The rapid expansion of global cloud wide-area networks (WANs) has posed a challenge for commercial optimization engines to efficiently solve network traffic engineering (TE) problems at scale. Existing acceleration strategies decompose TE optimization into concurrent subproblems but realize limited parallelism due to an inherent tradeoff between run time and allocation performance. We present Teal, a learning-based TE algorithm that leverages the parallel processing power of GPUs to accelerate TE control. First, Teal designs a flow-centric graph neural network (GNN) to capture WAN connectivity and network flows, learning flow features as inputs to downstream allocation. Second, to reduce the problem scale and make learning tractable, Teal employs a multi-agent reinforcement learning (RL) algorithm to independently allocate each traffic demand while optimizing a central TE objective. Finally, Teal fine-tunes allocations with ADMM (Alternating Direction Method of Multipliers), a highly parallelizable optimization algorithm for reducing constraint violations such as overutilized links. We evaluate Teal using traffic matrices from Microsoft's WAN. On a large WAN topology with >1,700 nodes, Teal generates near-optimal flow allocations while running several orders of magnitude faster than the production optimization engine. Compared with other TE acceleration schemes, Teal satisfies 6--32% more traffic demand and yields 197--625x speedups.

A central issue lying at the heart of online reinforcement learning (RL) is data efficiency. While a number of recent works achieved asymptotically minimal regret in online RL, the optimality of these results is only guaranteed in a ``large-sample'' regime, imposing enormous burn-in cost in order for their algorithms to operate optimally. How to achieve minimax-optimal regret without incurring any burn-in cost has been an open problem in RL theory. We settle this problem for the context of finite-horizon inhomogeneous Markov decision processes. Specifically, we prove that a modified version of Monotonic Value Propagation (MVP), a model-based algorithm proposed by \cite{zhang2020reinforcement}, achieves a regret on the order of (modulo log factors) \begin{equation*} \min\big\{ \sqrt{SAH^3K}, \,HK \big\}, \end{equation*} where $S$ is the number of states, $A$ is the number of actions, $H$ is the planning horizon, and $K$ is the total number of episodes. This regret matches the minimax lower bound for the entire range of sample size $K\geq 1$, essentially eliminating any burn-in requirement. It also translates to a PAC sample complexity (i.e., the number of episodes needed to yield $\varepsilon$-accuracy) of $\frac{SAH^3}{\varepsilon^2}$ up to log factor, which is minimax-optimal for the full $\varepsilon$-range. Further, we extend our theory to unveil the influences of problem-dependent quantities like the optimal value/cost and certain variances. The key technical innovation lies in the development of a new regret decomposition strategy and a novel analysis paradigm to decouple complicated statistical dependency -- a long-standing challenge facing the analysis of online RL in the sample-hungry regime.

Goal-based wealth management (GBWM), also known as goal-based investing [1], is a relatively new class of approaches to wealth management that focus on attaining specific financial objectives (goals). As opposed to more traditional approaches to wealth management, in which the notion of expected profit and loss (PnL) plays a central role, GBWM revolves around maximizing the probability of goal attainment. Common investment goals include saving for college tuition, retirement, or purchasing a home. Recent years have seen an uptick in the popularity of GBWM [2], particularly through the use of target date funds (TDFs). TDFs, also known as life-cycle funds [3] or target-retirement funds, are mutual funds or exchange-traded funds that provide investors with an asset allocation aimed at fulfilling a target (goal) by a specified target date (e.g. a retirement date).

Ultra-reliable low latency communications (URLLC) service is envisioned to enable use cases with strict reliability and latency requirements in 5G. One approach for enabling URLLC services is to leverage Reinforcement Learning (RL) to efficiently allocate wireless resources. However, with conventional RL methods, the decision variables (though being deployed at various network layers) are typically optimized in the same control loop, leading to significant practical limitations on the control loop's delay as well as excessive signaling and energy consumption. In this paper, we propose a multi-agent Hierarchical RL (HRL) framework that enables the implementation of multi-level policies with different control loop timescales. Agents with faster control loops are deployed closer to the base station, while the ones with slower control loops are at the edge or closer to the core network providing high-level guidelines for low-level actions. On a use case from the prior art, with our HRL framework, we optimized the maximum number of retransmissions and transmission power of industrial devices. Our extensive simulation results on the factory automation scenario show that the HRL framework achieves better performance as the baseline single-agent RL method, with significantly less overhead of signal transmissions and delay compared to the one-agent RL methods.

In reinforcement learning (RL), rewards of states are typically considered additive, and following the Markov assumption, they are $\textit{independent}$ of states visited previously. In many important applications, such as coverage control, experiment design and informative path planning, rewards naturally have diminishing returns, i.e., their value decreases in light of similar states visited previously. To tackle this, we propose $\textit{submodular RL}$ (SubRL), a paradigm which seeks to optimize more general, non-additive (and history-dependent) rewards modelled via submodular set functions which capture diminishing returns. Unfortunately, in general, even in tabular settings, we show that the resulting optimization problem is hard to approximate. On the other hand, motivated by the success of greedy algorithms in classical submodular optimization, we propose SubPO, a simple policy gradient-based algorithm for SubRL that handles non-additive rewards by greedily maximizing marginal gains. Indeed, under some assumptions on the underlying Markov Decision Process (MDP), SubPO recovers optimal constant factor approximations of submodular bandits. Moreover, we derive a natural policy gradient approach for locally optimizing SubRL instances even in large state- and action- spaces. We showcase the versatility of our approach by applying SubPO to several applications, such as biodiversity monitoring, Bayesian experiment design, informative path planning, and coverage maximization. Our results demonstrate sample efficiency, as well as scalability to high-dimensional state-action spaces.

