A key barrier to using reinforcement learning (RL) in many real-world applications is the requirement of a large number of system interactions to learn a good control policy. Off-policy and Offline RL methods have been proposed to reduce the number of interactions with the physical environment by learning control policies from historical data. However, their performances suffer from the lack of exploration and the distributional shifts in trajectories once controllers are updated. Moreover, most RL methods require that all states are directly observed, which is difficult to be attained in many settings. To overcome these challenges, we propose a trajectory generation algorithm, which adaptively generates new trajectories as if the system is being operated and explored under the updated control policies. Motivated by the fundamental lemma for linear systems, assuming sufficient excitation, we generate trajectories from linear combinations of historical trajectories. For linear feedback control, we prove that the algorithm generates trajectories with the exact distribution as if they are sampled from the real system using the updated control policy. In particular, the algorithm extends to systems where the states are not directly observed. Experiments show that the proposed method significantly reduces the number of sampled data needed for RL algorithms.
In task-oriented dialogs (TOD), reinforcement learning (RL) algorithms train a model to directly optimize response for task-related metrics. However, RL needs to perform exploration, which can be time-consuming due to the slow auto-regressive sequence generation process. We investigate an approach to create a more efficient RL-based algorithm to improve TOD performance in an offline setting. First, we use a faster generation procedure that samples from independent next-word distributions after training the language model (LM) with supervised learning. We then introduce a fine-grained reward function to help the model focus on learning key information in a dialog, by measuring the importance and semantic closeness of each generated token. Experiments on the MultiWoZ dataset show our new training algorithm, Keywords Reinforcement Learning with Next-word Sampling (KRLS), achieves state-of-the-art performance on the end-to-end response generation task, with a 15% training time reduction compared to a standard RL algorithm using auto-regressive generation.
Offline reinforcement learning (RL) have received rising interest due to its appealing data efficiency. The present study addresses behavior estimation, a task that lays the foundation of many offline RL algorithms. Behavior estimation aims at estimating the policy with which training data are generated. In particular, this work considers a scenario where the data are collected from multiple sources. In this case, neglecting data heterogeneity, existing approaches for behavior estimation suffers from behavior misspecification. To overcome this drawback, the present study proposes a latent variable model to infer a set of policies from data, which allows an agent to use as behavior policy the policy that best describes a particular trajectory. This model provides with a agent fine-grained characterization for multi-source data and helps it overcome behavior misspecification. This work also proposes a learning algorithm for this model and illustrates its practical usage via extending an existing offline RL algorithm. Lastly, with extensive evaluation this work confirms the existence of behavior misspecification and the efficacy of the proposed model.
Constrained multiagent reinforcement learning (C-MARL) is gaining importance as MARL algorithms find new applications in real-world systems ranging from energy systems to drone swarms. Most C-MARL algorithms use a primal-dual approach to enforce constraints through a penalty function added to the reward. In this paper, we study the structural effects of this penalty term on the MARL problem. First, we show that the standard practice of using the constraint function as the penalty leads to a weak notion of safety. However, by making simple modifications to the penalty term, we can enforce meaningful probabilistic (chance and conditional value at risk) constraints. Second, we quantify the effect of the penalty term on the value function, uncovering an improved value estimation procedure. We use these insights to propose a constrained multiagent advantage actor critic (C-MAA2C) algorithm. Simulations in a simple constrained multiagent environment affirm that our reinterpretation of the primal-dual method in terms of probabilistic constraints is effective, and that our proposed value estimate accelerates convergence to a safe joint policy.
World models power some of the most efficient reinforcement learning algorithms. In this work, we showcase that they can be harnessed for continual learning - a situation when the agent faces changing environments. World models typically employ a replay buffer for training, which can be naturally extended to continual learning. We systematically study how different selective experience replay methods affect performance, forgetting, and transfer. We also provide recommendations regarding various modeling options for using world models. The best set of choices is called Continual-Dreamer, it is task-agnostic and utilizes the world model for continual exploration. Continual-Dreamer is sample efficient and outperforms state-of-the-art task-agnostic continual reinforcement learning methods on Minigrid and Minihack benchmarks.
