We propose a discrete time graphon game formulation on continuous state and action spaces using a representative player to study stochastic games with heterogeneous interaction among agents. This formulation admits both philosophical and mathematical advantages, compared to a widely adopted formulation using a continuum of players. We prove the existence and uniqueness of the graphon equilibrium with mild assumptions, and show that this equilibrium can be used to construct an approximate solution for finite player game on networks, which is challenging to analyze and solve due to curse of dimensionality. An online oracle-free learning algorithm is developed to solve the equilibrium numerically, and sample complexity analysis is provided for its convergence.

In this paper, we study safe data collection for the purpose of policy evaluation in tabular Markov decision processes (MDPs). In policy evaluation, we are given a \textit{target} policy and asked to estimate the expected cumulative reward it will obtain. Policy evaluation requires data and we are interested in the question of what \textit{behavior} policy should collect the data for the most accurate evaluation of the target policy. While prior work has considered behavior policy selection, in this paper, we additionally consider a safety constraint on the behavior policy. Namely, we assume there exists a known default policy that incurs a particular expected cost when run and we enforce that the cumulative cost of all behavior policies ran is better than a constant factor of the cost that would be incurred had we always run the default policy. We first show that there exists a class of intractable MDPs where no safe oracle algorithm with knowledge about problem parameters can efficiently collect data and satisfy the safety constraints. We then define the tractability condition for an MDP such that a safe oracle algorithm can efficiently collect data and using that we prove the first lower bound for this setting. We then introduce an algorithm SaVeR for this problem that approximates the safe oracle algorithm and bound the finite-sample mean squared error of the algorithm while ensuring it satisfies the safety constraint. Finally, we show in simulations that SaVeR produces low MSE policy evaluation while satisfying the safety constraint.

Predicting the behavior of AI-driven agents is particularly challenging without a preexisting model. In our paper, we address this by treating AI agents as nonlinear dynamical systems and adopting a probabilistic perspective to predict their statistical behavior using the Perron-Frobenius (PF) operator. We formulate the approximation of the PF operator as an entropy minimization problem, which can be solved by leveraging the Markovian property of the operator and decomposing its spectrum. Our data-driven methodology simultaneously approximates the PF operator to perform prediction of the evolution of the agents and also predicts the terminal probability density of AI agents, such as robotic systems and generative models. We demonstrate the effectiveness of our prediction model through extensive experiments on practical systems driven by AI algorithms.

The simulation-based testing is essential for safely implementing autonomous vehicles (AVs) on roads, necessitating simulated traffic environments that dynamically interact with the Vehicle Under Test (VUT). This study introduces a VUT-Centered environmental Dynamics Inference (VCDI) model for realistic, interactive, and diverse background traffic simulation. VCDI is built on a Transformer-based trajectory inference model to generate trajectories for background objects. Serving the purpose of AV testing, VCDI additionally considers VUT-centered interactivity and scenario diversity using a conditional inference framework. First, the VUT future motion is taken as an augmented model input to bridge the interaction between VUT and background objects. Second, to enrich the scenario diversity, a Bayesian-network-based cost function module is designed. The module, learned in a distributional form, captures the uncertainty of the VUT's strategy, triggering various scenario evolution. Experimental results validate VCDI's trajectory-level simulation precision which outperforms the state-of-the-art trajectory prediction work. The flexibility of the distributional cost function allows VCDI to provide diverse-yet-realistic scenarios for AV testing. We demonstrate such capability by modifying the anticipation to VUT's cost-based strategy and thus achieve multiple testing scenarios with explainable background traffic evolution.

Sample efficiency remains a key challenge in multi-agent reinforcement learning (MARL). A promising approach is to learn a meaningful latent representation space through auxiliary learning objectives alongside the MARL objective to aid in learning a successful control policy. In our work, we present MAPO-LSO (Multi-Agent Policy Optimization with Latent Space Optimization) which applies a form of comprehensive representation learning devised to supplement MARL training. Specifically, MAPO-LSO proposes a multi-agent extension of transition dynamics reconstruction and self-predictive learning that constructs a latent state optimization scheme that can be trivially extended to current state-of-the-art MARL algorithms. Empirical results demonstrate MAPO-LSO to show notable improvements in sample efficiency and learning performance compared to its vanilla MARL counterpart without any additional MARL hyperparameter tuning on a diverse suite of MARL tasks.

Vehicle motion planning is an essential component of autonomous driving technology. Current rule-based vehicle motion planning methods perform satisfactorily in common scenarios but struggle to generalize to long-tailed situations. Meanwhile, learning-based methods have yet to achieve superior performance over rule-based approaches in large-scale closed-loop scenarios. To address these issues, we propose PlanAgent, the first mid-to-mid planning system based on a Multi-modal Large Language Model (MLLM). MLLM is used as a cognitive agent to introduce human-like knowledge, interpretability, and common-sense reasoning into the closed-loop planning. Specifically, PlanAgent leverages the power of MLLM through three core modules. First, an Environment Transformation module constructs a Bird's Eye View (BEV) map and a lane-graph-based textual description from the environment as inputs. Second, a Reasoning Engine module introduces a hierarchical chain-of-thought from scene understanding to lateral and longitudinal motion instructions, culminating in planner code generation. Last, a Reflection module is integrated to simulate and evaluate the generated planner for reducing MLLM's uncertainty. PlanAgent is endowed with the common-sense reasoning and generalization capability of MLLM, which empowers it to effectively tackle both common and complex long-tailed scenarios. Our proposed PlanAgent is evaluated on the large-scale and challenging nuPlan benchmarks. A comprehensive set of experiments convincingly demonstrates that PlanAgent outperforms the existing state-of-the-art in the closed-loop motion planning task. Codes will be soon released.

