We present a novel tool BertRLFuzzer, a BERT and Reinforcement Learning (RL) based fuzzer aimed at finding security vulnerabilities for Web applications. BertRLFuzzer works as follows: given a set of seed inputs, the fuzzer performs grammar-adhering and attack-provoking mutation operations on them to generate candidate attack vectors. The key insight of BertRLFuzzer is the use of RL with a BERT model as an agent to guide the fuzzer to efficiently learn grammar-adhering and attack-provoking mutation operators. In order to establish the efficacy of BertRLFuzzer we compare it against a total of 13 black box and white box fuzzers over a benchmark of 9 victim websites with over 16K LOC. We observed a significant improvement relative to the nearest competing tool in terms of time to first attack (54% less), new vulnerabilities found (17 new vulnerabilities), and attack rate (4.4% more attack vectors generated).
Popular reinforcement learning (RL) algorithms tend to produce a unimodal policy distribution, which weakens the expressiveness of complicated policy and decays the ability of exploration. The diffusion probability model is powerful to learn complicated multimodal distributions, which has shown promising and potential applications to RL. In this paper, we formally build a theoretical foundation of policy representation via the diffusion probability model and provide practical implementations of diffusion policy for online model-free RL. Concretely, we character diffusion policy as a stochastic process, which is a new approach to representing a policy. Then we present a convergence guarantee for diffusion policy, which provides a theory to understand the multimodality of diffusion policy. Furthermore, we propose the DIPO which is an implementation for model-free online RL with DIffusion POlicy. To the best of our knowledge, DIPO is the first algorithm to solve model-free online RL problems with the diffusion model. Finally, extensive empirical results show the effectiveness and superiority of DIPO on the standard continuous control Mujoco benchmark.
Multi-task Imitation Learning (MIL) aims to train a policy capable of performing a distribution of tasks based on multi-task expert demonstrations, which is essential for general-purpose robots. Existing MIL algorithms suffer from low data efficiency and poor performance on complex long-horizontal tasks. We develop Multi-task Hierarchical Adversarial Inverse Reinforcement Learning (MH-AIRL) to learn hierarchically-structured multi-task policies, which is more beneficial for compositional tasks with long horizons and has higher expert data efficiency through identifying and transferring reusable basic skills across tasks. To realize this, MH-AIRL effectively synthesizes context-based multi-task learning, AIRL (an IL approach), and hierarchical policy learning. Further, MH-AIRL can be adopted to demonstrations without the task or skill annotations (i.e., state-action pairs only) which are more accessible in practice. Theoretical justifications are provided for each module of MH-AIRL, and evaluations on challenging multi-task settings demonstrate superior performance and transferability of the multi-task policies learned with MH-AIRL as compared to SOTA MIL baselines.
Deep Reinforcement Learning (DRL) has received a lot of attention from the research community in recent years. As the technology moves away from game playing to practical contexts, such as autonomous vehicles and robotics, it is crucial to evaluate the quality of DRL agents. In this paper, we propose a search-based approach to test such agents. Our approach, implemented in a tool called Indago, trains a classifier on failure and non-failure environment (i.e., pass) configurations resulting from the DRL training process. The classifier is used at testing time as a surrogate model for the DRL agent execution in the environment, predicting the extent to which a given environment configuration induces a failure of the DRL agent under test. The failure prediction acts as a fitness function, guiding the generation towards failure environment configurations, while saving computation time by deferring the execution of the DRL agent in the environment to those configurations that are more likely to expose failures. Experimental results show that our search-based approach finds 50% more failures of the DRL agent than state-of-the-art techniques. Moreover, such failures are, on average, 78% more diverse; similarly, the behaviors of the DRL agent induced by failure configurations are 74% more diverse.
Diffusion models are a class of flexible generative models trained with an approximation to the log-likelihood objective. However, most use cases of diffusion models are not concerned with likelihoods, but instead with downstream objectives such as human-perceived image quality or drug effectiveness. In this paper, we investigate reinforcement learning methods for directly optimizing diffusion models for such objectives. We describe how posing denoising as a multi-step decision-making problem enables a class of policy gradient algorithms, which we refer to as denoising diffusion policy optimization (DDPO), that are more effective than alternative reward-weighted likelihood approaches. Empirically, DDPO is able to adapt text-to-image diffusion models to objectives that are difficult to express via prompting, such as image compressibility, and those derived from human feedback, such as aesthetic quality. Finally, we show that DDPO can improve prompt-image alignment using feedback from a vision-language model without the need for additional data collection or human annotation. The project's website can be found at http://rl-diffusion.github.io .
