A central task in control theory, artificial intelligence, and formal methods is to synthesize reward-maximizing strategies for agents that operate in partially unknown environments. In environments modeled by gray-box Markov decision processes (MDPs), the impact of the agents' actions are known in terms of successor states but not the stochastics involved. In this paper, we devise a strategy synthesis algorithm for gray-box MDPs via reinforcement learning that utilizes interval MDPs as internal model. To compete with limited sampling access in reinforcement learning, we incorporate two novel concepts into our algorithm, focusing on rapid and successful learning rather than on stochastic guarantees and optimality: lower confidence bound exploration reinforces variants of already learned practical strategies and action scoping reduces the learning action space to promising actions. We illustrate benefits of our algorithms by means of a prototypical implementation applied on examples from the AI and formal methods communities.
相關內容
The proximal policy optimization (PPO) algorithm stands as one of the most prosperous methods in the field of reinforcement learning (RL). Despite its success, the theoretical understanding of PPO remains deficient. Specifically, it is unclear whether PPO or its optimistic variants can effectively solve linear Markov decision processes (MDPs), which are arguably the simplest models in RL with function approximation. To bridge this gap, we propose an optimistic variant of PPO for episodic adversarial linear MDPs with full-information feedback, and establish a $\tilde{\mathcal{O}}(d^{3/4}H^2K^{3/4})$ regret for it. Here $d$ is the ambient dimension of linear MDPs, $H$ is the length of each episode, and $K$ is the number of episodes. Compared with existing policy-based algorithms, we achieve the state-of-the-art regret bound in both stochastic linear MDPs and adversarial linear MDPs with full information. Additionally, our algorithm design features a novel multi-batched updating mechanism and the theoretical analysis utilizes a new covering number argument of value and policy classes, which might be of independent interest.
Recent years have witnessed a huge demand for artificial intelligence and machine learning applications in wireless edge networks to assist individuals with real-time services. Owing to the practical setting and privacy preservation of federated learning (FL), it is a suitable and appealing distributed learning paradigm to deploy these applications at the network edge. Despite the many successful efforts made to apply FL to wireless edge networks, the adopted algorithms mostly follow the same spirit as FedAvg, thereby heavily suffering from the practical challenges of label deficiency and device heterogeneity. These challenges not only decelerate the model training in FL but also downgrade the application performance. In this paper, we focus on the algorithm design and address these challenges by investigating the personalized semi-supervised FL problem and proposing an effective algorithm, namely FedCPSL. In particular, the techniques of pseudo-labeling, and interpolation-based model personalization are judiciously combined to provide a new problem formulation for personalized semi-supervised FL. The proposed FedCPSL algorithm adopts novel strategies, including adaptive client variance reduction, local momentum, and normalized global aggregation, to combat the challenge of device heterogeneity and boost algorithm convergence. Moreover, the convergence property of FedCPSL is thoroughly analyzed and shows that FedCPSL is resilient to both statistical and system heterogeneity, obtaining a sublinear convergence rate. Experimental results on image classification tasks are also presented to demonstrate that the proposed approach outperforms its counterparts in terms of both convergence speed and application performance.
Prior beliefs about the latent function to shape inductive biases can be incorporated into a Gaussian Process (GP) via the kernel. However, beyond kernel choices, the decision-making process of GP models remains poorly understood. In this work, we contribute an analysis of the loss landscape for GP models using methods from physics. We demonstrate $\nu$-continuity for Matern kernels and outline aspects of catastrophe theory at critical points in the loss landscape. By directly including $\nu$ in the hyperparameter optimisation for Matern kernels, we find that typical values of $\nu$ are far from optimal in terms of performance, yet prevail in the literature due to the increased computational speed. We also provide an a priori method for evaluating the effect of GP ensembles and discuss various voting approaches based on physical properties of the loss landscape. The utility of these approaches is demonstrated for various synthetic and real datasets. Our findings provide an enhanced understanding of the decision-making process behind GPs and offer practical guidance for improving their performance and interpretability in a range of applications.
Random Search is one of the most widely-used method for Hyperparameter Optimization, and is critical to the success of deep learning models. Despite its astonishing performance, little non-heuristic theory has been developed to describe the underlying working mechanism. This paper gives a theoretical accounting of Random Search. We introduce the concept of \emph{scattering dimension} that describes the landscape of the underlying function, and quantifies the performance of random search. We show that, when the environment is noise-free, the output of random search converges to the optimal value in probability at rate $ \widetilde{\mathcal{O}} \left( \left( \frac{1}{T} \right)^{ \frac{1}{d_s} } \right) $, where $ d_s \ge 0 $ is the scattering dimension of the underlying function. When the observed function values are corrupted by bounded $iid$ noise, the output of random search converges to the optimal value in probability at rate $ \widetilde{\mathcal{O}} \left( \left( \frac{1}{T} \right)^{ \frac{1}{d_s + 1} } \right) $. In addition, based on the principles of random search, we introduce an algorithm, called BLiN-MOS, for Lipschitz bandits in doubling metric spaces that are also endowed with a Borel measure, and show that BLiN-MOS achieves a regret rate of order $ \widetilde{\mathcal{O}} \left( T^{ \frac{d_z}{d_z + 1} } \right) $, where $d_z$ is the zooming dimension of the problem instance. Our results show that under certain conditions, the known information-theoretical lower bounds for Lipschitz bandits $\Omega \left( T^{\frac{d_z+1}{d_z+2}} \right)$ can be improved.
