There has been growing progress on theoretical analyses for provably efficient learning in MDPs with linear function approximation, but much of the existing work has made strong assumptions to enable exploration by conventional exploration frameworks. Typically these assumptions are stronger than what is needed to find good solutions in the batch setting. In this work, we show how under a more standard notion of low inherent Bellman error, typically employed in least-square value iteration-style algorithms, we can provide strong PAC guarantees on learning a near optimal value function provided that the linear space is sufficiently "explorable". We present a computationally tractable algorithm for the reward-free setting and show how it can be used to learn a near optimal policy for any (linear) reward function, which is revealed only once learning has completed. If this reward function is also estimated from the samples gathered during pure exploration, our results also provide same-order PAC guarantees on the performance of the resulting policy for this setting.

Self-imitation learning motivated by lower-bound Q-learning is a novel and effective approach for off-policy learning. In this work, we propose a n-step lower bound which generalizes the original return-based lower-bound Q-learning, and introduce a new family of self-imitation learning algorithms. To provide a formal motivation for the potential performance gains provided by self-imitation learning, we show that n-step lower bound Q-learning achieves a trade-off between fixed point bias and contraction rate, drawing close connections to the popular uncorrected n-step Q-learning. We finally show that n-step lower bound Q-learning is a more robust alternative to return-based self-imitation learning and uncorrected n-step, over a wide range of continuous control benchmark tasks.

We present a novel framework for design space search on analog circuit sizing using deep reinforcement learning (DRL). Nowadays, analog circuit design is a manual routine that requires heavy design efforts due to the absence of automation tools, motivating the urge to develop one. Prior approaches cast this process as an optimization problem. They use global search strategies based on DRL with complex network architectures. Nonetheless, the models are hard to converge and neglected various working conditions of PVT (process, voltage, temperature).In this work, we reduce the problem to a constraint satisfaction problem, where a local strategy is adopted. Thus, a simple feed-forward network with few layers can be used to implement a model-based reinforcement learning agent. To evaluate the value of the our framework in production, we cooperate with R&Ds in an IC design company. On circuits with TSMC advanced 5 and 6nm process, our agents can deliver PPA (performance, power, area) beyond human level. Furthermore, the product will be taped out in the near future.

While REPS is elegantly derived from a principled We propose a new reinforcement learning algorithm linear-programing (LP) formulation of optimal control derived from a regularized linearprogramming in MDPs, it has the serious shortcoming that its faithful formulation of optimal control implementation requires access to the true MDP in MDPs. The method is closely related to for both the policy evaluation and improvement steps, the classic Relative Entropy Policy Search even at deployment time. The usual way to address (REPS) algorithm of Peters et al. (2010), with this limitation is to use an empirical approximation to the key difference that our method introduces the policy evaluation step and to project the policy a Q-function that enables efficient exact from the improvement step into a parametric space model-free implementation. The main (Deisenroth et al., 2013), losing all the theoretical feature of our algorithm (called Q-REPS) is guarantees of REPS in the process.

Machine learning models have widely been used in fraud detection systems. Most of the research and development efforts have been concentrated on improving the performance of the fraud scoring models. Yet, the downstream fraud alert systems still have limited to no model adoption and rely on manual steps. Alert systems are pervasively used across all payment channels in retail banking and play an important role in the overall fraud detection process. Current fraud detection systems end up with large numbers of dropped alerts due to their inability to account for the alert processing capacity. Ideally, alert threshold selection enables the system to maximize the fraud detection while balancing the upstream fraud scores and the available bandwidth of the alert processing teams. However, in practice, fixed thresholds that are used for their simplicity do not have this ability. In this paper, we propose an enhanced threshold selection policy for fraud alert systems. The proposed approach formulates the threshold selection as a sequential decision making problem and uses Deep Q-Network based reinforcement learning. Experimental results show that this adaptive approach outperforms the current static solutions by reducing the fraud losses as well as improving the operational efficiency of the alert system.

Deep reinforcement learning (RL) agents trained in a limited set of environments tend to suffer overfitting and fail to generalize to unseen testing environments. To improve their generalizability, data augmentation approaches (e.g. cutout and random convolution) are previously explored to increase the data diversity. However, we find these approaches only locally perturb the observations regardless of the training environments, showing limited effectiveness on enhancing the data diversity and the generalization performance. In this work, we introduce a simple approach, named mixreg, which trains agents on a mixture of observations from different training environments and imposes linearity constraints on the observation interpolations and the supervision (e.g. associated reward) interpolations. Mixreg increases the data diversity more effectively and helps learn smoother policies. We verify its effectiveness on improving generalization by conducting extensive experiments on the large-scale Procgen benchmark. Results show mixreg outperforms the well-established baselines on unseen testing environments by a large margin. Mixreg is simple, effective and general. It can be applied to both policy-based and value-based RL algorithms. Code is available at https://github.com/kaixin96/mixreg .

Deep Reinforcement Learning Course is a free course (articles and videos) about Deep Reinforcement Learning, where we'll learn the main algorithms, and how to implement them in Tensorflow and PyTorch.

When autonomous agents interact in the same environment, they must often cooperate to achieve their goals. One way for agents to cooperate effectively is to form a team, make a binding agreement on a joint plan, and execute it. However, when agents are self-interested, the gains from team formation must be allocated appropriately to incentivize agreement. Various approaches for multi-agent negotiation have been proposed, but typically only work for particular negotiation protocols. More general methods usually require human input or domain-specific data, and so do not scale. To address this, we propose a framework for training agents to negotiate and form teams using deep reinforcement learning. Importantly, our method makes no assumptions about the specific negotiation protocol, and is instead completely experience driven. We evaluate our approach on both non-spatial and spatially extended team-formation negotiation environments, demonstrating that our agents beat hand-crafted bots and reach negotiation outcomes consistent with fair solutions predicted by cooperative game theory. Additionally, we investigate how the physical location of agents influences negotiation outcomes.

Despite great interest in recent research, in particular in China [1, 2] micromanagement of sensors by centralized command and control drives possible inefficiencies and risk into operations. Tactical decision making and execution by headquarters usually fail to achieve the speed necessary to meet rapid changes. Collaborative radars with C2 must provide decision superiority despite the attempts of an adversary to disrupt OODA cycles at all level of operations. Artificial intelligence can make a contribution for the purposes of coordinated conduct of the action, by improving the response time to threats and optimizing the allocation and the distribution of tasks within elementary smart radars. In order to address this problem, Thales and the private research lab NukkAI have been collaborating to introduce novel approaches for netted radars. Thales provided the simulation modeling the multi-radar target allocation problem and NukkAI proposed two novel reward-based learning approaches for the problem. In this paper, we present these two approaches: Evolutionary Single-Target Ordering (ESTO), which is based on evolution strategies and an RL approach based on Actor-Critic methods. To make the RL method tractable in practice, we introduce a simplification of the problem that we prove to be equivalent to solving the initial formulation. We evaluate our solutions on diverse scenarios of the aforementioned simulation.

Many real-world physical control systems are required to satisfy constraints upon deployment. Furthermore, real-world systems are often subject to effects such as non-stationarity, wear-and-tear, uncalibrated sensors and so on. Such effects effectively perturb the system dynamics and can cause a policy trained successfully in one domain to perform poorly when deployed to a perturbed version of the same domain. This can affect a policy's ability to maximize future rewards as well as the extent to which it satisfies constraints. We refer to this as constrained model misspecification. We present an algorithm with theoretical guarantees that mitigates this form of misspecification, and showcase its performance in multiple Mujoco tasks from the Real World Reinforcement Learning (RWRL) suite.