In combination with Reinforcement Learning, Monte-Carlo Tree Search has shown to outperform human grandmasters in games such as Chess, Shogi and Go with little to no prior domain knowledge. However, most classical use cases only feature up to two players. Scaling the search to an arbitrary number of players presents a computational challenge, especially if decisions have to be planned over a longer time horizon. In this work, we investigate techniques that transform general-sum multiplayer games into single-player and two-player games that consider other agents to act according to given opponent models. For our evaluation, we focus on the challenging Pommerman environment which involves partial observability, a long time horizon and sparse rewards. In combination with our search methods, we investigate the phenomena of opponent modeling using heuristics and self-play. Overall, we demonstrate the effectiveness of our multiplayer search variants both in a supervised learning and reinforcement learning setting.
Millions of slum dwellers suffer from poor accessibility to urban services due to inadequate road infrastructure within slums, and road planning for slums is critical to the sustainable development of cities. Existing re-blocking or heuristic methods are either time-consuming which cannot generalize to different slums, or yield sub-optimal road plans in terms of accessibility and construction costs. In this paper, we present a deep reinforcement learning based approach to automatically layout roads for slums. We propose a generic graph model to capture the topological structure of a slum, and devise a novel graph neural network to select locations for the planned roads. Through masked policy optimization, our model can generate road plans that connect places in a slum at minimal construction costs. Extensive experiments on real-world slums in different countries verify the effectiveness of our model, which can significantly improve accessibility by 14.3% against existing baseline methods. Further investigations on transferring across different tasks demonstrate that our model can master road planning skills in simple scenarios and adapt them to much more complicated ones, indicating the potential of applying our model in real-world slum upgrading. The code and data are available at https://github.com/tsinghua-fib-lab/road-planning-for-slums.
Among the great successes of Reinforcement Learning (RL), self-play algorithms play an essential role in solving competitive games. Current self-play algorithms optimize the agent to maximize expected win-rates against its current or historical copies, making it often stuck in the local optimum and its strategy style simple and homogeneous. A possible solution is to improve the diversity of policies, which helps the agent break the stalemate and enhances its robustness when facing different opponents. However, enhancing diversity in the self-play algorithms is not trivial. In this paper, we aim to introduce diversity from the perspective that agents could have diverse risk preferences in the face of uncertainty. Specifically, we design a novel reinforcement learning algorithm called Risk-sensitive Proximal Policy Optimization (RPPO), which smoothly interpolates between worst-case and best-case policy learning and allows for policy learning with desired risk preferences. Seamlessly integrating RPPO with population-based self-play, agents in the population optimize dynamic risk-sensitive objectives with experiences from playing against diverse opponents. Empirical results show that our method achieves comparable or superior performance in competitive games and that diverse modes of behaviors emerge. Our code is public online at \url{https://github.com/Jackory/RPBT}.
Learning reinforcement learning (RL)-based recommenders from historical user-item interaction sequences is vital to generate high-reward recommendations and improve long-term cumulative benefits. However, existing RL recommendation methods encounter difficulties (i) to estimate the value functions for states which are not contained in the offline training data, and (ii) to learn effective state representations from user implicit feedback due to the lack of contrastive signals. In this work, we propose contrastive state augmentations (CSA) for the training of RL-based recommender systems. To tackle the first issue, we propose four state augmentation strategies to enlarge the state space of the offline data. The proposed method improves the generalization capability of the recommender by making the RL agent visit the local state regions and ensuring the learned value functions are similar between the original and augmented states. For the second issue, we propose introducing contrastive signals between augmented states and the state randomly sampled from other sessions to improve the state representation learning further. To verify the effectiveness of the proposed CSA, we conduct extensive experiments on two publicly accessible datasets and one dataset collected from a real-life e-commerce platform. We also conduct experiments on a simulated environment as the online evaluation setting. Experimental results demonstrate that CSA can effectively improve recommendation performance.
Lifelong learning agents aim to learn multiple tasks sequentially over a lifetime. This involves the ability to exploit previous knowledge when learning new tasks and to avoid forgetting. Modulating masks, a specific type of parameter isolation approach, have recently shown promise in both supervised and reinforcement learning. While lifelong learning algorithms have been investigated mainly within a single-agent approach, a question remains on how multiple agents can share lifelong learning knowledge with each other. We show that the parameter isolation mechanism used by modulating masks is particularly suitable for exchanging knowledge among agents in a distributed and decentralized system of lifelong learners. The key idea is that the isolation of specific task knowledge to specific masks allows agents to transfer only specific knowledge on-demand, resulting in robust and effective distributed lifelong learning. We assume fully distributed and asynchronous scenarios with dynamic agent numbers and connectivity. An on-demand communication protocol ensures agents query their peers for specific masks to be transferred and integrated into their policies when facing each task. Experiments indicate that on-demand mask communication is an effective way to implement distributed lifelong reinforcement learning and provides a lifelong learning benefit with respect to distributed RL baselines such as DD-PPO, IMPALA, and PPO+EWC. The system is particularly robust to connection drops and demonstrates rapid learning due to knowledge exchange.
