Natural intelligence processes experience as a continuous stream, sensing, acting, and learning moment-by-moment in real time. Streaming learning, the modus operandi of classic reinforcement learning (RL) algorithms like Q-learning and TD, mimics natural learning by using the most recent sample without storing it. This approach is also ideal for resource-constrained, communication-limited, and privacy-sensitive applications. However, in deep RL, learners almost always use batch updates and replay buffers, making them computationally expensive and incompatible with streaming learning. Although the prevalence of batch deep RL is often attributed to its sample efficiency, a more critical reason for the absence of streaming deep RL is its frequent instability and failure to learn, which we refer to as stream barrier. This paper introduces the stream-x algorithms, the first class of deep RL algorithms to overcome stream barrier for both prediction and control and match sample efficiency of batch RL. Through experiments in Mujoco Gym, DM Control Suite, and Atari Games, we demonstrate stream barrier in existing algorithms and successful stable learning with our stream-x algorithms: stream Q, stream AC, and stream TD, achieving the best model-free performance in DM Control Dog environments. A set of common techniques underlies the stream-x algorithms, enabling their success with a single set of hyperparameters and allowing for easy extension to other algorithms, thereby reviving streaming RL.

Modern deep policy gradient methods achieve effective performance on simulated robotic tasks, but they all require large replay buffers or expensive batch updates, or both, making them incompatible for real systems with resource-limited computers. We show that these methods fail catastrophically when limited to small replay buffers or during incremental learning, where updates only use the most recent sample without batch updates or a replay buffer. We propose a novel incremental deep policy gradient method -- Action Value Gradient (AVG) and a set of normalization and scaling techniques to address the challenges of instability in incremental learning. On robotic simulation benchmarks, we show that AVG is the only incremental method that learns effectively, often achieving final performance comparable to batch policy gradient methods. This advancement enabled us to show for the first time effective deep reinforcement learning with real robots using only incremental updates, employing a robotic manipulator and a mobile robot.

Many failures in deep continual and reinforcement learning are associated with increasing magnitudes of the weights, making them hard to change and potentially causing overfitting. While many methods address these learning failures, they often change the optimizer or the architecture, a complexity that hinders widespread adoption in various systems. In this paper, we focus on learning failures that are associated with increasing weight norm and we propose a simple technique that can be easily added on top of existing learning systems: clipping neural network weights to limit them to a specific range. We study the effectiveness of weight clipping in a series of supervised and reinforcement learning experiments. Our empirical results highlight the benefits of weight clipping for generalization, addressing loss of plasticity and policy collapse, and facilitating learning with a large replay ratio.

While many methods address these two issues separately, only a few currently deal with both simultaneously. In this paper, we introduce Utility-based Perturbed Gradient Descent (UPGD) as a novel approach for the continual learning of representations. UPGD combines gradient updates with perturbations, where it applies smaller modifications to more useful units, protecting them from forgetting, and larger modifications to less useful units, rejuvenating their plasticity. We use a challenging streaming learning setup where continual learning problems have hundreds of non-stationarities and unknown task boundaries. We show that many existing methods suffer from at least one of the issues, predominantly manifested by their decreasing accuracy over tasks. On the other hand, UPGD continues to improve performance and surpasses or is competitive with all methods in all problems. Finally, in extended reinforcement learning experiments with PPO, we show that while Adam exhibits a performance drop after initial learning, UPGD avoids it by addressing both continual learning issues. Continual learning remains a significant hurdle for artificial intelligence, despite advancements in natural language processing, games, and computer vision. Catastrophic forgetting (McCloskey & Cohen 1989, Hetherington & Seidenberg 1989) in neural networks is widely recognized as a major challenge of continual learning (De Lange et al. 2021). The phenomenon manifests as the failure of gradient-based methods like SGD or Adam to retain or leverage past knowledge due to forgetting or overwriting previously learned units (Kirkpatrick et al. 2017). This issue also raises a concern for reusing large practical models, where finetuning them for new tasks causes significant forgetting of pretrained models (Chen et al. 2020, He et al. 2021). Methods for mitigating catastrophic forgetting are primarily designed for specific settings. These include settings with independently and identically distributed (i.i.d.) samples, tasks fully contained within a batch or dataset, growing memory requirements, known task boundaries, storing past samples, and offline evaluation. Such setups are often impractical in situations where continual learning is paramount, such as on-device learning. For example, retaining samples may not be possible due to the limitation of computational resources (Hayes et al. 2019, Hayes et al. 2020, Hayes & Kannan 2022, Wang et al. 2023) or concerns over data privacy (Van de Ven et al. 2020). In the challenging and practical setting of streaming learning, catastrophic forgetting is more severe and remains largely unaddressed (Hayes et al. 2019). In streaming learning, samples are presented to the learner as they arise, which is non-i.i.d. in most practical problems.

Modern representation learning methods often struggle to adapt quickly under non-stationarity because they suffer from catastrophic forgetting and decaying plasticity. Such problems prevent learners from fast adaptation since they may forget useful features or have difficulty learning new ones. Hence, these methods are rendered ineffective for continual learning. This paper proposes Utility-based Perturbed Gradient Descent (UPGD), an online learning algorithm well-suited for continual learning agents. UPGD protects useful weights or features from forgetting and perturbs less useful ones based on their utilities. Our empirical results show that UPGD helps reduce forgetting and maintain plasticity, enabling modern representation learning methods to work effectively in continual learning.

Second-order optimization uses curvature information about the objective function, which can help in faster convergence. However, such methods typically require expensive computation of the Hessian matrix, preventing their usage in a scalable way. The absence of efficient ways of computation drove the most widely used methods to focus on first-order approximations that do not capture the curvature information. In this paper, we develop HesScale, a scalable approach to approximating the diagonal of the Hessian matrix, to incorporate second-order information in a computationally efficient manner. We show that HesScale has the same computational complexity as backpropagation. Our results on supervised classification show that HesScale achieves high approximation accuracy, allowing for scalable and efficient second-order optimization. First-order optimization offers a cheap and efficient way of performing local progress in optimization problems by using gradient information. However, their performance suffers from instability or slow progress when used in ill-conditioned landscapes. Such a problem is present because firstorder methods do not capture curvature information which causes two interrelated issues. First, the updates in first-order have incorrect units (Duchi et al. 2011), which creates a scaling issue.

Multi-agent interaction is a fundamental aspect of autonomous driving in the real world. Despite more than a decade of research and development, the problem of how to competently interact with diverse road users in diverse scenarios remains largely unsolved. Learning methods have much to offer towards solving this problem. But they require a realistic multi-agent simulator that generates diverse and competent driving interactions. To meet this need, we develop a dedicated simulation platform called SMARTS (Scalable Multi-Agent RL Training School). SMARTS supports the training, accumulation, and use of diverse behavior models of road users. These are in turn used to create increasingly more realistic and diverse interactions that enable deeper and broader research on multi-agent interaction. In this paper, we describe the design goals of SMARTS, explain its basic architecture and its key features, and illustrate its use through concrete multi-agent experiments on interactive scenarios. We open-source the SMARTS platform and the associated benchmark tasks and evaluation metrics to encourage and empower research on multi-agent learning for autonomous driving. Our code is available at https://github.com/huawei-noah/SMARTS.