Power sector decarbonization plays a vital role in the upcoming energy transition towards a more sustainable future. Decentralized energy resources, such as Electric Vehicles (EV) and solar photovoltaic systems (PV), are continuously integrated in residential power systems, increasing the risk of bottlenecks in power distribution networks. This paper aims to address the challenge of domestic EV charging while prioritizing clean, solar energy consumption. Real Time-of-Use tariffs are treated as a price-based Demand Response (DR) mechanism that can incentivize end-users to optimally shift EV charging load in hours of high solar PV generation with the use of Deep Reinforcement Learning (DRL). Historical measurements from the Pecan Street dataset are analyzed to shape a flexibility potential reward to describe end-user charging preferences. Experimental results show that the proposed DQN EV optimal charging policy is able to reduce electricity bills by an average 11.5\% by achieving an average utilization of solar power 88.4

We empirically investigate how pre-training on data of different modalities, such as language and vision, affects fine-tuning of Transformer-based models to Mujoco offline reinforcement learning tasks. Analysis of the internal representation reveals that the pre-trained Transformers acquire largely different representations before and after pre-training, but acquire less information of data in fine-tuning than the randomly initialized one. A closer look at the parameter changes of the pre-trained Transformers reveals that their parameters do not change that much and that the bad performance of the model pre-trained with image data could partially come from large gradients and gradient clipping. To study what information the Transformer pre-trained with language data utilizes, we fine-tune this model with no context provided, finding that the model learns efficiently even without context information. Subsequent follow-up analysis supports the hypothesis that pre-training with language data is likely to make the Transformer get context-like information and utilize it to solve the downstream task.

Online reinforcement learning is concerned with training an agent on-the-fly via dynamic interaction with the environment. Here, due to the specifics of the application, it is not generally possible to perform long pre-training, as it is commonly done in off-line, model-free approaches, which are akin to dynamic programming. Such applications may be found more frequently in industry, rather than in pure digital fields, such as cloud services, video games, database management, etc., where reinforcement learning has been demonstrating success. Online reinforcement learning, in contrast, is more akin to classical control, which utilizes some model knowledge about the environment. Stability of the closed-loop (agent plus the environment) is a major challenge for such online approaches. In this paper, we tackle this problem by a special fusion of online reinforcement learning with elements of classical control, namely, based on the Lyapunov theory of stability. The idea is to start the agent at once, without pre-training, and learn approximately optimal policy under specially designed constraints, which guarantee stability. The resulting approach was tested in an extensive experimental study with a mobile robot. A nominal parking controller was used as a baseline. It was observed that the suggested agent could always successfully park the robot, while significantly improving the cost. While many approaches may be exploited for mobile robot control, we suggest that the experiments showed the promising potential of online reinforcement learning agents based on Lyapunov-like constraints. The presented methodology may be utilized in safety-critical, industrial applications where stability is necessary.

Motivated by the emerging needs of personalized preventative intervention in many healthcare applications, we consider a multi-stage, dynamic decision-making problem in the online setting with unknown model parameters. To deal with the pervasive issue of small sample size in personalized planning, we develop a novel data-pooling reinforcement learning (RL) algorithm based on a general perturbed value iteration framework. Our algorithm adaptively pools historical data, with three main innovations: (i) the weight of pooling ties directly to the performance of decision (measured by regret) as opposed to estimation accuracy in conventional methods; (ii) no parametric assumptions are needed between historical and current data; and (iii) requiring data-sharing only via aggregate statistics, as opposed to patient-level data. Our data-pooling algorithm framework applies to a variety of popular RL algorithms, and we establish a theoretical performance guarantee showing that our pooling version achieves a regret bound strictly smaller than that of the no-pooling counterpart. We substantiate the theoretical development with empirically better performance of our algorithm via a case study in the context of post-discharge intervention to prevent unplanned readmissions, generating practical insights for healthcare management. In particular, our algorithm alleviates privacy concerns about sharing health data, which (i) opens the door for individual organizations to levering public datasets or published studies to better manage their own patients; and (ii) provides the basis for public policy makers to encourage organizations to share aggregate data to improve population health outcomes for the broader community.

