In today's networked world, Digital Twin Networks (DTNs) are revolutionizing how we understand and optimize physical networks. These networks, also known as 'Digital Twin Networks (DTNs)' or 'Networks Digital Twins (NDTs),' encompass many physical networks, from cellular and wireless to optical and satellite. They leverage computational power and AI capabilities to provide virtual representations, leading to highly refined recommendations for real-world network challenges. Within DTNs, tasks include network performance enhancement, latency optimization, energy efficiency, and more. To achieve these goals, DTNs utilize AI tools such as Machine Learning (ML), Deep Learning (DL), Reinforcement Learning (RL), Federated Learning (FL), and graph-based approaches. However, data quality, scalability, interpretability, and security challenges necessitate strategies prioritizing transparency, fairness, privacy, and accountability. This chapter delves into the world of AI-driven traffic analysis within DTNs. It explores DTNs' development efforts, tasks, AI models, and challenges while offering insights into how AI can enhance these dynamic networks. Through this journey, readers will gain a deeper understanding of the pivotal role AI plays in the ever-evolving landscape of networked systems.

Reinforcement learning (RL) is a powerful machine learning technique that has been successfully applied to a wide variety of problems. However, it can be unpredictable and produce suboptimal results in complicated learning environments. This is especially true when multiple agents learn simultaneously, which creates a complex system that is often analytically intractable. Our work considers the fundamental framework of Q-learning in Public Goods Games, where RL individuals must work together to achieve a common goal. This setting allows us to study the tragedy of the commons and free rider effects in AI cooperation, an emerging field with potential to resolve challenging obstacles to the wider application of artificial intelligence. While this social dilemma has been mainly investigated through traditional and evolutionary game theory, our approach bridges the gap between these two by studying agents with an intermediate level of intelligence. Specifically, we consider the influence of learning parameters on cooperation levels in simulations and a limiting system of differential equations, as well as the effect of evolutionary pressures on exploration rate in both of these models. We find selection for higher and lower levels of exploration, as well as attracting values, and a condition that separates these in a restricted class of games. Our work enhances the theoretical understanding of evolutionary Q-learning, and extends our knowledge of the evolution of machine behavior in social dilemmas.

This paper introduces a new approach to enhance the robustness of humanoid walking under strong perturbations, such as substantial pushes. Effective recovery from external disturbances requires bipedal robots to dynamically adjust their stepping strategies, including footstep positions and timing. Unlike most advanced walking controllers that restrict footstep locations to a predefined convex region, substantially limiting recoverable disturbances, our method leverages reinforcement learning to dynamically adjust the permissible footstep region, expanding it to a larger, effectively non-convex area and allowing cross-over stepping, which is crucial for counteracting large lateral pushes. Additionally, our method adapts footstep timing in real time to further extend the range of recoverable disturbances. Based on these adjustments, feasible footstep positions and DCM trajectory are planned by solving a QP. Finally, we employ a DCM controller and an inverse dynamics whole-body control framework to ensure the robot effectively follows the trajectory.

We study the problem of estimating the \emph{value} of the largest mean among $K$ distributions via samples from them (rather than estimating \emph{which} distribution has the largest mean), which arises from various machine learning tasks including Q-learning and Monte Carlo tree search. While there have been a few proposed algorithms, their performance analyses have been limited to their biases rather than a precise error metric. In this paper, we propose a novel algorithm called HAVER (Head AVERaging) and analyze its mean squared error. Our analysis reveals that HAVER has a compelling performance in two respects. First, HAVER estimates the maximum mean as well as the oracle who knows the identity of the best distribution and reports its sample mean. Second, perhaps surprisingly, HAVER exhibits even better rates than this oracle when there are many distributions near the best one. Both of these improvements are the first of their kind in the literature, and we also prove that the naive algorithm that reports the largest empirical mean does not achieve these bounds. Finally, we confirm our theoretical findings via numerical experiments including bandits and Q-learning scenarios where HAVER outperforms baseline methods.

In offline reinforcement learning (RL), addressing the out-of-distribution (OOD) action issue has been a focus, but we argue that there exists an OOD state issue that also impairs performance yet has been underexplored. Such an issue describes the scenario when the agent encounters states out of the offline dataset during the test phase, leading to uncontrolled behavior and performance degradation. To this end, we propose SCAS, a simple yet effective approach that unifies OOD state correction and OOD action suppression in offline RL. Technically, SCAS achieves value-aware OOD state correction, capable of correcting the agent from OOD states to high-value in-distribution states. Theoretical and empirical results show that SCAS also exhibits the effect of suppressing OOD actions. On standard offline RL benchmarks, SCAS achieves excellent performance without additional hyperparameter tuning. Moreover, benefiting from its OOD state correction feature, SCAS demonstrates enhanced robustness against environmental perturbations.