While reinforcement learning (RL) has achieved great success in acquiring complex skills solely from environmental interactions, it assumes that resets to the initial state are readily available at the end of each episode. Such an assumption hinders the autonomous learning of embodied agents due to the time-consuming and cumbersome workarounds for resetting in the physical world. Hence, there has been a growing interest in autonomous RL (ARL) methods that are capable of learning from non-episodic interactions. However, existing works on ARL are limited by their reliance on prior data and are unable to learn in environments where task-relevant interactions are sparse. In contrast, we propose a demonstration-free ARL algorithm via Implicit and Bi-directional Curriculum (IBC). With an auxiliary agent that is conditionally activated upon learning progress and a bidirectional goal curriculum based on optimal transport, our method outperforms previous methods, even the ones that leverage demonstrations.
As multi-agent reinforcement learning (MARL) systems are increasingly deployed throughout society, it is imperative yet challenging for users to understand the emergent behaviors of MARL agents in complex environments. This work presents an approach for generating policy-level contrastive explanations for MARL to answer a temporal user query, which specifies a sequence of tasks completed by agents with possible cooperation. The proposed approach encodes the temporal query as a PCTL logic formula and checks if the query is feasible under a given MARL policy via probabilistic model checking. Such explanations can help reconcile discrepancies between the actual and anticipated multi-agent behaviors. The proposed approach also generates correct and complete explanations to pinpoint reasons that make a user query infeasible. We have successfully applied the proposed approach to four benchmark MARL domains (up to 9 agents in one domain). Moreover, the results of a user study show that the generated explanations significantly improve user performance and satisfaction.
This research paper presents an experimental approach to using the Reptile algorithm for reinforcement learning to train a neural network to play Super Mario Bros. We implement the Reptile algorithm using the Super Mario Bros Gym library and TensorFlow in Python, creating a neural network model with a single convolutional layer, a flatten layer, and a dense layer. We define the optimizer and use the Reptile class to create an instance of the Reptile meta-learning algorithm. We train the model using multiple tasks and episodes, choosing actions using the current weights of the neural network model, taking those actions in the environment, and updating the model weights using the Reptile algorithm. We evaluate the performance of the algorithm by printing the total reward for each episode. In addition, we compare the performance of the Reptile algorithm approach to two other popular reinforcement learning algorithms, Proximal Policy Optimization (PPO) and Deep Q-Network (DQN), applied to the same Super Mario Bros task. Our results demonstrate that the Reptile algorithm provides a promising approach to few-shot learning in video game AI, with comparable or even better performance than the other two algorithms, particularly in terms of moves vs distance that agent performs for 1M episodes of training. The results shows that best total distance for world 1-2 in the game environment were ~1732 (PPO), ~1840 (DQN) and ~2300 (RAMario). Full code is available at https://github.com/s4nyam/RAMario.
Few multi-agent reinforcement learning (MARL) research on Google Research Football (GRF) focus on the 11v11 multi-agent full-game scenario and to the best of our knowledge, no open benchmark on this scenario has been released to the public. In this work, we fill the gap by providing a population-based MARL training pipeline and hyperparameter settings on multi-agent football scenario that outperforms the bot with difficulty 1.0 from scratch within 2 million steps. Our experiments serve as a reference for the expected performance of Independent Proximal Policy Optimization (IPPO), a state-of-the-art multi-agent reinforcement learning algorithm where each agent tries to maximize its own policy independently across various training configurations. Meanwhile, we open-source our training framework Light-MALib which extends the MALib codebase by distributed and asynchronized implementation with additional analytical tools for football games. Finally, we provide guidance for building strong football AI with population-based training and release diverse pretrained policies for benchmarking. The goal is to provide the community with a head start for whoever experiment their works on GRF and a simple-to-use population-based training framework for further improving their agents through self-play. The implementation is available at https://github.com/Shanghai-Digital-Brain-Laboratory/DB-Football.