A central problem in computational biophysics is protein structure prediction, i.e., finding the optimal folding of a given amino acid sequence. This problem has been studied in a classical abstract model, the HP model, where the protein is modeled as a sequence of H (hydrophobic) and P (polar) amino acids on a lattice. The objective is to find conformations maximizing H-H contacts. It is known that even in this reduced setting, the problem is intractable (NP-hard). In this work, we apply deep reinforcement learning (DRL) to the two-dimensional HP model. We can obtain the conformations of best known energies for benchmark HP sequences with lengths from 20 to 50. Our DRL is based on a deep Q-network (DQN). We find that a DQN based on long short-term memory (LSTM) architecture greatly enhances the RL learning ability and significantly improves the search process. DRL can sample the state space efficiently, without the need of manual heuristics. Experimentally we show that it can find multiple distinct best-known solutions per trial. This study demonstrates the effectiveness of deep reinforcement learning in the HP model for protein folding.
Recent advancements in reinforcement learning algorithms have opened doors for researchers to operate and optimize building energy management systems autonomously. However, the lack of an easily configurable building dynamical model and energy management task simulation and evaluation platform has arguably slowed the progress in developing advanced and dedicated reinforcement learning (RL) and control algorithms for building operation tasks. Here we propose "BEAR", a physics-principled Building Environment for Control And Reinforcement Learning. The platform allows researchers to benchmark both model-based and model-free controllers using a broad collection of standard building models in Python without co-simulation using external building simulators. In this paper, we discuss the design of this platform and compare it with other existing building simulation frameworks. We demonstrate the compatibility and performance of BEAR with different controllers, including both model predictive control (MPC) and several state-of-the-art RL methods with two case studies.
Constrained reinforcement learning (RL) is an area of RL whose objective is to find an optimal policy that maximizes expected cumulative return while satisfying a given constraint. Most of the previous constrained RL works consider expected cumulative sum cost as the constraint. However, optimization with this constraint cannot guarantee a target probability of outage event that the cumulative sum cost exceeds a given threshold. This paper proposes a framework, named Quantile Constrained RL (QCRL), to constrain the quantile of the distribution of the cumulative sum cost that is a necessary and sufficient condition to satisfy the outage constraint. This is the first work that tackles the issue of applying the policy gradient theorem to the quantile and provides theoretical results for approximating the gradient of the quantile. Based on the derived theoretical results and the technique of the Lagrange multiplier, we construct a constrained RL algorithm named Quantile Constrained Policy Optimization (QCPO). We use distributional RL with the Large Deviation Principle (LDP) to estimate quantiles and tail probability of the cumulative sum cost for the implementation of QCPO. The implemented algorithm satisfies the outage probability constraint after the training period.
Transfer of recent advances in deep reinforcement learning to real-world applications is hindered by high data demands and thus low efficiency and scalability. Through independent improvements of components such as replay buffers or more stable learning algorithms, and through massively distributed systems, training time could be reduced from several days to several hours for standard benchmark tasks. However, while rewards in simulated environments are well-defined and easy to compute, reward evaluation becomes the bottleneck in many real-world environments, e.g., in molecular optimization tasks, where computationally demanding simulations or even experiments are required to evaluate states and to quantify rewards. Therefore, training might become prohibitively expensive without an extensive amount of computational resources and time. We propose to alleviate this problem by replacing costly ground-truth rewards with rewards modeled by neural networks, counteracting non-stationarity of state and reward distributions during training with an active learning component. We demonstrate that using our proposed ACRL method (Actively learning Costly rewards for Reinforcement Learning), it is possible to train agents in complex real-world environments orders of magnitudes faster. By enabling the application of reinforcement learning methods to new domains, we show that we can find interesting and non-trivial solutions to real-world optimization problems in chemistry, materials science and engineering.