This paper aims at presenting a new application of information geometry to reinforcement learning focusing on dynamic treatment resumes. In a standard framework of reinforcement learning, a Q-function is defined as the conditional expectation of a reward given a state and an action for a single-stage situation. We introduce an equivalence relation, called the policy equivalence, in the space of all the Q-functions. A class of information divergence is defined in the Q-function space for every stage. The main objective is to propose an estimator of the optimal policy function by a method of minimum information divergence based on a dataset of trajectories. In particular, we discuss the $\gamma$-power divergence that is shown to have an advantageous property such that the $\gamma$-power divergence between policy-equivalent Q-functions vanishes. This property essentially works to seek the optimal policy, which is discussed in a framework of a semiparametric model for the Q-function. The specific choices of power index $\gamma$ give interesting relationships of the value function, and the geometric and harmonic means of the Q-function. A numerical experiment demonstrates the performance of the minimum $\gamma$-power divergence method in the context of dynamic treatment regimes.

Building autonomous agents able to grow open-ended repertoires of skills across their lives is a fundamental goal of artificial intelligence (AI). A promising developmental approach recommends the design of intrinsically motivated agents that learn new skills by generating and pursuing their own goals -- autotelic agents. But despite recent progress, existing algorithms still show serious limitations in terms of goal diversity, exploration, generalisation or skill composition. This perspective calls for the immersion of autotelic agents into rich socio-cultural worlds, an immensely important attribute of our environment that shapes human cognition but is mostly omitted in modern AI. Inspired by the seminal work of Vygotsky, we propose Vygotskian autotelic agents -- agents able to internalise their interactions with others and turn them into cognitive tools. We focus on language and show how its structure and informational content may support the development of new cognitive functions in artificial agents as it does in humans. We justify the approach by uncovering several examples of new artificial cognitive functions emerging from interactions between language and embodiment in recent works at the intersection of deep reinforcement learning and natural language processing. Looking forward, we highlight future opportunities and challenges for Vygotskian Autotelic AI research, including the use of language models as cultural models supporting artificial cognitive development.

This paper aims to develop resilient transmission mechanisms to suitably distribute traffic across multiple paths in an arbitrary millimeter-wave (mmWave) network. The main contributions include: (a) the development of proactive transmission mechanisms that build resilience against network disruptions in advance, while achieving a high end-to-end packet rate; (b) the design of a heuristic path selection algorithm that efficiently selects (in polynomial time in the network size) multiple proactively resilient paths with high packet rates; and (c) the development of a hybrid scheduling algorithm that combines the proposed path selection algorithm with a deep reinforcement learning (DRL) based online approach for decentralized adaptation to blocked links and failed paths. To achieve resilience to link failures, a stateof-the-art Soft Actor-Critic DRL algorithm, which adapts the information flow through the network, is investigated. The proposed scheduling algorithm robustly adapts to link failures over different topologies, channel and blockage realizations while offering a superior performance to alternative algorithms. M. G. Dogan and C. Fragouli are with the Electrical and Computer Engineering Department at the University of California, Los Angeles, CA 90095 USA (e-mail: {minedogan96, christina.fragouli}@ucla.edu). The research carried out at UCLA was supported in part by the Army Research Laboratory under Co-Operative Agreement W911NF-17-2-0196 and by the U.S. National Science Foundation (NSF) awards 442521-FC-22071 and 442521-FC-21454. M. Cardone is with the Electrical and Computer Engineering Department of the University of Minnesota, MN 55404 USA (e-mail: cardo089@umn.edu). The work of M. Cardone was supported in part by the NSF under Grants CCF-2045237 and CNS-2146838. Part of this work was presented at the 2021 IEEE Military Communications Conference [1] and at the 2022 IEEE International Symposium on Information Theory [2]. Millimeter Wave (mmWave) (and beyond) is an enabling technology that is playing an increasingly important role in our wireless infrastructure by expanding the available spectrum and enabling multi-gigabit services [3]-[5]. A number of use cases are currently built around multihop mmWave networks, such as Facebook's Terragraph network [6] that uses flexible mmWave backbones to connect clusters of base stations. Other example scenarios include private networks, such as in shopping centers, airports and enterprises; mmWave mesh networks that use mmWave links as backhaul in dense urban scenarios; military applications employing mobile hot spots; and mmWave based vehicle-to-everything (V2X) services, such as cooperative perception [7]-[9].

