CyberSec Research

We introduce the General Incentives-based Framework for Fairness (GIFF), a novel approach for fair multi-agent resource allocation that infers fair decision-making from standard value functions. In resource-constrained settings, agents optimizing for efficiency often create inequitable outcomes. Our approach leverages the action-value (Q-)function to balance efficiency and fairness without requiring additional training. Specifically, our method computes a local fairness gain for each action and introduces a counterfactual advantage correction term to discourage over-allocation to already well-off agents. This approach is formalized within a centralized control setting, where an arbitrator uses the GIFF-modified Q-values to solve an allocation problem. Empirical evaluations across diverse domains, including dynamic ridesharing, homelessness prevention, and a complex job allocation task-demonstrate that our framework consistently outperforms strong baselines and can discover far-sighted, equitable policies. The framework's effectiveness is supported by a theoretical foundation; we prove its fairness surrogate is a principled lower bound on the true fairness improvement and that its trade-off parameter offers monotonic tuning. Our findings establish GIFF as a robust and principled framework for leveraging standard reinforcement learning components to achieve more equitable outcomes in complex multi-agent systems.

The electricity sector transition requires substantial increases in residential demand response capacity, yet Home Energy Management Systems (HEMS) adoption remains limited by user interaction barriers requiring translation of everyday preferences into technical parameters. While large language models have been applied to energy systems as code generators and parameter extractors, no existing implementation deploys LLMs as autonomous coordinators managing the complete workflow from natural language input to multi-appliance scheduling. This paper presents an agentic AI HEMS where LLMs autonomously coordinate multi-appliance scheduling from natural language requests to device control, achieving optimal scheduling without example demonstrations. A hierarchical architecture combining one orchestrator with three specialist agents uses the ReAct pattern for iterative reasoning, enabling dynamic coordination without hardcoded workflows while integrating Google Calendar for context-aware deadline extraction. Evaluation across three open-source models using real Austrian day-ahead electricity prices reveals substantial capability differences. Llama-3.3-70B successfully coordinates all appliances across all scenarios to match cost-optimal benchmarks computed via mixed-integer linear programming, while other models achieve perfect single-appliance performance but struggle to coordinate all appliances simultaneously. Progressive prompt engineering experiments demonstrate that analytical query handling without explicit guidance remains unreliable despite models' general reasoning capabilities. We open-source the complete system including orchestration logic, agent prompts, tools, and web interfaces to enable reproducibility, extension, and future research.

While Multi-Agent Systems (MAS) excel at complex tasks, their growing autonomy with operational complexity often leads to critical inefficiencies, such as excessive token consumption and failures arising from misinformation. Existing methods primarily focus on post-hoc failure attribution, lacking proactive, real-time interventions to enhance robustness and efficiency. To this end, we introduce SupervisorAgent, a lightweight and modular framework for runtime, adaptive supervision that operates without altering the base agent's architecture. Triggered by an LLM-free adaptive filter, SupervisorAgent intervenes at critical junctures to proactively correct errors, guide inefficient behaviors, and purify observations. On the challenging GAIA benchmark, SupervisorAgent reduces the token consumption of the Smolagent framework by an average of 29.45% without compromising its success rate. Extensive experiments across five additional benchmarks (math reasoning, code generation, and question answering) and various SoTA foundation models validate the broad applicability and robustness of our approach. The code is available at https://github.com/LINs-lab/SupervisorAgent.

People tend to walk in groups, and interactions with those groups have a significant impact on crowd behavior and pedestrian traffic dynamics. Social norms can be seen as unwritten rules regulating people interactions in social settings. This article studies people interactions with groups and the emergence of group proxemics. Group zones, zone occupancy counts and people clearance from the group are studied using naturalistic data. Analysis indicate potential presence of three different zones in addition to the public zone. People tend to remain in the public zone and only progressively get closer to groups, and those closer approaches happen in a low frequency and for brief periods of time.

