Partial observability of the underlying states generally presents significant challenges for reinforcement learning (RL). In practice, certain \emph{privileged information}, e.g., the access to states from simulators, has been exploited in training and has achieved prominent empirical successes. To better understand the benefits of privileged information, we revisit and examine several simple and practically used paradigms in this setting. Specifically, we first formalize the empirical paradigm of \emph{expert distillation} (also known as \emph{teacher-student} learning), demonstrating its pitfall in finding near-optimal policies. We then identify a condition of the partially observable environment, the \emph{deterministic filter condition}, under which expert distillation achieves sample and computational complexities that are \emph{both} polynomial. Furthermore, we investigate another useful empirical paradigm of \emph{asymmetric actor-critic}, and focus on the more challenging setting of observable partially observable Markov decision processes. We develop a belief-weighted asymmetric actor-critic algorithm with polynomial sample and quasi-polynomial computational complexities, in which one key component is a new provable oracle for learning belief states that preserve \emph{filter stability} under a misspecified model, which may be of independent interest. Finally, we also investigate the provable efficiency of partially observable multi-agent RL (MARL) with privileged information. We develop algorithms featuring \emph{centralized-training-with-decentralized-execution}, a popular framework in empirical MARL, with polynomial sample and (quasi-)polynomial computational complexities in both paradigms above. Compared with a few recent related theoretical studies, our focus is on understanding practically inspired algorithmic paradigms, without computationally intractable oracles.

Traditionally, cognition has been considered a uniquely human capability involving perception, memory, learning, reasoning, and problem-solving. However, recent research shows that cognition is a fundamental ability shared by all living beings, from single cells to complex organisms. This chapter takes an info-computational approach (ICON), viewing natural structures as information and the processes of change in these structures as computations. It is a relational framework dependent on the perspective of a cognizing observer/cognizer. Informational structures are properties of the material substrate, and when focusing on the behavior of the substrate, we discuss morphological computing (MC). ICON and MC are complementary perspectives for a cognizer. Information and computation are inseparably connected with cognition. This chapter explores research connecting nature as a computational structure for a cognizer, with morphological computation, morphogenesis, agency, extended cognition, and extended evolutionary synthesis, using examples of the free energy principle and active inference. It introduces theoretical and practical approaches challenging traditional computational models of cognition limited to abstract symbol processing, highlighting the computational capacities inherent in the material substrate (embodiment). Understanding the embodiment of cognition through its morphological computational basis is crucial for biology, evolution, intelligence theory, AI, robotics, and other fields.

Dynamic game theory is an increasingly popular tool for modeling multi-agent, e.g. human-robot, interactions. Game-theoretic models presume that each agent wishes to minimize a private cost function that depends on others' actions. These games typically evolve over a fixed time horizon, which specifies the degree to which all agents care about the distant future. In practical settings, however, decision-makers may vary in their degree of short-sightedness. We conjecture that quantifying and estimating each agent's short-sightedness from online data will enable safer and more efficient interactions with other agents. To this end, we frame this inference problem as an inverse dynamic game. We consider a specific parametrization of each agent's objective function that smoothly interpolates myopic and farsighted planning. Games of this form are readily transformed into parametric mixed complementarity problems; we exploit the directional differentiability of solutions to these problems with respect to their hidden parameters in order to solve for agents' short-sightedness. We conduct several experiments simulating human behavior at a real-world crosswalk. The results of these experiments clearly demonstrate that by explicitly inferring agents' short-sightedness, we can recover more accurate game-theoretic models, which ultimately allow us to make better predictions of agents' behavior. Specifically, our results show up to a 30% more accurate prediction of myopic behavior compared to the baseline.

As the ubiquity of smart mobile devices continues to rise, Optical Camera Communication systems have gained more attention as a solution for efficient and private data streaming. This system utilizes optical cameras to receive data from digital screens via visible light. Despite their promise, most of them are hindered by dynamic factors such as screen refreshing and rapid camera motion. CMOS cameras, often serving as the receivers, suffer from limited frame rates and motion-induced image blur, which degrade overall performance. To address these challenges, this paper unveils a novel system that utilizes event cameras. We introduce a dynamic visual marker and design event-based tracking algorithms to achieve fast localization and data streaming. Remarkably, the event camera's unique capabilities mitigate issues related to screen refresh rates and camera motion, enabling a high throughput of up to 114 Kbps in static conditions, and a 1 cm localization accuracy with 1% bit error rate under various camera motions.

InvestESG is a novel multi-agent reinforcement learning (MARL) benchmark designed to study the impact of Environmental, Social, and Governance (ESG) disclosure mandates on corporate climate investments. Supported by both PyTorch and JAX implementation, the benchmark models an intertemporal social dilemma where companies balance short-term profit losses from climate mitigation efforts and long-term benefits from reducing climate risk, while ESG-conscious investors attempt to influence corporate behavior through their investment decisions, in a scalable and hardware-accelerated manner. Companies allocate capital across mitigation, greenwashing, and resilience, with varying strategies influencing climate outcomes and investor preferences. Our experiments show that without ESG-conscious investors with sufficient capital, corporate mitigation efforts remain limited under the disclosure mandate. However, when a critical mass of investors prioritizes ESG, corporate cooperation increases, which in turn reduces climate risks and enhances long-term financial stability. Additionally, providing more information about global climate risks encourages companies to invest more in mitigation, even without investor involvement. Our findings align with empirical research using real-world data, highlighting MARL's potential to inform policy by providing insights into large-scale socio-economic challenges through efficient testing of alternative policy and market designs.

