Deep reinforcement learning agents achieve state-of-the-art performance in a wide range of simulated control tasks. However, successful applications to real-world problems remain limited. One reason for this dichotomy is because the learnt policies are not robust to observation noise or adversarial attacks. In this paper, we investigate the robustness of deep RL policies to a single small state perturbation in deterministic continuous control tasks. We demonstrate that RL policies can be deterministically chaotic, as small perturbations to the system state have a large impact on subsequent state and reward trajectories. This unstable non-linear behaviour has two consequences: first, inaccuracies in sensor readings, or adversarial attacks, can cause significant performance degradation; second, even policies that show robust performance in terms of rewards may have unpredictable behaviour in practice. These two facets of chaos in RL policies drastically restrict the application of deep RL to real-world problems. To address this issue, we propose an improvement on the successful Dreamer V3 architecture, implementing Maximal Lyapunov Exponent regularisation. This new approach reduces the chaotic state dynamics, rendering the learnt policies more resilient to sensor noise or adversarial attacks and thereby improving the suitability of deep reinforcement learning for real-world applications.

Vision-language models (VLMs) have tremendous potential for grounding language, and thus enabling language-conditioned agents (LCAs) to perform diverse tasks specified with text. This has motivated the study of LCAs based on reinforcement learning (RL) with rewards given by rendering images of an environment and evaluating those images with VLMs. If single-task RL is employed, such approaches are limited by the cost and time required to train a policy for each new task. Multi-task RL (MTRL) is a natural alternative, but requires a carefully designed corpus of training tasks and does not always generalize reliably to new tasks. Therefore, this paper introduces a novel decomposition of the problem of building an LCA: first find an environment configuration that has a high VLM score for text describing a task; then use a (pretrained) goal-conditioned policy to reach that configuration. We also explore several enhancements to the speed and quality of VLM-based LCAs, notably, the use of distilled models, and the evaluation of configurations from multiple viewpoints to resolve the ambiguities inherent in a single 2D view. We demonstrate our approach on the Humanoid environment, showing that it results in LCAs that outperform MTRL baselines in zero-shot generalization, without requiring any textual task descriptions or other forms of environment-specific annotation during training. Videos and an interactive demo can be found at https://europe.naverlabs.com/text2control

Algorithmic agents are used in a variety of competitive decision settings, notably in making pricing decisions in contexts that range from online retail to residential home rentals. Business managers, algorithm designers, legal scholars, and regulators alike are all starting to consider the ramifications of "algorithmic collusion." We study the emergent behavior of multi-armed bandit machine learning algorithms used in situations where agents are competing, but they have no information about the strategic interaction they are engaged in. Using a general-form repeated Prisoner's Dilemma game, agents engage in online learning with no prior model of game structure and no knowledge of competitors' states or actions (e.g., no observation of competing prices). We show that these context-free bandits, with no knowledge of opponents' choices or outcomes, still will consistently learn collusive behavior - what we call "naive collusion." We primarily study this system through an analytical model and examine perturbations to the model through simulations. Our findings have several notable implications for regulators. First, calls to limit algorithms from conditioning on competitors' prices are insufficient to prevent algorithmic collusion. This is a direct result of collusion arising even in the naive setting. Second, symmetry in algorithms can increase collusion potential. This highlights a new, simple mechanism for "hub-and-spoke" algorithmic collusion. A central distributor need not imbue its algorithm with supra-competitive tendencies for apparent collusion to arise; it can simply arise by using certain (common) machine learning algorithms. Finally, we highlight that collusive outcomes depend starkly on the specific algorithm being used, and we highlight market and algorithmic conditions under which it will be unknown a priori whether collusion occurs.

Reinforcement learning (RL) controllers are flexible and performant but rarely guarantee safety. Safety filters impart hard safety guarantees to RL controllers while maintaining flexibility. However, safety filters can cause undesired behaviours due to the separation between the controller and the safety filter, often degrading performance and robustness. In this paper, we analyze several modifications to incorporating the safety filter in training RL controllers rather than solely applying it during evaluation. The modifications allow the RL controller to learn to account for the safety filter, improving performance. This paper presents a comprehensive analysis of training RL with safety filters, featuring simulated and real-world experiments with a Crazyflie 2.0 drone. We examine how various training modifications and hyperparameters impact performance, sample efficiency, safety, and chattering. Our findings serve as a guide for practitioners and researchers focused on safety filters and safe RL.

This study explores the potential for artificial agents to develop core consciousness, as proposed by Antonio Damasio's theory of consciousness. According to Damasio, the emergence of core consciousness relies on the integration of a self model, informed by representations of emotions and feelings, and a world model. We hypothesize that an artificial agent, trained via reinforcement learning (RL) in a virtual environment, can develop preliminary forms of these models as a byproduct of its primary task. The agent's main objective is to learn to play a video game and explore the environment. To evaluate the emergence of world and self models, we employ probes-feedforward classifiers that use the activations of the trained agent's neural networks to predict the spatial positions of the agent itself. Our results demonstrate that the agent can form rudimentary world and self models, suggesting a pathway toward developing machine consciousness. This research provides foundational insights into the capabilities of artificial agents in mirroring aspects of human consciousness, with implications for future advancements in artificial intelligence.