This article investigates the pedestrian group as an emergent agent. The article explores empirical data to derive emergent agency and formation state spaces and outline recurring patterns of walking behavior. In this analysis, pedestrian trajectories extracted from surveillance videos are used along with manually annotated pedestrian group memberships. We conducted manual expert evaluation of observed groups, produced new manual annotations for relevant events pertaining to group behavior and extracted metrics relevant group formation. This information along with quantitative analysis was used to model the life-cycle and formation of the group agent. Those models give structure to expectations around walking behavior of groups; from pedestrian walking independently to the emergence of a collective intention where group members tended to maintain bounded distance between each other. Disturbances to this bounded distance often happened in association with changes in either their agency or their formation states. We summarized the patterns of behavior along with the sequences of state transitions into abstract patterns, which can aid in the development of more detailed group agents in simulation and in the design of engineering systems to interact with such groups.

Recently, deep multi-agent reinforcement learning (MARL) has demonstrated promising performance for solving challenging tasks, such as long-term dependencies and non-Markovian environments. Its success is partly attributed to conditioning policies on large fixed context length. However, such large fixed context lengths may lead to limited exploration efficiency and redundant information. In this paper, we propose a novel MARL framework to obtain adaptive and effective contextual information. Specifically, we design a central agent that dynamically optimizes context length via temporal gradient analysis, enhancing exploration to facilitate convergence to global optima in MARL. Furthermore, to enhance the adaptive optimization capability of the context length, we present an efficient input representation for the central agent, which effectively filters redundant information. By leveraging a Fourier-based low-frequency truncation method, we extract global temporal trends across decentralized agents, providing an effective and efficient representation of the MARL environment. Extensive experiments demonstrate that the proposed method achieves state-of-the-art (SOTA) performance on long-term dependency tasks, including PettingZoo, MiniGrid, Google Research Football (GRF), and StarCraft Multi-Agent Challenge v2 (SMACv2).

While a multi-agent approach based on large language models (LLMs) represents a promising strategy to surpass the capabilities of single models, its success is critically dependent on synergistic team composition. However, forming optimal teams is a significant challenge, as the inherent opacity of most models obscures the internal characteristics necessary for effective collaboration. In this paper, we propose an interaction-centric framework for automatic team composition that does not require any prior knowledge including their internal architectures, training data, or task performances. Our method constructs a "language model graph" that maps relationships between models from the semantic coherence of pairwise conversations, and then applies community detection to identify synergistic model clusters. Our experiments with diverse LLMs demonstrate that the proposed method discovers functionally coherent groups that reflect their latent specializations. Priming conversations with specific topics identified synergistic teams which outperform random baselines on downstream benchmarks and achieve comparable accuracy to that of manually-curated teams based on known model specializations. Our findings provide a new basis for the automated design of collaborative multi-agent LLM teams.

Generative AI (GenAI) is rapidly transforming software engineering (SE) practices, influencing how SE processes are executed, as well as how software systems are developed, operated, and evolved. This paper applies design science research to build a roadmap for GenAI-augmented SE. The process consists of three cycles that incrementally integrate multiple sources of evidence, including collaborative discussions from the FSE 2025 "Software Engineering 2030" workshop, rapid literature reviews, and external feedback sessions involving peers. McLuhan's tetrads were used as a conceptual instrument to systematically capture the transforming effects of GenAI on SE processes and software products.The resulting roadmap identifies four fundamental forms of GenAI augmentation in SE and systematically characterizes their related research challenges and opportunities. These insights are then consolidated into a set of future research directions. By grounding the roadmap in a rigorous multi-cycle process and cross-validating it among independent author teams and peers, the study provides a transparent and reproducible foundation for analyzing how GenAI affects SE processes, methods and tools, and for framing future research within this rapidly evolving area. Based on these findings, the article finally makes ten predictions for SE in the year 2030.

