Interaction-Oriented Programming (IOP) is an approach to building a multiagent system by modeling the interactions between its roles via a flexible interaction protocol and implementing agents to realize the interactions of the roles they play in the protocol. In recent years, we have developed an extensive suite of software that enables multiagent system developers to apply IOP. These include tools for efficiently verifying protocols for properties such as liveness and safety and middleware that simplifies the implementation of agents. This paper presents some of that software suite.

Existing asynchronous MARL methods based on MacDec-POMDP typically construct training trajectory buffers by simply sampling limited and biased data at the endpoints of macro-actions, and directly apply conventional MARL methods on the buffers. As a result, these methods lead to an incomplete and inaccurate representation of the macro-action execution process, along with unsuitable credit assignments. To solve these problems, the T emporal Macro-action Value F actorization (ToMacVF) is proposed to achieve fine-grained temporal credit assignment for macro-action contributions. A centralized training buffer, called Macro-action S egmented Joint Experience R eplay Trajectory (Mac-SJERT), is designed to incorporate with ToMacVF to collect accurate and complete macro-action execution information, supporting a more comprehensive and precise representation of the macro-action process. To ensure principled and fine-grained asynchronous value factorization, the consistency requirement between joint and individual macro-action selection called Tempo ral Mac ro-action based IGM (To-Mac-IGM) is formalized, proving that it generalizes the synchronous cases. Based on To-Mac-IGM, a modularized ToMacVF architecture, which satisfies CTDE principle, is designed to conveniently integrate previous value factorization methods. Next, the ToMacVF algorithm is devised as an implementation of the ToMacVF architecture. Experimental results demonstrate that, compared to asynchronous baselines, our ToMacVF algorithm not only achieves optimal performance but also exhibits strong adaptability and robustness across various asynchronous multi-agent experimental scenarios.

We propose a novel cooperative herding strategy through backstepping control barrier functions (CBFs), which coordinates multiple herders to herd a group of evaders safely towards a designated goal region. For the herding system with heterogeneous groups involving herders and evaders, the behavior of the evaders can only be influenced indirectly by the herders' motion, especially when the evaders follow an inverse dynamics model and respond solely to repulsive interactions from the herders. This indirect interaction mechanism inherently renders the overall system underactuated. To address this issue, we first construct separate CBFs for the dual objectives of goal reaching and collision avoidance, which ensure both herding completion and safety guarantees. Then, we reformulate the underactuated herding dynamics into a control-affine structure and employ a backstepping approach to recursively design control inputs for the hierarchical barrier functions, avoiding taking derivatives of the higher-order system. Finally, we present a cooperative herding strategy based on backstepping CBFs that allow herders to safely herd multiple evaders into the goal region. In addition, centralized and decentralized implementations of the proposed algorithm are developed, further enhancing its flexibility and applicability. Extensive simulations and real-world experiments validate the effectiveness and safety of the proposed strategy in multi-robot herding.

We introduce RegionFocus, a visual test-time scaling approach for Vision Language Model Agents. Understanding webpages is challenging due to the visual complexity of GUI images and the large number of interface elements, making accurate action selection difficult. Our approach dynamically zooms in on relevant regions, reducing background clutter and improving grounding accuracy. To support this process, we propose an image-as-map mechanism that visualizes key landmarks at each step, providing a transparent action record and enables the agent to effectively choose among action candidates. Even with a simple region selection strategy, we observe significant performance gains of 28+\% on Screenspot-pro and 24+\% on WebVoyager benchmarks on top of two state-of-the-art open vision language model agents, UI-TARS and Qwen2.5-VL, highlighting the effectiveness of visual test-time scaling in interactive settings. We achieve a new state-of-the-art grounding performance of 61.6\% on the ScreenSpot-Pro benchmark by applying RegionFocus to a Qwen2.5-VL-72B model. Our code will be released publicly at https://github.com/tiangeluo/RegionFocus.

Cognitive abilities, which include ideas like intentionality and consciousness, have long been viewed in Western philosophy as exclusive to the human realm. Intent is roundly considered justifiable only with minimum requirements for self-awareness or situational comprehension. However, such hard line views have softened gradually with modern enlightenment, and more of us are likely to accept that terms such as'agency', 'intelligence', and even'emotion' can apply for other species too. Even plants lean into sunlight in an intentional way; the identification of an intention doesn't have to arise from the plant to be true. Latterly their possibility has been extended even to artificial systems, which some find more acceptable, though a modern version of the privilege argument persists in a distinction between'simple' machinery and'complex' biology, which many believe still holds some principled leap in understanding. Ideological'blood-brain barriers', like these, continue to undermine efforts to form a rational causal explanation of intent, leading extremists to clutch at esoteric straws like quantum mechanics or complexity theory to account for perceived magic. In this note, I address another apparent schism that may shed light on these questions: the difference between process dynamics (the realm of physics) and interpretive semantics (the realm of linguistics and philosophy), and the suggestion that (deep down) intentionality might be a relatively simple phenomenon with an energetic explanation (as trust has been shown to be [9]). The recent acceptance of attention mechanisms in Large Language Models is related example [19, 22].

