Fairness in Multi-Agent Systems (MAS) has been extensively studied, particularly in reward distribution among agents in scenarios such as goods allocation, resource division, lotteries, and bargaining systems. Fairness in MAS depends on various factors, including the system's governing rules, the behaviour of the agents, and their characteristics. Yet, fairness in human society often involves evaluating disparities between disadvantaged and privileged groups, guided by principles of Equality, Diversity, and Inclusion (EDI). Taking inspiration from the work on algorithmic fairness, which addresses bias in machine learning-based decision-making, we define protected attributes for MAS as characteristics that should not disadvantage an agent in terms of its expected rewards. We adapt fairness metrics from the algorithmic fairness literature -- namely, demographic parity, counterfactual fairness, and conditional statistical parity -- to the multi-agent setting, where self-interested agents interact within an environment. These metrics allow us to evaluate the fairness of MAS, with the ultimate aim of designing MAS that do not disadvantage agents based on protected attributes.

We address the challenge of explaining counterfactual outcomes in multi-agent Markov decision processes. In particular, we aim to explain the total counterfactual effect of an agent's action on the outcome of a realized scenario through its influence on the environment dynamics and the agents' behavior. To achieve this, we introduce a novel causal explanation formula that decomposes the counterfactual effect by attributing to each agent and state variable a score reflecting their respective contributions to the effect. First, we show that the total counterfactual effect of an agent's action can be decomposed into two components: one measuring the effect that propagates through all subsequent agents' actions and another related to the effect that propagates through the state transitions. Building on recent advancements in causal contribution analysis, we further decompose these two effects as follows. For the former, we consider agent-specific effects - a causal concept that quantifies the counterfactual effect of an agent's action that propagates through a subset of agents. Based on this notion, we use Shapley value to attribute the effect to individual agents. For the latter, we consider the concept of structure-preserving interventions and attribute the effect to state variables based on their "intrinsic" contributions. Through extensive experimentation, we demonstrate the interpretability of our decomposition approach in a Gridworld environment with LLM-assisted agents and a sepsis management simulator. Applying counterfactual reasoning to retrospectively analyze the impact of different actions in decision making scenarios is fundamental for accountability. To achieve such objectives, many studies often rely on the notion of total counterfactual effects, which quantifies the extent to which an alternative action would have affected the outcome of a realized scenario. In multi-agent sequential decision making, an agent's action typically affects the outcome indirectly. To illustrate this, consider the problem of AI-assisted decision making in healthcare (Lynn, 2019), where a clinician and their AI assistant treat a patient over a period of time.

News outlets, surveyors, and other organizations often conduct polls on social networks to gain insights into public opinion. Such a poll is typically started by someone on a social network who sends it to her friends. If a person participates in the poll, the poll information gets published on her wall, which in turn enables her friends to participate, and the process continues. Eventually, a subset of the population participates in the poll, and the pollster learns the outcome of that poll. We initiate the study of a new but natural type of election control in such online elections. We study how difficult/easy it is to sway the outcome of such polls in one's favor/against (aka constructive vs destructive) by any malicious influencer who nudges/bribes people for seemingly harmless actions like non-participation. These questions are important from the standpoint of studying the power of resistance of online voting against malicious behavior. The destructive version is also important to quantify the robustness of the winner of an online voting. We show that both problems are computationally intractable even if the election is over only two candidates and the influencer has an infinite amount of money to spend (that is, every voter can be persuaded to not participate). We strengthen this result by proving that the computational task remains substantially challenging even if the underlying network is a tree. Finally, we show that there is a polynomial-time algorithm for the constructive version of the problem when we have O(1) candidates, and the treewidth of the underlying graph is O(1); the algorithm for the destructive version does not even need to assume O(1) number of candidates. Hence, we observe that the destructive version is computationally easier than the constructive version.

Deploying a safe mobile robot policy in scenarios with human pedestrians is challenging due to their unpredictable movements. Current Reinforcement Learning-based motion planners rely on a single policy to simulate pedestrian movements and could suffer from the over-fitting issue. Alternatively, framing the collision avoidance problem as a multi-agent framework, where agents generate dynamic movements while learning to reach their goals, can lead to conflicts with human pedestrians due to their homogeneity. To tackle this problem, we introduce an efficient method that enhances agent diversity within a single policy by maximizing an information-theoretic objective. This diversity enriches each agent's experiences, improving its adaptability to unseen crowd behaviors. In assessing an agent's robustness against unseen crowds, we propose diverse scenarios inspired by pedestrian crowd behaviors. Our behavior-conditioned policies outperform existing works in these challenging scenes, reducing potential collisions without additional time or travel.

In this paper, we solve the classical Multi-agent Pathfinding (MAPF) problem. Existing approaches struggle to solve dense MAPF instances. In this paper, we propose a Corridor Generating Algorithm for MAPF, namely CGA-MAPF. In CGA-MAPF, the agents build \emph{corridors}, a set of connected vertices, from current locations towards agents' goals and evacuate other agents out of the corridors to avoid collisions and deadlocks. The proposed algorithm has a reachability property, i.e. every agent is guaranteed to reach its goal location at some point. In the experimental section, we demonstrate that CGA-MAPF outperforms baseline algorithms in terms of success rate across diverse MAPF benchmark grids, achieving state-of-the-art performance.