Large Language Models (LLMs) often perpetuate biases in pronoun usage, leading to misrepresentation or exclusion of queer individuals. This paper addresses the specific problem of biased pronoun usage in LLM outputs, particularly the inappropriate use of traditionally gendered pronouns ("he," "she") when inclusive language is needed to accurately represent all identities. We introduce a collaborative agent pipeline designed to mitigate these biases by analyzing and optimizing pronoun usage for inclusivity. Our multi-agent framework includes specialized agents for both bias detection and correction. Experimental evaluations using the Tango dataset-a benchmark focused on gender pronoun usage-demonstrate that our approach significantly improves inclusive pronoun classification, achieving a 32.6 percentage point increase over GPT-4o in correctly disagreeing with inappropriate traditionally gendered pronouns $(\chi^2 = 38.57, p < 0.0001)$. These results accentuate the potential of agent-driven frameworks in enhancing fairness and inclusivity in AI-generated content, demonstrating their efficacy in reducing biases and promoting socially responsible AI.

This study aims to place Lorenzo Magnanis Eco-Cognitive Computationalism within the broader context of current work on information, computation, and cognition. Traditionally, cognition was believed to be exclusive to humans and a result of brain activity. However, recent studies reveal it as a fundamental characteristic of all life forms, ranging from single cells to complex multicellular organisms and their networks. Yet, the literature and general understanding of cognition still largely remain human-brain-focused, leading to conceptual gaps and incoherency. This paper presents a variety of computational (information processing) approaches, including an info-computational approach to cognition, where natural structures represent information and dynamical processes on natural structures are regarded as computation, relative to an observing cognizing agent. We model cognition as a web of concurrent morphological computations, driven by processes of self-assembly, self-organisation, and autopoiesis across physical, chemical, and biological domains. We examine recent findings linking morphological computation, morphogenesis, agency, basal cognition, extended evolutionary synthesis, and active inference. We establish a connection to Magnanis Eco-Cognitive Computationalism and the idea of computational domestication of ignorant entities. Novel theoretical and applied insights question the boundaries of conventional computational models of cognition. The traditional models prioritize symbolic processing and often neglect the inherent constraints and potentialities in the physical embodiment of agents on different levels of organization. Gaining a better info-computational grasp of cognitive embodiment is crucial for the advancement of fields such as biology, evolutionary studies, artificial intelligence, robotics, medicine, and more.

In this paper, we introduce an algorithm designed to address the problem of time-optimal formation reshaping in three-dimensional environments while preventing collisions between agents. The utility of the proposed approach is particularly evident in mobile robotics, where agents benefit from being organized and navigated in formation for a variety of real-world applications requiring frequent alterations in formation shape for efficient navigation or task completion. Given the constrained operational time inherent to battery-powered mobile robots, the time needed to complete the formation reshaping process is crucial for their efficient operation, especially in case of multi-rotor Unmanned Aerial Vehicles (UAVs). The proposed Collision-Aware Time-Optimal formation Reshaping Algorithm (CAT-ORA) builds upon the Hungarian algorithm for the solution of the robot-to-goal assignment implementing the inter-agent collision avoidance through direct constraints on mutually exclusive robot-goal pairs combined with a trajectory generation approach minimizing the duration of the reshaping process. Theoretical validations confirm the optimality of CAT-ORA, with its efficacy further showcased through simulations, and a real-world outdoor experiment involving 19 UAVs. Thorough numerical analysis shows the potential of CAT-ORA to decrease the time required to perform complex formation reshaping tasks by up to 49%, and 12% on average compared to commonly used methods in randomly generated scenarios.

Reinforcement Learning (RL) has become a popular learning framework to solve sequential decision making problems in unknown environments, and has achieved tremendous success in a wide array of domains such as playing the game of Go (Silver et al., 2016), robotic control (Kober et al., 2013), and autonomous driving (Kiran et al., 2022; Lin et al., 2023). A critical feature of most real-world systems is their uncertain nature, and consequently RL has emerged as a powerful tool for learning optimal policies for multi-agent systems to operate in unknown environments (Kim & Giannakis, 2017; Zhang et al., 2021; Lin et al., 2024; Anand & Qu, 2024). While the early literature on RL predominantly focused on the single-agent setting, multi-agent reinforcement learning (MARL) has also recently achieved impressive successes in a broad range of areas, such as coordination of robotic swarms (Preiss et al., 2017), self-driving vehicles (DeWeese & Qu, 2024), real-time bidding (Jin et al., 2018), ride-sharing (Li et al., 2019), and stochastic games (Jin et al., 2020). Despite growing interest in multi-agent RL (MARL), extending RL to multi-agent settings poses significant computational challenges due to the curse of dimensionality (Sayin et al., 2021). Even if the individual agents' state or action spaces are small, the global state space or action space can take values from a set with size that is exponentially large as a function of the number of agents.

Advancements in Large Language Models (LLMs) have opened transformative possibilities for human-robot interaction, especially in collaborative environments. However, Real-time human-AI collaboration requires agents to adapt to unseen human behaviors while maintaining effective communication dynamically. Existing benchmarks fall short in evaluating such adaptability for embodied agents, focusing mostly on the task performance of the agent itself. To address this gap, we propose a novel benchmark that assesses agents' reactive adaptability and instantaneous communication capabilities at every step. Based on this benchmark, we propose a Monitor-then-Adapt framework (MonTA), combining strong adaptability and communication with real-time execution. MonTA contains three key LLM modules, a lightweight \textit{Monitor} for monitoring the need for adaptation in high frequency, and two proficient \textit{Adapters} for subtask and path adaptation reasoning in low frequency. Our results demonstrate that MonTA outperforms other baseline agents on our proposed benchmark. Further user studies confirm the high reasonability adaptation plan and consistent language instruction provided by our framework.