Traffic congestion in urban road networks leads to longer trip times and higher emissions, especially during peak periods. While the Shortest Path First (SPF) algorithm is optimal for a single vehicle in a static network, it performs poorly in dynamic, multi-vehicle settings, often worsening congestion by routing all vehicles along identical paths. We address dynamic vehicle routing through a multi-agent reinforcement learning (MARL) framework for coordinated, network-aware fleet navigation. We first propose Adaptive Navigation (AN), a decentralized MARL model where each intersection agent provides routing guidance based on (i) local traffic and (ii) neighborhood state modeled using Graph Attention Networks (GAT). To improve scalability in large networks, we further propose Hierarchical Hub-based Adaptive Navigation (HHAN), an extension of AN that assigns agents only to key intersections (hubs). Vehicles are routed hub-to-hub under agent control, while SPF handles micro-routing within each hub region. For hub coordination, HHAN adopts centralized training with decentralized execution (CTDE) under the Attentive Q-Mixing (A-QMIX) framework, which aggregates asynchronous vehicle decisions via attention. Hub agents use flow-aware state features that combine local congestion and predictive dynamics for proactive routing. Experiments on synthetic grids and real urban maps (Toronto, Manhattan) show that AN reduces average travel time versus SPF and learning baselines, maintaining 100% routing success. HHAN scales to networks with hundreds of intersections, achieving up to 15.9% improvement under heavy traffic. These findings highlight the potential of network-constrained MARL for scalable, coordinated, and congestion-aware routing in intelligent transportation systems.

Algorithmic collusion has emerged as a central question in AI: Will the interaction between different AI agents deployed in markets lead to collusion? More generally, understanding how emergent behavior, be it a cartel or market dominance from more advanced bots, affects the market overall is an important research question. We propose a hierarchical multi-agent reinforcement learning framework to study algorithmic collusion in market making. The framework includes a self-interested market maker (Agent~A), which is trained in an uncertain environment shaped by an adversary, and three bottom-layer competitors: the self-interested Agent~B1 (whose objective is to maximize its own PnL), the competitive Agent~B2 (whose objective is to minimize the PnL of its opponent), and the hybrid Agent~B$^\star$, which can modulate between the behavior of the other two. To analyze how these agents shape the behavior of each other and affect market outcomes, we propose interaction-level metrics that quantify behavioral asymmetry and system-level dynamics, while providing signals potentially indicative of emergent interaction patterns. Experimental results show that Agent~B2 secures dominant performance in a zero-sum setting against B1, aggressively capturing order flow while tightening average spreads, thus improving market execution efficiency. In contrast, Agent~B$^\star$ exhibits a self-interested inclination when co-existing with other profit-seeking agents, securing dominant market share through adaptive quoting, yet exerting a milder adverse impact on the rewards of Agents~A and B1 compared to B2. These findings suggest that adaptive incentive control supports more sustainable strategic co-existence in heterogeneous agent environments and offers a structured lens for evaluating behavioral design in algorithmic trading systems.

Automating the co-design of a robot's morphology and control is a long-standing challenge due to the vast design space and the tight coupling between body and behavior. We introduce Debate2Create (D2C), a framework in which large language model (LLM) agents engage in a structured dialectical debate to jointly optimize a robot's design and its reward function. In each round, a design agent proposes targeted morphological modifications, and a control agent devises a reward function tailored to exploit the new design. A panel of pluralistic judges then evaluates the design-control pair in simulation and provides feedback that guides the next round of debate. Through iterative debates, the agents progressively refine their proposals, producing increasingly effective robot designs. Notably, D2C yields diverse and specialized morphologies despite no explicit diversity objective. On a quadruped locomotion benchmark, D2C discovers designs that travel 73% farther than the default, demonstrating that structured LLM-based debate can serve as a powerful mechanism for emergent robot co-design. Our results suggest that multi-agent debate, when coupled with physics-grounded feedback, is a promising new paradigm for automated robot design.