In the last few years, researchers have applied machine learning strategies in the context of vehicular platoons to increase the safety and efficiency of cooperative transportation. Reinforcement Learning methods have been employed in the longitudinal spacing control of Cooperative Adaptive Cruise Control systems, but to date, none of those studies have addressed problems of disturbance rejection in such scenarios. Characteristics such as uncertain parameters in the model and external interferences may prevent agents from reaching null-spacing errors when traveling at cruising speed. On the other hand, complex communication topologies lead to specific training processes that can not be generalized to other contexts, demanding re-training every time the configuration changes. Therefore, in this paper, we propose an approach to generalize the training process of a vehicular platoon, such that the acceleration command of each agent becomes independent of the network topology. Also, we have modeled the acceleration input as a term with integral action, such that the Artificial Neural Network is capable of learning corrective actions when the states are disturbed by unknown effects. We illustrate the effectiveness of our proposal with experiments using different network topologies, uncertain parameters, and external forces. Comparative analyses, in terms of the steady-state error and overshoot response, were conducted against the state-of-the-art literature. The findings offer new insights concerning generalization and robustness of using Reinforcement Learning in the control of autonomous platoons.

Amazon, Nimble, Plus One, Waymo, and Zoox use remote human supervision of robot fleets in applications ranging from self-driving taxis to automated warehouse fulfillment [1, 2, 3, 4, 5]. These robots intermittently cede control during task execution to remote human supervisors for corrective interventions. The interventions take place either during learning, when they are used to improve the robot policy, or during execution, when the policy is no longer updated but robots can still request human assistance when needed to improve reliability. In the continual learning setting, these occur simultaneously: the robot policy has been deployed but continues to be updated indefinitely with additional intervention data. Furthermore, any individual robot can share its intervention data with the rest of the fleet. As opposed to robot swarms that must coordinate with each other to achieve a common objective, a robot fleet is a set of independent robots simultaneously executing the same control policy in parallel environments. We refer to the setting of a robot fleet learning via interactive requests for human supervision (see Figure 1) as Interactive Fleet Learning (IFL). Of central importance in IFL is the supervisor allocation problem: how should limited human supervision be allocated to robots in a manner that maximizes the throughput of the fleet?

Large-scale language models often learn behaviors that are misaligned with user expectations. Generated text may contain offensive or toxic language, contain significant repetition, or be of a different sentiment than desired by the user. We consider the task of unlearning these misalignments by fine-tuning the language model on signals of what not to do. We introduce Quantized Reward Konditioning (Quark), an algorithm for optimizing a reward function that quantifies an (un)wanted property, while not straying too far from the original model. Quark alternates between (i) collecting samples with the current language model, (ii) sorting them into quantiles based on reward, with each quantile identified by a reward token prepended to the language model's input, and (iii) using a standard language modeling loss on samples from each quantile conditioned on its reward token, while remaining nearby the original language model via a KL-divergence penalty. By conditioning on a high-reward token at generation time, the model generates text that exhibits less of the unwanted property. For unlearning toxicity, negative sentiment, and repetition, our experiments show that Quark outperforms both strong baselines and state-of-the-art reinforcement learning methods like PPO (Schulman et al. 2017), while relying only on standard language modeling primitives.