As medical ultrasound is becoming a prevailing examination approach nowadays, robotic ultrasound systems can facilitate the scanning process and prevent professional sonographers from repetitive and tedious work. Despite the recent progress, it is still a challenge to enable robots to autonomously accomplish the ultrasound examination, which is largely due to the lack of a proper task representation method, and also an adaptation approach to generalize learned skills across different patients. To solve these problems, we propose the latent task representation and the robotic skills adaptation for autonomous ultrasound in this paper. During the offline stage, the multimodal ultrasound skills are merged and encapsulated into a low-dimensional probability model through a fully self-supervised framework, which takes clinically demonstrated ultrasound images, probe orientations, and contact forces into account. During the online stage, the probability model will select and evaluate the optimal prediction. For unstable singularities, the adaptive optimizer fine-tunes them to near and stable predictions in high-confidence regions. Experimental results show that the proposed approach can generate complex ultrasound strategies for diverse populations and achieve significantly better quantitative results than our previous method.

Recent advances in Reinforcement Learning (RL) have led to many exciting applications. These advancements have been driven by improvements in both algorithms and engineering, which have resulted in faster training of RL agents. We present marl-jax, a multi-agent reinforcement learning software package for training and evaluating social generalization of the agents. The package is designed for training a population of agents in multi-agent environments and evaluating their ability to generalize to diverse background agents. It is built on top of DeepMind's JAX ecosystem~\cite{deepmind2020jax} and leverages the RL ecosystem developed by DeepMind. Our framework marl-jax is capable of working in cooperative and competitive, simultaneous-acting environments with multiple agents. The package offers an intuitive and user-friendly command-line interface for training a population and evaluating its generalization capabilities. In conclusion, marl-jax provides a valuable resource for researchers interested in exploring social generalization in the context of MARL. The open-source code for marl-jax is available at: \href{https://github.com/kinalmehta/marl-jax}{https://github.com/kinalmehta/marl-jax}

Short-term load forecasting (STLF) plays a significant role in the operation of electricity trading markets. Considering the growing concern of data privacy, federated learning (FL) is increasingly adopted to train STLF models for utility companies (UCs) in recent research. Inspiringly, in wholesale markets, as it is not realistic for power plants (PPs) to access UCs' data directly, FL is definitely a feasible solution of obtaining an accurate STLF model for PPs. However, due to FL's distributed nature and intense competition among UCs, defects increasingly occur and lead to poor performance of the STLF model, indicating that simply adopting FL is not enough. In this paper, we propose a DRL-assisted FL approach, DEfect-AwaRe federated soft actor-critic (DearFSAC), to robustly train an accurate STLF model for PPs to forecast precise short-term utility electricity demand. Firstly. we design a STLF model based on long short-term memory (LSTM) using just historical load data and time data. Furthermore, considering the uncertainty of defects occurrence, a deep reinforcement learning (DRL) algorithm is adopted to assist FL by alleviating model degradation caused by defects. In addition, for faster convergence of FL training, an auto-encoder is designed for both dimension reduction and quality evaluation of uploaded models. In the simulations, we validate our approach on real data of Helsinki's UCs in 2019. The results show that DearFSAC outperforms all the other approaches no matter if defects occur or not.

As Reinforcement Learning (RL) agents are increasingly employed in diverse decision-making problems using reward preferences, it becomes important to ensure that policies learned by these frameworks in mapping observations to a probability distribution of the possible actions are explainable. However, there is little to no work in the systematic understanding of these complex policies in a contrastive manner, i.e., what minimal changes to the policy would improve/worsen its performance to a desired level. In this work, we present COUNTERPOL, the first framework to analyze RL policies using counterfactual explanations in the form of minimal changes to the policy that lead to the desired outcome. We do so by incorporating counterfactuals in supervised learning in RL with the target outcome regulated using desired return. We establish a theoretical connection between Counterpol and widely used trust region-based policy optimization methods in RL. Extensive empirical analysis shows the efficacy of COUNTERPOL in generating explanations for (un)learning skills while keeping close to the original policy. Our results on five different RL environments with diverse state and action spaces demonstrate the utility of counterfactual explanations, paving the way for new frontiers in designing and developing counterfactual policies.

While many real-world problems that might benefit from reinforcement learning, these problems rarely fit into the MDP mold: interacting with the environment is often expensive and specifying reward functions is challenging. Motivated by these challenges, prior work has developed data-driven approaches that learn entirely from samples from the transition dynamics and examples of high-return states. These methods typically learn a reward function from high-return states, use that reward function to label the transitions, and then apply an offline RL algorithm to these transitions. While these methods can achieve good results on many tasks, they can be complex, often requiring regularization and temporal difference updates. In this paper, we propose a method for offline, example-based control that learns an implicit model of multi-step transitions, rather than a reward function. We show that this implicit model can represent the Q-values for the example-based control problem. Across a range of state-based and image-based offline control tasks, our method outperforms baselines that use learned reward functions; additional experiments demonstrate improved robustness and scaling with dataset size.