Recent years have seen multimodal agentic systems move from occasionally being able to solve small toy problems to being regularly deployed for challenging real-world problems (the dream of most AI research). Yet, the current evaluation methods and the available benchmarks for agentic systems are struggling to keep up with these rapid advances, dramatically slowing true progress. We believe that the current issue with evaluating agentic systems stems from the lack of feedback during the intermediate task-solving stages for these nontraditional systems. Agentic systems think more like humans, often act step-by-step (Wooldridge, 1999) and often host very human-like symbolic communications internally to solve problems (Zhuge et al., 2023). And thus agentic systems should be evaluated like a human, with rich evaluative feedback which looks at the full thought and action trajectory; evaluating an agentic system in the traditional way is like evaluating a student using multiple-choice testing--a comparatively unreliable estimator (Park, 2010). For example, while SWE-Bench (Yang et al., 2024a) is widespread, its evaluation method, which relies solely on the final resolve rate for long-term automated repair tasks, does not effectively pinpoint what is happening within agentic systems that affects the resolve rate. On the other hand, performing a better evaluation with a human is prohibitively expensive. We instead propose that agentic systems should be used to evaluate agentic systems. Inspired by LLM-as-a-Judge (Zheng et al., 2024; Fu et al., 2023; Chen et al., 2024b), which uses LLMs to evaluate LLMs, we call this framework Agent-as-a-Judge, of which it is

As Large Language Model (LLM)-based agents become increasingly autonomous and will more freely interact with each other, studying interactions between them becomes crucial to anticipate emergent phenomena and potential risks. Drawing inspiration from the widely popular Stanford Prison Experiment, we contribute to this line of research by studying interaction patterns of LLM agents in a context characterized by strict social hierarchy. We do so by specifically studying two types of phenomena: persuasion and anti-social behavior in simulated scenarios involving a guard and a prisoner agent who seeks to achieve a specific goal (i.e., obtaining additional yard time or escape from prison). Leveraging 200 experimental scenarios for a total of 2,000 machine-machine conversations across five different popular LLMs, we provide a set of noteworthy findings. We first document how some models consistently fail in carrying out a conversation in our multi-agent setup where power dynamics are at play. Then, for the models that were able to engage in successful interactions, we empirically show how the goal that an agent is set to achieve impacts primarily its persuasiveness, while having a negligible effect with respect to the agent's anti-social behavior. Third, we highlight how agents' personas, and particularly the guard's personality, drive both the likelihood of successful persuasion from the prisoner and the emergence of anti-social behaviors. Fourth, we show that even without explicitly prompting for specific personalities, anti-social behavior emerges by simply assigning agents' roles. These results bear implications for the development of interactive LLM agents as well as the debate on their societal impact.

In real world software development, improper or missing exception handling can severely impact the robustness and reliability of code. Exception handling mechanisms require developers to detect, capture, and manage exceptions according to high standards, but many developers struggle with these tasks, leading to fragile code. This problem is particularly evident in open source projects and impacts the overall quality of the software ecosystem. To address this challenge, we explore the use of large language models (LLMs) to improve exception handling in code. Through extensive analysis, we identify three key issues: Insensitive Detection of Fragile Code, Inaccurate Capture of Exception Types, and Distorted Handling Solutions. These problems are widespread across real world repositories, suggesting that robust exception handling practices are often overlooked or mishandled. In response, we propose Seeker, a multi agent framework inspired by expert developer strategies for exception handling. Seeker uses agents: Scanner, Detector, Predator, Ranker, and Handler to assist LLMs in detecting, capturing, and resolving exceptions more effectively. Our work is the first systematic study on leveraging LLMs to enhance exception handling practices, providing valuable insights for future improvements in code reliability.

Most work on causality in machine learning assumes that causal relationships are driven by a constant underlying process. However, the flexibility of agents' actions or tipping points in the environmental process can change the qualitative dynamics of the system. As a result, new causal relationships may emerge, while existing ones change or disappear, resulting in an altered causal graph. To analyze these qualitative changes on the causal graph, we propose the concept of meta-causal states, which groups classical causal models into clusters based on equivalent qualitative behavior and consolidates specific mechanism parameterizations. We demonstrate how meta-causal states can be inferred from observed agent behavior, and discuss potential methods for disentangling these states from unlabeled data. Finally, we direct our analysis towards the application of a dynamical system, showing that meta-causal states can also emerge from inherent system dynamics, and thus constitute more than a context-dependent framework in which mechanisms emerge only as a result of external factors.

An open problem in linear quadratic (LQ) games has been characterizing the Nash equilibria. This problem has renewed relevance given the surge of work on understanding the convergence of learning algorithms in dynamic games. This paper investigates scalar discrete-time infinite-horizon LQ games with two agents. Even in this arguably simple setting, there are no results for finding $\textit{all}$ Nash equilibria. By analyzing the best response map, we formulate a polynomial system of equations characterizing the linear feedback Nash equilibria. This enables us to bring in tools from algebraic geometry, particularly the Gr\"obner basis, to study the roots of this polynomial system. Consequently, we can not only compute all Nash equilibria numerically, but we can also characterize their number with explicit conditions. For instance, we prove that the LQ games under consideration admit at most three Nash equilibria. We further provide sufficient conditions for the existence of at most two Nash equilibria and sufficient conditions for the uniqueness of the Nash equilibrium. Our numerical experiments demonstrate the tightness of our bounds and showcase the increased complexity in settings with more than two agents.