns under different charging strategies and fleet sizes. The results show that charging station locations depend mostly on the spatial distribution of installation costs and charging demands. Moreover, charging strategies and fleet size affect the charging patterns and the required number of chargers as well as fleet performance.

Siting and Sizing of Charging Infrastructure for Shared Autonomous Electric Fleets

Non-stationary environments are challenging for reinforcement learning algorithms. If the state transition and/or reward functions change based on latent factors, the agent is effectively tasked with optimizing a behavior that maximizes performance over a possibly infinite random sequence of Markov Decision Processes (MDPs), each of which drawn from some unknown distribution. We call each such MDP a context. Most related works make strong assumptions such as knowledge about the distribution over contexts, the existence of pre-training phases, or a priori  knowledge about the number, sequence, or boundaries between contexts. We introduce an algorithm that efficiently learns policies in non-stationary environments. It analyzes a possibly infinite stream of data and computes, in real-time, high-confidence change-point detection statistics that reflect whether novel, specialized policies need to be created and deployed to tackle novel contexts, or whether previously-optimized ones might be reused. We show that (i)  this algorithm minimizes the delay until unforeseen changes to a context are detected, thereby allowing for rapid responses; and (ii) it bounds the rate of false alarm, which is important in order to minimize regret. Our method constructs a mixture model composed of a (possibly infinite) ensemble of probabilistic dynamics predictors that model the different modes of the distribution over underlying latent MDPs. We evaluate our algorithm on high-dimensional continuous reinforcement learning problems and show that it outperforms state-of-the-art (model-free and model-based) RL algorithms, as well as state-of-the-art meta-learning methods specially designed to deal with non-stationarity.

Minimum-Delay Adaptation in Non-Stationary Reinforcement Learning via Online High-Confidence Change-Point Detection

We introduce a novel method to aggregate Bipolar Argumentation (BA) Frameworks expressing opinions by different parties in debates. We use Bipolar Assumption-based Argumentation (ABA) as an all-encompassing formalism for BA under different semantics. By leveraging on recent results on judgement aggregation in Social Choice Theory, we prove several preservation results, both positive and negative, for relevant properties of Bipolar ABA.

Aggregating Bipolar Opinions

We introduce a new numerical framework to learn optimal bidding strategies in repeated auctions when the seller uses past bids to optimize her mechanism. Crucially, we do not assume that the bidders know which optimization mechanism is used by the seller. We recover essentially all state-of-the-art analytical results for the single-item framework derived previously in the setup where the bidder knows the optimization mechanism used by the seller and extend our approach to multi-item settings, in which no optimal shading strategies were previously known. Our approach yields substantial increases in bidder utility in all settings and has a strong potential for practical usage since it provides a simple way to optimize bidding strategies on modern marketplaces where buyers face unknown data-driven mechanisms.

Adversarial Learning in Revenue-Maximizing Auctions

We study mechanism design for nonexcludable and excludable binary public project problems. We aim to maximize the expected number of consumers and the expected agents' welfare. For the nonexcludable public project model, we identify a sufficient condition on the prior distribution for the conservative equal costs mechanism to be the optimal strategy-proof and individually rational mechanism. For general distributions, we propose a dynamic program that solves for the optimal mechanism. For the excludable public project model, we identify a similar sufficient condition for the serial cost sharing mechanism to be optimal for 2 and 3 agents. We derive a numerical upper bound. Experiments show that for several common distributions, the serial cost sharing mechanism is close to optimality.

Mechanism Design for Public Projects via Neural Networks

In AI, Multi-Agent Systems are able to model many kind of collective behavior and have therefore a wide range of application. In this paper, we propose a logical framework (Logic of "Inferable") which enable reasoning about whether a group of agents can perform an action, highlighting the concepts of cost of actions and of budget that agents have available to perform actions. The focus is on modeling the group dynamics of cooperative agents.

The Price is (Probably) Right: Learning Market Equilibria from Samples

Downloads

Next from AAMAS 2021

ADT2AMAS: Managing Agents in Attack-Defence Scenarios

Similar lecture

Committee Selection using Attribute Approvals

The Price is (Probably) Right: Learning Market Equilibria from Samples

.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

ADT2AMAS: Managing Agents in Attack-Defence Scenarios

Similar lecture

Committee Selection using Attribute Approvals

Downloads