Ensuring the safety of autonomous vehicles (AVs) requires identifying rare but critical failure cases that on-road testing alone cannot discover. High-fidelity simulations provide a scalable alternative, but automatically generating realistic and diverse traffic scenarios that can effectively stress test AV motion planners remains a key challenge. This paper introduces CRASH - Challenging Reinforcement-learning based Adversarial scenarios for Safety Hardening - an adversarial deep reinforcement learning framework to address this issue. First CRASH can control adversarial Non Player Character (NPC) agents in an AV simulator to automatically induce collisions with the Ego vehicle, falsifying its motion planner. We also propose a novel approach, that we term safety hardening, which iteratively refines the motion planner by simulating improvement scenarios against adversarial agents, leveraging the failure cases to strengthen the AV stack. CRASH is evaluated on a simplified two-lane highway scenario, demonstrating its ability to falsify both rule-based and learning-based planners with collision rates exceeding 90%. Additionally, safety hardening reduces the Ego vehicle's collision rate by 26%. While preliminary, these results highlight RL-based safety hardening as a promising approach for scenario-driven simulation testing for autonomous vehicles.

An agent trained within a closed system can master any desired capability, as long as the following three conditions hold: (a) it receives sufficiently informative and aligned feedback, (b) its coverage of experience/data is broad enough, and (c) it has sufficient capacity and resource. In this position paper, we justify these conditions, and consider what limitations arise from (a) and (b) in closed systems, when assuming that (c) is not a bottleneck. Considering the special case of agents with matching input and output spaces (namely, language), we argue that such pure recursive self-improvement, dubbed "Socratic learning", can boost performance vastly beyond what is present in its initial data or knowledge, and is only limited by time, as well as gradual misalignment concerns. Furthermore, we propose a constructive framework to implement it, based on the notion of language games.

We present PROGRESSOR, a novel framework that learns a task-agnostic reward function from videos, enabling policy training through goal-conditioned reinforcement learning (RL) without manual supervision. Underlying this reward is an estimate of the distribution over task progress as a function of the current, initial, and goal observations that is learned in a self-supervised fashion. Crucially, PROGRESSOR refines rewards adversarially during online RL training by pushing back predictions for out-of-distribution observations, to mitigate distribution shift inherent in non-expert observations. Utilizing this progress prediction as a dense reward together with an adversarial push-back, we show that PROGRESSOR enables robots to learn complex behaviors without any external supervision. Pretrained on large-scale egocentric human video from EPIC-KITCHENS, PROGRESSOR requires no fine-tuning on in-domain task-specific data for generalization to real-robot offline RL under noisy demonstrations, outperforming contemporary methods that provide dense visual reward for robotic learning. Our findings highlight the potential of PROGRESSOR for scalable robotic applications where direct action labels and task-specific rewards are not readily available.

In the realm of emerging real-time networked applications like cyber-physical systems (CPS), the Age of Information (AoI) has merged as a pivotal metric for evaluating the timeliness. To meet the high computational demands, such as those in intelligent manufacturing within CPS, mobile edge computing (MEC) presents a promising solution for optimizing computing and reducing AoI. In this work, we study the timeliness of computational-intensive updates and explores jointly optimize the task updating and offloading policies to minimize AoI. Specifically, we consider edge load dynamics and formulate a task scheduling problem to minimize the expected time-average AoI. The fractional objective introduced by AoI and the semi-Markov game nature of the problem render this challenge particularly difficult, with existing approaches not directly applicable. To this end, we present a comprehensive framework to fractional reinforcement learning (RL). We first introduce a fractional single-agent RL framework and prove its linear convergence. We then extend this to a fractional multi-agent RL framework with a convergence analysis. To tackle the challenge of asynchronous control in semi-Markov game, we further design an asynchronous model-free fractional multi-agent RL algorithm, where each device makes scheduling decisions with the hybrid action space without knowing the system dynamics and decisions of other devices. Experimental results show that our proposed algorithms reduce the average AoI by up to 52.6% compared with the best baseline algorithm in our experiments.

Due to the inherent lack of transparency in deep neural networks, it is challenging for deep reinforcement learning (DRL) agents to gain trust and acceptance from users, especially in safety-critical applications such as medical diagnosis and military operations. Existing methods for explaining an agent's decision either require to retrain the agent using models that support explanation generation or rely on perturbation-based techniques to reveal the significance of different input features in the decision making process. However, retraining the agent may compromise its integrity and performance, while perturbation-based methods have limited performance and lack knowledge accumulation or learning capabilities. Moreover, since each perturbation is performed independently, the joint state of the perturbed inputs may not be physically meaningful. To address these challenges, we introduce $\textbf{VisionMask}$, a standalone explanation model trained end-to-end to identify the most critical regions in the agent's visual input that can explain its actions. VisionMask is trained in a self-supervised manner without relying on human-generated labels. Importantly, its training does not alter the agent model, hence preserving the agent's performance and integrity. We evaluate VisionMask on Super Mario Bros (SMB) and three Atari games. Compared to existing methods, VisionMask achieves a 14.9% higher insertion accuracy and a 30.08% higher F1-Score in reproducing original actions from the selected visual explanations. We also present examples illustrating how VisionMask can be used for counterfactual analysis.