Increasing traffic demands, higher levels of automation, and communication enhancements provide novel design opportunities for future air traffic controllers (ATCs). This article presents a novel deep reinforcement learning (DRL) controller to aid conflict resolution for autonomous free flight. Although DRL has achieved important advancements in this field, the existing works pay little attention to the explainability and safety issues related to DRL controllers, particularly the safety under adversarial attacks. To address those two issues, we design a fully explainable DRL framework wherein we: 1) decompose the coupled Q value learning model into a safety-awareness and efficiency (reach the target) one; and 2) use information from surrounding intruders as inputs, eliminating the needs of central controllers. In our simulated experiments, we show that by decoupling the safety-awareness and efficiency, we can exceed performance on free flight control tasks while dramatically improving explainability on practical. In addition, the safety Q learning module provides rich information about the safety situation of environments. To study the safety under adversarial attacks, we additionally propose an adversarial attack strategy that can impose both safety-oriented and efficiency-oriented attacks. The adversarial aims to minimize safety/efficiency by only attacking the agent at a few time steps. In the experiments, our attack strategy increases as many collisions as the uniform attack (i.e., attacking at every time step) by only attacking the agent four times less often, which provide insights into the capabilities and restrictions of the DRL in future ATC designs. The source code is publicly available at https://github.com/WLeiiiii/Gym-ATC-Attack-Project.
We are interested in learning scalable agents for reinforcement learning that can learn from large-scale, diverse sequential data similar to current large vision and language models. To this end, this paper presents masked decision prediction (MaskDP), a simple and scalable self-supervised pretraining method for reinforcement learning (RL) and behavioral cloning (BC). In our MaskDP approach, we employ a masked autoencoder (MAE) to state-action trajectories, wherein we randomly mask state and action tokens and reconstruct the missing data. By doing so, the model is required to infer masked-out states and actions and extract information about dynamics. We find that masking different proportions of the input sequence significantly helps with learning a better model that generalizes well to multiple downstream tasks. In our empirical study, we find that a MaskDP model gains the capability of zero-shot transfer to new BC tasks, such as single and multiple goal reaching, and it can zero-shot infer skills from a few example transitions. In addition, MaskDP transfers well to offline RL and shows promising scaling behavior w.r.t. to model size. It is amenable to data-efficient finetuning, achieving competitive results with prior methods based on autoregressive pretraining.
imitation provides open-source implementations of imitation and reward learning algorithms in PyTorch. We include three inverse reinforcement learning (IRL) algorithms, three imitation learning algorithms and a preference comparison algorithm. The implementations have been benchmarked against previous results, and automated tests cover 98% of the code. Moreover, the algorithms are implemented in a modular fashion, making it simple to develop novel algorithms in the framework. Our source code, including documentation and examples, is available at https://github.com/HumanCompatibleAI/imitation
Careful prompt design is critical to the use of large language models in zero-shot or few-shot learning. As a consequence, there is a growing interest in automated methods to design optimal prompts. In this work, we propose Test-time Prompt Editing using Reinforcement learning (TEMPERA). In contrast to prior prompt generation methods, TEMPERA can efficiently leverage prior knowledge, is adaptive to different queries and provides an interpretable prompt for every query. To achieve this, we design a novel action space that allows flexible editing of the initial prompts covering a wide set of commonly-used components like instructions, few-shot exemplars, and verbalizers. The proposed method achieves significant gains compared with recent SoTA approaches like prompt tuning, AutoPrompt, and RLPrompt, across a variety of tasks including sentiment analysis, topic classification, natural language inference, and reading comprehension. Our method achieves 5.33x on average improvement in sample efficiency when compared to the traditional fine-tuning methods.
This effort is focused on examining the behavior of reinforcement learning systems in personalization environments and detailing the differences in policy entropy associated with the type of learning algorithm utilized. We demonstrate that Policy Optimization agents often possess low-entropy policies during training, which in practice results in agents prioritizing certain actions and avoiding others. Conversely, we also show that Q-Learning agents are far less susceptible to such behavior and generally maintain high-entropy policies throughout training, which is often preferable in real-world applications. We provide a wide range of numerical experiments as well as theoretical justification to show that these differences in entropy are due to the type of learning being employed.
In recent years, spiking neural networks (SNNs) have been used in reinforcement learning (RL) due to their low power consumption and event-driven features. However, spiking reinforcement learning (SRL), which suffers from fixed coding methods, still faces the problems of high latency and poor versatility. In this paper, we use learnable matrix multiplication to encode and decode spikes, improving the flexibility of the coders and thus reducing latency. Meanwhile, we train the SNNs using the direct training method and use two different structures for online and offline RL algorithms, which gives our model a wider range of applications. Extensive experiments have revealed that our method achieves optimal performance with ultra-low latency (as low as 0.8% of other SRL methods) and excellent energy efficiency (up to 5X the DNNs) in different algorithms and different environments.