In heterogeneous multi-agent reinforcement learning (MARL), achieving monotonic improvement plays a pivotal role in enhancing performance. The HAPPO algorithm proposes a feasible solution by introducing a sequential update scheme, which requires independent learning with No Parameter-sharing (NoPS). However, heterogeneous MARL generally requires Partial Parameter-sharing (ParPS) based on agent grouping to achieve high cooperative performance. Our experiments prove that directly combining ParPS with the sequential update scheme leads to the policy updating baseline drift problem, thereby failing to achieve improvement. To solve the conflict between monotonic improvement and ParPS, we propose the Optimal Marginal Deterministic Policy Gradient (OMDPG) algorithm. First, we replace the sequentially computed $Q_ψ^s(s,a_{1:i})$ with the Optimal Marginal Q (OMQ) function $ϕ_ψ^*(s,a_{1:i})$ derived from Q-functions. This maintains MAAD's monotonic improvement while eliminating the conflict through optimal joint action sequences instead of sequential policy ratio calculations. Second, we introduce the Generalized Q Critic (GQC) as the critic function, employing pessimistic uncertainty-constrained loss to optimize different Q-value estimations. This provides the required Q-values for OMQ computation and stable baselines for actor updates. Finally, we implement a Centralized Critic Grouped Actor (CCGA) architecture that simultaneously achieves ParPS in local policy networks and accurate global Q-function computation. Experimental results in SMAC and MAMuJoCo environments demonstrate that OMDPG outperforms various state-of-the-art MARL baselines.

Many of society's most pressing challenges, from pandemic response to supply chain disruptions to climate adaptation, emerge from the collective behavior of millions of autonomous agents making decisions over time. Large Population Models (LPMs) offer an approach to understand these complex systems by simulating entire populations with realistic behaviors and interactions at unprecedented scale. LPMs extend traditional modeling approaches through three key innovations: computational methods that efficiently simulate millions of agents simultaneously, mathematical frameworks that learn from diverse real-world data streams, and privacy-preserving communication protocols that bridge virtual and physical environments. This allows researchers to observe how agent behavior aggregates into system-level outcomes and test interventions before real-world implementation. While current AI advances primarily focus on creating "digital humans" with sophisticated individual capabilities, LPMs develop "digital societies" where the richness of interactions reveals emergent phenomena. By bridging individual agent behavior and population-scale dynamics, LPMs offer a complementary path in AI research illuminating collective intelligence and providing testing grounds for policies and social innovations before real-world deployment. We discuss the technical foundations and some open problems here. LPMs are implemented by the AgentTorch framework (github.com/AgentTorch/AgentTorch)

Autonomous vehicles (AVs) are becoming increasingly popular, with their applications now extending beyond just a mode of transportation to serving as mobile actuators of a traffic flow to control flow dynamics. This contrasts with traditional fixed-location actuators, such as traffic signals, and is referred to as Lagrangian traffic control. However, designing effective Lagrangian traffic control policies for AVs that generalize across traffic scenarios introduces a major challenge. Real-world traffic environments are highly diverse, and developing policies that perform robustly across such diverse traffic scenarios is challenging. It is further compounded by the joint complexity of the multi-agent nature of traffic systems, mixed motives among participants, and conflicting optimization objectives subject to strict physical and external constraints. To address these challenges, we introduce Multi-Residual Mixture of Expert Learning (MRMEL), a novel framework for Lagrangian traffic control that augments a given suboptimal nominal policy with a learned residual while explicitly accounting for the structure of the traffic scenario space. In particular, taking inspiration from residual reinforcement learning, MRMEL augments a suboptimal nominal AV control policy by learning a residual correction, but at the same time dynamically selects the most suitable nominal policy from a pool of nominal policies conditioned on the traffic scenarios and modeled as a mixture of experts. We validate MRMEL using a case study in cooperative eco-driving at signalized intersections in Atlanta, Dallas Fort Worth, and Salt Lake City, with real-world data-driven traffic scenarios. The results show that MRMEL consistently yields superior performance-achieving an additional 4%-9% reduction in aggregate vehicle emissions relative to the strongest baseline in each setting.

Recent advances in Large Language Models (LLM) have led to a new class of autonomous agents, renewing and expanding interest in the area. LLM-powered Multiagent Systems (MAS) have thus emerged, both for assistive and simulation purposes, yet tools for realistic human behavior simulation -- with its distinctive challenges and opportunities -- remain underdeveloped. Existing MAS libraries and tools lack fine-grained persona specifications, population sampling facilities, experimentation support, and integrated validation, among other key capabilities, limiting their utility for behavioral studies, social simulation, and related applications. To address these deficiencies, in this work we introduce TinyTroupe, a simulation toolkit enabling detailed persona definitions (e.g., nationality, age, occupation, personality, beliefs, behaviors) and programmatic control via numerous LLM-driven mechanisms. This allows for the concise formulation of behavioral problems of practical interest, either at the individual or group level, and provides effective means for their solution. TinyTroupe's components are presented using representative working examples, such as brainstorming and market research sessions, thereby simultaneously clarifying their purpose and demonstrating their usefulness. Quantitative and qualitative evaluations of selected aspects are also provided, highlighting possibilities, limitations, and trade-offs. The approach, though realized as a specific Python implementation, is meant as a novel conceptual contribution, which can be partially or fully incorporated in other contexts. The library is available as open source at https://github.com/microsoft/tinytroupe.

Artificial intelligence (AI) systems often interact with multiple agents. The regulation of such AI systems often requires that {\em a priori\/} guarantees of fairness and robustness be satisfied. With stochastic models of agents' responses to the outputs of AI systems, such {\em a priori\/} guarantees require non-trivial reasoning about the corresponding stochastic systems. Here, we present an open-source PyTorch-based toolkit for the use of stochastic control techniques in modelling interconnections of AI systems and properties of their repeated uses. It models robustness and fairness desiderata in a closed-loop fashion, and provides {\em a priori\/} guarantees for these interconnections. The PyTorch-based toolkit removes much of the complexity associated with the provision of fairness guarantees for closed-loop models of multi-agent systems.