Large language models (LLMs) encode a vast amount of world knowledge acquired from massive text datasets. Recent studies have demonstrated that LLMs can assist an algorithm agent in solving complex sequential decision making tasks in embodied environments by providing high-level instructions. However, interacting with LLMs can be time-consuming, as in many practical scenarios, they require a significant amount of storage space that can only be deployed on remote cloud server nodes. Additionally, using commercial LLMs can be costly since they may charge based on usage frequency. In this paper, we explore how to enable efficient and cost-effective interactions between the agent and an LLM. We propose a reinforcement learning based mediator model that determines when it is necessary to consult LLMs for high-level instructions to accomplish a target task. Experiments on 4 MiniGrid environments that entail planning sub-goals demonstrate that our method can learn to solve target tasks with only a few necessary interactions with an LLM, significantly reducing interaction costs in testing environments, compared with baseline methods. Experimental results also suggest that by learning a mediator model to interact with the LLM, the agent's performance becomes more robust against both exploratory and stochastic environments.
The distributed flocking control of collective aerial vehicles has extraordinary advantages in scalability and reliability, \emph{etc.} However, it is still challenging to design a reliable, efficient, and responsive flocking algorithm. In this paper, a distributed predictive flocking framework is presented based on a Markov random field (MRF). The MRF is used to characterize the optimization problem that is eventually resolved by discretizing the input space. Potential functions are employed to describe the interactions between aerial vehicles and as indicators of flight performance. The dynamic constraints are taken into account in the candidate feasible trajectories which correspond to random variables. Numerical simulation shows that compared with some existing latest methods, the proposed algorithm has better-flocking cohesion and control efficiency performances. Experiments are also conducted to demonstrate the feasibility of the proposed algorithm.
Modern communication networks are increasingly equipped with in-network computational capabilities and services. Routing in such networks is significantly more complicated than the traditional routing. A legitimate route for a flow not only needs to have enough communication and computation resources, but also has to conform to various application-specific routing constraints. This paper presents a comprehensive study on routing optimization problems in networks with embedded computational services. We develop a set of routing optimization models and derive low-complexity heuristic routing algorithms for diverse computation scenarios. For dynamic demands, we also develop an online routing algorithm with performance guarantees. Through evaluations over emerging applications on real topologies, we demonstrate that our models can be flexibly customized to meet the diverse routing requirements of different computation applications. Our proposed heuristic algorithms significantly outperform baseline algorithms and can achieve close-to-optimal performance in various scenarios.
Operating on the principles of quantum mechanics, quantum algorithms hold the promise for solving problems that are beyond the reach of the best-available classical algorithms. An integral part of realizing such speedup is the implementation of quantum queries, which read data into forms that quantum computers can process. Quantum random access memory (QRAM) is a promising architecture for realizing quantum queries. However, implementing QRAM in practice poses significant challenges, including query latency, memory capacity and fault-tolerance. In this paper, we propose the first end-to-end system architecture for QRAM. First, we introduce a novel QRAM that hybridizes two existing implementations and achieves asymptotically superior scaling in space (qubit number) and time (circuit depth). Like in classical virtual memory, our construction enables queries to a virtual address space larger than what is actually available in hardware. Second, we present a compilation framework to synthesize, map, and schedule QRAM circuits on realistic hardware. For the first time, we demonstrate how to embed large-scale QRAM on a 2D Euclidean space, such as a grid layout, with minimal routing overhead. Third, we show how to leverage the intrinsic biased-noise resilience of the proposed QRAM for implementation on either Noisy Intermediate-Scale Quantum (NISQ) or Fault-Tolerant Quantum Computing (FTQC) hardware. Finally, we validate these results numerically via both classical simulation and quantum hardware experimentation. Our novel Feynman-path-based simulator allows for efficient simulation of noisy QRAM circuits at a larger scale than previously possible. Collectively, our results outline the set of software and hardware controls needed to implement practical QRAM.
Adversarial attack is a technique for deceiving Machine Learning (ML) models, which provides a way to evaluate the adversarial robustness. In practice, attack algorithms are artificially selected and tuned by human experts to break a ML system. However, manual selection of attackers tends to be sub-optimal, leading to a mistakenly assessment of model security. In this paper, a new procedure called Composite Adversarial Attack (CAA) is proposed for automatically searching the best combination of attack algorithms and their hyper-parameters from a candidate pool of \textbf{32 base attackers}. We design a search space where attack policy is represented as an attacking sequence, i.e., the output of the previous attacker is used as the initialization input for successors. Multi-objective NSGA-II genetic algorithm is adopted for finding the strongest attack policy with minimum complexity. The experimental result shows CAA beats 10 top attackers on 11 diverse defenses with less elapsed time (\textbf{6 $\times$ faster than AutoAttack}), and achieves the new state-of-the-art on $l_{\infty}$, $l_{2}$ and unrestricted adversarial attacks.
Object detection typically assumes that training and test data are drawn from an identical distribution, which, however, does not always hold in practice. Such a distribution mismatch will lead to a significant performance drop. In this work, we aim to improve the cross-domain robustness of object detection. We tackle the domain shift on two levels: 1) the image-level shift, such as image style, illumination, etc, and 2) the instance-level shift, such as object appearance, size, etc. We build our approach based on the recent state-of-the-art Faster R-CNN model, and design two domain adaptation components, on image level and instance level, to reduce the domain discrepancy. The two domain adaptation components are based on H-divergence theory, and are implemented by learning a domain classifier in adversarial training manner. The domain classifiers on different levels are further reinforced with a consistency regularization to learn a domain-invariant region proposal network (RPN) in the Faster R-CNN model. We evaluate our newly proposed approach using multiple datasets including Cityscapes, KITTI, SIM10K, etc. The results demonstrate the effectiveness of our proposed approach for robust object detection in various domain shift scenarios.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191