Decision Transformers have recently emerged as a new and compelling paradigm for offline Reinforcement Learning (RL), completing a trajectory in an autoregressive way. While improvements have been made to overcome initial shortcomings, online finetuning of decision transformers has been surprisingly under-explored. The widely adopted state-of-the-art Online Decision Transformer (ODT) still struggles when pretrained with low-reward offline data. In this paper, we theoretically analyze the online-finetuning of the decision transformer, showing that the commonly used Return-To-Go (RTG) that's far from the expected return hampers the online fine-tuning process. This problem, however, is well-addressed by the value function and advantage of standard RL algorithms. As suggested by our analysis, in our experiments, we hence find that simply adding TD3 gradients to the finetuning process of ODT effectively improves the online finetuning performance of ODT, especially if ODT is pretrained with low-reward offline data. These findings provide new directions to further improve decision transformers.

A financial market is a marketplace where financial instruments such as stocks and bonds are bought and sold (Fama 1970). Individuals and organizations can play crucial roles in financial markets to facilitate the allocation of capital. Market participants face diverse challenges, such as portfolio management, which aims to maximize investment returns over time, and market-making, which seeks to profit from the bid-ask spread while managing inventory risk. As the volume of financial data has increased dramatically over time, new opportunities and challenges have arisen in the analysis process, leading to the increased adoption of advanced Machine Learning (ML) models. Reinforcement Learning (RL)(Sutton & Barto 2018), as one of the main categories of ML, has revolutionized the field of artificial intelligence by empowering agents to interact with the environment and allowing them to learn and improve their performance. The success of RL has been demonstrated in various fields, including games, robots, mobile health (Nash Jr 1950, Kalman 1960, Murphy 2003), etc. In finance, applications such as market making, portfolio management, and order execution can benefit from the ability of RL algorithms to learn and adapt to changing environments. Compared to traditional models that rely on statistical techniques and econometric methods such as time series models (ARMA, ARIMA), factor models, and panel models, the RL framework empowers agents to learn decision-making by interacting with an environment and deducing the consequences of past actions to maximize cumulative rewards (Charpentier et al. 2021).

The use of reinforcement learning (RL) to learn policies for just-in-time adaptive interventions (JITAIs) is of significant interest in many behavioral intervention domains including improving levels of physical activity. In a messaging-based physical activity JITAI, a mobile health app is typically used to send messages to a participant to encourage engagement in physical activity. In this setting, RL methods can be used to learn what intervention options to provide to a participant in different contexts. However, deploying RL methods in real physical activity adaptive interventions comes with challenges: the cost and time constraints of real intervention studies result in limited data to learn adaptive intervention policies. Further, commonly used RL simulation environments have dynamics that are of limited relevance to physical activity adaptive interventions and thus shed little light on what RL methods may be optimal for this challenging application domain. In this paper, we introduce StepCountJITAI, an RL environment designed to foster research on RL methods that address the significant challenges of policy learning for adaptive behavioral interventions.

Building quantum circuits that perform a given task is a notoriously difficult problem. Reinforcement learning has proven to be a powerful approach, but many limitations remain due to the exponential scaling of the space of possible operations on qubits. In this paper, we develop an algorithm that automatically learns composite gates ("$gadgets$") and adds them as additional actions to the reinforcement learning agent to facilitate the search, namely the Gadget Reinforcement Learning (GRL) algorithm. We apply our algorithm to finding parameterized quantum circuits (PQCs) that implement the ground state of a given quantum Hamiltonian, a well-known NP-hard challenge. In particular, we focus on the transverse field Ising model (TFIM), since understanding its ground state is crucial for studying quantum phase transitions and critical behavior, and serves as a benchmark for validating quantum algorithms and simulation techniques. We show that with GRL we can find very compact PQCs that improve the error in estimating the ground state of TFIM by up to $10^7$ fold and make it suitable for implementation on real hardware compared to a pure reinforcement learning approach. Moreover, GRL scales better with increasing difficulty and to larger systems. The generality of the algorithm shows the potential for applications to other settings, including optimization tailored to specific real-world quantum platforms.

Conventional methods for Bayesian optimization (BO) primarily involve one-step optimal decisions (e.g., maximizing expected improvement of the next step). To avoid myopic behavior, multi-step lookahead BO algorithms such as rollout strategies consider the sequential decision-making nature of BO, i.e., as a stochastic dynamic programming (SDP) problem, demonstrating promising results in recent years. However, owing to the curse of dimensionality, most of these methods make significant approximations or suffer scalability issues, e.g., being limited to two-step lookahead. This paper presents a novel reinforcement learning (RL)-based framework for multi-step lookahead BO in high-dimensional black-box optimization problems. The proposed method enhances the scalability and decision-making quality of multi-step lookahead BO by efficiently solving the SDP of the BO process in a near-optimal manner using RL. We first introduce an Attention-DeepSets encoder to represent the state of knowledge to the RL agent and employ off-policy learning to accelerate its initial training. We then propose a multi-task, fine-tuning procedure based on end-to-end (encoder-RL) on-policy learning. We evaluate the proposed method, EARL-BO (Encoder Augmented RL for Bayesian Optimization), on both synthetic benchmark functions and real-world hyperparameter optimization problems, demonstrating significantly improved performance compared to existing multi-step lookahead and high-dimensional BO methods.