An Agentic AI Workflow (AAW), also known as an LLM-based multi-agent system, is an autonomous system that assembles several LLM-based agents to work collaboratively towards a shared goal. The high autonomy, widespread adoption, and growing interest in such AAWs highlight the need for a deeper understanding of their operations, from both quality and security aspects. To this day, there are no existing methods to assess the influence of each agent on the AAW's final output. Adopting techniques from related fields is not feasible since existing methods perform only static structural analysis, which is unsuitable for inference time execution. We present Counterfactual-based Agent Influence Ranker (CAIR) - the first method for assessing the influence level of each agent on the AAW's output and determining which agents are the most influential. By performing counterfactual analysis, CAIR provides a task-agnostic analysis that can be used both offline and at inference time. We evaluate CAIR using an AAWs dataset of our creation, containing 30 different use cases with 230 different functionalities. Our evaluation showed that CAIR produces consistent rankings, outperforms baseline methods, and can easily enhance the effectiveness and relevancy of downstream tasks.

Formation control simplifies minimizing multi-robot cost functions by encoding a cost function as a shape the robots maintain. However, by reducing complex cost functions to formations, discrepancies arise between maintaining the shape and minimizing the original cost function. For example, a Diamond or Box formation shape is often used for protecting all members of the formation. When more information about the surrounding environment becomes available, a static shape often no longer minimizes the original protection cost. We propose a formation planner to reduce mismatch between a formation and the cost function while still leveraging efficient formation controllers. Our formation planner is a two-step optimization problem that identifies desired relative robot positions. We first solve a constrained problem to estimate non-linear and non-differentiable costs with a weighted sum of surrogate cost functions. We theoretically analyze this problem and identify situations where weights do not need to be updated. The weighted, surrogate cost function is then minimized using relative positions between robots. The desired relative positions are realized using a non-cooperative formation controller derived from Lyapunov's direct approach. We then demonstrate the efficacy of this approach for military-like costs such as protection and obstacle avoidance. In simulations, we show a formation planner can reduce a single cost by over 75%. When minimizing a variety of cost functions simultaneously, using a formation planner with adaptive weights can reduce the cost by 20-40%. Formation planning provides better performance by minimizing a surrogate cost function that closely approximates the original cost function instead of relying on a shape abstraction.

Multi-agent reinforcement learning (MARl) has achieved strong results in cooperative tasks but typically assumes fixed, fully controlled teams. Ad hoc teamwork (AHT) relaxes this by allowing collaboration with unknown partners, yet existing variants still presume shared conventions. We introduce Multil-party Ad Hoc Teamwork (MAHT), where controlled agents must coordinate with multiple mutually unfamiliar groups of uncontrolled teammates. To address this, we propose MARs, which builds a sparse skeleton graph and applies relational modeling to capture cross-group dvnamics. Experiments on MPE and starCralt ll show that MARs outperforms MARL and AHT baselines while converging faster.

Frequent natural disasters cause significant losses to human society, and timely, efficient collection of post-disaster environmental information is the foundation for effective rescue operations. Due to the extreme complexity of post-disaster environments, existing sensing technologies such as mobile crowdsensing suffer from weak environmental adaptability, insufficient professional sensing capabilities, and poor practicality of sensing solutions. Therefore, this paper explores a heterogeneous multi-agent online collaborative scheduling algorithm, HoCs-MPQ, to achieve efficient collection of post-disaster environmental information. HoCs-MPQ models collaboration and conflict relationships among multiple elements through weighted undirected graph construction, and iteratively solves the maximum weight independent set based on multi-priority queues, ultimately achieving collaborative sensing scheduling of time-dependent UA Vs, vehicles, and workers. Specifically, (1) HoCs-MPQ constructs weighted undirected graph nodes based on collaborative relationships among multiple elements and quantifies their weights, then models the weighted undirected graph based on conflict relationships between nodes; (2) HoCs-MPQ solves the maximum weight independent set based on iterated local search, and accelerates the solution process using multi-priority queues. Finally, we conducted detailed experiments based on extensive real-world and simulated data. The experiments show that, compared to baseline methods (e.g., HoCs-GREEDY, HoCs-K-WTA, HoCs-MADL, and HoCs-MARL), HoCs-MPQ improves task completion rates by an average of 54.13%, 23.82%, 14.12%, and 12.89% respectively, with computation time for single online autonomous scheduling decisions not exceeding 3 seconds.