Many multiagent systems in the real world include multiple types of agents with different abilities and functionality. Such heterogeneous multiagent systems have significant practical advantages. However, they also come with challenges compared with homogeneous systems for multiagent reinforcement learning, such as the non-stationary problem and the policy version iteration issue. This work proposes a general-purpose reinforcement learning algorithm named Heterogeneous League Training (HLT) to address heterogeneous multiagent problems. HLT keeps track of a pool of policies that agents have explored during training, gathering a league of heterogeneous policies to facilitate future policy optimization. Moreover, a hyper-network is introduced to increase the diversity of agent behaviors when collaborating with teammates having different levels of cooperation skills. We use heterogeneous benchmark tasks to demonstrate that (1) HLT promotes the success rate in cooperative heterogeneous tasks; (2) HLT is an effective approach to solving the policy version iteration problem; (3) HLT provides a practical way to assess the difficulty of learning each role in a heterogeneous team.
Distributed access control is a crucial component for massive machine type communication (mMTC). In this communication scenario, centralized resource allocation is not scalable because resource configurations have to be sent frequently from the base station to a massive number of devices. We investigate distributed reinforcement learning for resource selection without relying on centralized control. Another important feature of mMTC is the sporadic and dynamic change of traffic. Existing studies on distributed access control assume that traffic load is static or they are able to gradually adapt to the dynamic traffic. We minimize the adaptation period by training TinyQMIX, which is a lightweight multi-agent deep reinforcement learning model, to learn a distributed wireless resource selection policy under various traffic patterns before deployment. Therefore, the trained agents are able to quickly adapt to dynamic traffic and provide low access delay. Numerical results are presented to support our claims.
We propose to use reinforcement learning to inform transformer-based contextualized link prediction models by providing paths that are most useful for predicting the correct answer. This is in contrast to previous approaches, that either used reinforcement learning (RL) to directly search for the answer, or based their prediction on limited or randomly selected context. Our experiments on WN18RR and FB15k-237 show that contextualized link prediction models consistently outperform RL-based answer search, and that additional improvements (of up to 13.5% MRR) can be gained by combining RL with a link prediction model. The PyTorch implementation of the RL agent is available at https://github.com/marina-sp/reinform
We propose a provable defense mechanism against backdoor policies in reinforcement learning under subspace trigger assumption. A backdoor policy is a security threat where an adversary publishes a seemingly well-behaved policy which in fact allows hidden triggers. During deployment, the adversary can modify observed states in a particular way to trigger unexpected actions and harm the agent. We assume the agent does not have the resources to re-train a good policy. Instead, our defense mechanism sanitizes the backdoor policy by projecting observed states to a 'safe subspace', estimated from a small number of interactions with a clean (non-triggered) environment. Our sanitized policy achieves $\epsilon$ approximate optimality in the presence of triggers, provided the number of clean interactions is $O\left(\frac{D}{(1-\gamma)^4 \epsilon^2}\right)$ where $\gamma$ is the discounting factor and $D$ is the dimension of state space. Empirically, we show that our sanitization defense performs well on two Atari game environments.
Offline reinforcement learning aims to train a policy on a pre-recorded and fixed dataset without any additional environment interactions. There are two major challenges in this setting: (1) extrapolation error caused by approximating the value of state-action pairs not well-covered by the training data and (2) distributional shift between behavior and inference policies. One way to tackle these problems is to induce conservatism - i.e., keeping the learned policies closer to the behavioral ones. To achieve this, we build upon recent works on learning policies in latent action spaces and use a special form of Normalizing Flows for constructing a generative model, which we use as a conservative action encoder. This Normalizing Flows action encoder is pre-trained in a supervised manner on the offline dataset, and then an additional policy model - controller in the latent space - is trained via reinforcement learning. This approach avoids querying actions outside of the training dataset and therefore does not require additional regularization for out-of-dataset actions. We evaluate our method on various locomotion and navigation tasks, demonstrating that our approach outperforms recently proposed algorithms with generative action models on a large portion of datasets.