Optimization problems over dynamic networks have been extensively studied and widely used in the past decades to formulate numerous real-world problems. However, (1) traditional optimization-based approaches do not scale to large networks, and (2) the design of good heuristics or approximation algorithms often requires significant manual trial-and-error. In this work, we argue that data-driven strategies can automate this process and learn efficient algorithms without compromising optimality. To do so, we present network control problems through the lens of reinforcement learning and propose a graph network-based framework to handle a broad class of problems. Instead of naively computing actions over high-dimensional graph elements, e.g., edges, we propose a bi-level formulation where we (1) specify a desired next state via RL, and (2) solve a convex program to best achieve it, leading to drastically improved scalability and performance. We further highlight a collection of desirable features to system designers, investigate design decisions, and present experiments on real-world control problems showing the utility, scalability, and flexibility of our framework.
Recently, deep reinforcement learning (RL) has been proposed to learn the tractography procedure and train agents to reconstruct the structure of the white matter without manually curated reference streamlines. While the performances reported were competitive, the proposed framework is complex, and little is still known about the role and impact of its multiple parts. In this work, we thoroughly explore the different components of the proposed framework, such as the choice of the RL algorithm, seeding strategy, the input signal and reward function, and shed light on their impact. Approximately 7,400 models were trained for this work, totalling nearly 41,000 hours of GPU time. Our goal is to guide researchers eager to explore the possibilities of deep RL for tractography by exposing what works and what does not work with the category of approach. As such, we ultimately propose a series of recommendations concerning the choice of RL algorithm, the input to the agents, the reward function and more to help future work using reinforcement learning for tractography. We also release the open source codebase, trained models, and datasets for users and researchers wanting to explore reinforcement learning for tractography.
Clinical diagnosis guidelines aim at specifying the steps that may lead to a diagnosis. Inspired by guidelines, we aim to learn the optimal sequence of actions to perform in order to obtain a correct diagnosis from electronic health records. We apply various deep reinforcement learning algorithms to this task and experiment on a synthetic but realistic dataset to differentially diagnose anemia and its subtypes and particularly evaluate the robustness of various approaches to noise and missing data. Experimental results show that the deep reinforcement learning algorithms show competitive performance compared to the state-of-the-art methods with the added advantage that they enable the progressive generation of a pathway to the suggested diagnosis, which can both guide and explain the decision process.
There is a lack of standard benchmarks for Multi-Agent Reinforcement Learning (MARL) algorithms. The Starcraft Multi-Agent Challenge (SMAC) has been widely used in MARL research, but is built on top of a heavy, closed-source computer game, StarCraft II. Thus, SMAC is computationally expensive and requires knowledge and the use of proprietary tools specific to the game for any meaningful alteration or contribution to the environment. We introduce SMAClite -- a challenge based on SMAC that is both decoupled from Starcraft II and open-source, along with a framework which makes it possible to create new content for SMAClite without any special knowledge. We conduct experiments to show that SMAClite is equivalent to SMAC, by training MARL algorithms on SMAClite and reproducing SMAC results. We then show that SMAClite outperforms SMAC in both runtime speed and memory.
We apply reinforcement learning techniques to topic modeling by replacing the variational autoencoder in ProdLDA with a continuous action space reinforcement learning policy. We train the system with a policy gradient algorithm REINFORCE. Additionally, we introduced several modifications: modernize the neural network architecture, weight the ELBO loss, use contextual embeddings, and monitor the learning process via computing topic diversity and coherence for each training step. Experiments are performed on 11 data sets. Our unsupervised model outperforms all other unsupervised models and performs on par with or better than most models using supervised labeling. Our model is outperformed on certain data sets by a model using supervised labeling and contrastive learning. We have also conducted an ablation study to provide empirical evidence of performance improvements from changes we made to ProdLDA and found that the reinforcement learning formulation boosts performance.
Policy optimization methods with function approximation are widely used in multi-agent reinforcement learning. However, it remains elusive how to design such algorithms with statistical guarantees. Leveraging a multi-agent performance difference lemma that characterizes the landscape of multi-agent policy optimization, we find that the localized action value function serves as an ideal descent direction for each local policy. Motivated by the observation, we present a multi-agent PPO algorithm in which the local policy of each agent is updated similarly to vanilla PPO. We prove that with standard regularity conditions on the Markov game and problem-dependent quantities, our algorithm converges to the globally optimal policy at a sublinear rate. We extend our algorithm to the off-policy setting and introduce pessimism to policy evaluation, which aligns with experiments. To our knowledge, this is the first provably convergent multi-agent PPO algorithm in cooperative Markov games.