We are given a bipartite graph $G = \left( A \cup B, E \right)$. In the one-sided model, every $a \in A$ (often called agents) ranks its neighbours $z \in N_{a}$ strictly, and no $b \in B$ has any preference order over its neighbours $y \in N_{b}$, and vertices in $B$ abstain from casting their votes to matchings. In the two-sided model with one-sided ties, every $a \in A$ ranks its neighbours $z \in N_{a}$ strictly, and every $b \in B$ puts all of its neighbours into a single large tie, i.e., $b \in B$ prefers every $y \in N_{b}$ equally. In this two-sided model with one-sided ties, when two matchings compete in a majority election, $b \in B$ abstains from casting its vote for a matching when both the matchings saturate $b$ or both leave $b$ unsaturated; else $b$ prefers the matching where it is saturated. A popular matching $M$ is \emph{robust} if it remains popular among multiple instances. We have analysed the cases when a robust popular matching exists in the one-sided model where only one agent alters her preference order among the instances, and we have proposed a polynomial-time algorithm to decide if there exists a robust popular matching when instances differ only with respect to the preference orders of a single agent. We give a simple characterisation of popular matchings in the two-sided model with one-sided ties. We show that in the two-sided model with one-sided ties, if the input instances differ only with respect to the preference orders of a single agent, there is a polynomial-time algorithm to decide whether there exists a robust popular matching. We have been able to decide the stable matching problem in bipartite graphs $G = (A \cup B, E)$ where \textit{both} sides have weak preferences (ties allowed), with the restriction that every tie has length at most $k$.

In the rapidly evolving research on artificial intelligence (AI) the demand for fast, computationally efficient, and scalable solutions has increased in recent years. The problem of optimizing the computing resources for distributed machine learning (ML) and optimization is considered in this paper. Given a set of data distributed over a network of computing-nodes/servers, the idea is to optimally assign the CPU (central processing unit) usage while simultaneously training each computing node locally via its own share of data. This formulates the problem as a co-optimization setup to (i) optimize the data processing and (ii) optimally allocate the computing resources. The information-sharing network among the nodes might be time-varying, but with balanced weights to ensure consensus-type convergence of the algorithm. The algorithm is all-time feasible, which implies that the computing resource-demand balance constraint holds at all iterations of the proposed solution. Moreover, the solution allows addressing possible log-scale quantization over the information-sharing channels to exchange log-quantized data. For some example applications, distributed support-vector-machine (SVM) and regression are considered as the ML training models. Results from perturbation theory, along with Lyapunov stability and eigen-spectrum analysis, are used to prove the convergence towards the optimal case. As compared to existing CPU scheduling solutions, the proposed algorithm improves the cost optimality gap by more than $50\%$.

Artificial Intelligence is reshaping America's \$9.4 trillion labor market, with cascading effects that extend far beyond visible technology sectors. When AI transforms quality control tasks in automotive plants, consequences spread through logistics networks, supply chains, and local service economies. Yet traditional workforce metrics cannot capture these ripple effects: they measure employment outcomes after disruption occurs, not where AI capabilities overlap with human skills before adoption crystallizes. Project Iceberg addresses this gap using Large Population Models to simulate the human-AI labor market, representing 151 million workers as autonomous agents executing over 32,000 skills and interacting with thousands of AI tools. It introduces the Iceberg Index, a skills-centered metric that measures the wage value of skills AI systems can perform within each occupation. The Index captures technical exposure, where AI can perform occupational tasks, not displacement outcomes or adoption timelines. Analysis shows that visible AI adoption concentrated in computing and technology (2.2% of wage value, approx \$211 billion) represents only the tip of the iceberg. Technical capability extends far below the surface through cognitive automation spanning administrative, financial, and professional services (11.7%, approx \$1.2 trillion). This exposure is fivefold larger and geographically distributed across all states rather than confined to coastal hubs. Traditional indicators such as GDP, income, and unemployment explain less than 5% of this skills-based variation, underscoring why new indices are needed to capture exposure in the AI economy. By simulating how these capabilities may spread under scenarios, Iceberg enables policymakers and business leaders to identify exposure hotspots, prioritize investments, and test interventions before committing billions to implementation