To adapt cities to the issues of climate change and public health, urban policies are trying to encourage soft mobility [14] in order to reduce traffic and pollution, via financial incentives or new infrastructure. However, mobility evolves very slowly, and the share of the car remains significant (74% in France [9]), despite increased public awareness of global warming, and increased concern for ecology. The pandemic offered an opportunity to explore the impact of reduced car mobility and new urban planning policies, for instance with temporary cycle paths [19]. But these public policies normally take longer to implement and are not always well accepted by the car-loving population; many of these temporary cycle paths were gradually returned to cars after the end of the lockdowns [6]. Many explaining factors of this inertia of mobility and reluctance to shift from the car are already known, both contextual, such as a lack of alternatives (limited public transportation options), individual constraints (transporting children or tools), or higher costs of newer or electric vehicles...); and psychological, such as the difficulty to change habits [8, 17], individualism [12], or influence of cognitive biases [15, 13].

Asynchronous Federated Learning (AFL) methods have emerged as promising alternatives to their synchronous counterparts bounded by the slowest agent, yet they add additional challenges in convergence guarantees, fairness with respect to compute heterogeneity, and incorporation of staleness in aggregated updates. Specifically, AFL biases model training heavily towards agents who can produce updates faster, leaving slower agents behind, who often also have differently distributed data which is not learned by the global model. Naively upweighting introduces incentive issues, where true fast updating agents may falsely report updates at a slower speed to increase their contribution to model training. We introduce FedStaleWeight, an algorithm addressing fairness in aggregating asynchronous client updates by employing average staleness to compute fair re-weightings. FedStaleWeight reframes asynchronous federated learning aggregation as a mechanism design problem, devising a weighting strategy that incentivizes truthful compute speed reporting without favoring faster update-producing agents by upweighting agent updates based on staleness. Leveraging only observed agent update staleness, FedStaleWeight results in more equitable aggregation on a per-agent basis. We both provide theoretical convergence guarantees in the smooth, non-convex setting and empirically compare FedStaleWeight against the commonly used asynchronous FedBuff with gradient averaging, demonstrating how it achieves stronger fairness, expediting convergence to a higher global model accuracy. Finally, we provide an open-source test bench to facilitate exploration of buffered AFL aggregation strategies, fostering further research in asynchronous federated learning paradigms.

In the rapidly evolving landscape of eCommerce, Artificial Intelligence (AI) based pricing algorithms, particularly those utilizing Reinforcement Learning (RL), are becoming increasingly prevalent. This rise has led to an inextricable pricing situation with the potential for market collusion. Our research employs an experimental oligopoly model of repeated price competition, systematically varying the environment to cover scenarios from basic economic theory to subjective consumer demand preferences. We also introduce a novel demand framework that enables the implementation of various demand models, allowing for a weighted blending of different models. In contrast to existing research in this domain, we aim to investigate the strategies and emerging pricing patterns developed by the agents, which may lead to a collusive outcome. Furthermore, we investigate a scenario where agents cannot observe their competitors' prices. Finally, we provide a comprehensive legal analysis across all scenarios. Our findings indicate that RL-based AI agents converge to a collusive state characterized by the charging of supracompetitive prices, without necessarily requiring inter-agent communication. Implementing alternative RL algorithms, altering the number of agents or simulation settings, and restricting the scope of the agents' observation space does not significantly impact the collusive market outcome behavior.

Human-robot collaboration, in which the robot intelligently assists the human with the upcoming task, is an appealing objective. To achieve this goal, the agent needs to be equipped with a fundamental collaborative navigation ability, where the agent should reason human intention by observing human activities and then navigate to the human's intended destination in advance of the human. However, this vital ability has not been well studied in previous literature. To fill this gap, we propose a collaborative navigation (CoNav) benchmark. Our CoNav tackles the critical challenge of constructing a 3D navigation environment with realistic and diverse human activities. To achieve this, we design a novel LLM-based humanoid animation generation framework, which is conditioned on both text descriptions and environmental context. The generated humanoid trajectory obeys the environmental context and can be easily integrated into popular simulators. We empirically find that the existing navigation methods struggle in CoNav task since they neglect the perception of human intention. To solve this problem, we propose an intention-aware agent for reasoning both long-term and short-term human intention. The agent predicts navigation action based on the predicted intention and panoramic observation. The emergent agent behavior including observing humans, avoiding human collision, and navigation reveals the efficiency of the proposed datasets and agents.