Deep reinforcement learning (DRL) is a promising method to learn control policies for robots only from demonstration and experience. To cover the whole dynamic behaviour of the robot, DRL training is an active exploration process typically performed in simulation environments. Although this simulation training is cheap and fast, applying DRL algorithms to real-world settings is difficult. If agents are trained until they perform safely in simulation, transferring them to physical systems is difficult due to the sim-to-real gap caused by the difference between the simulation dynamics and the physical robot. In this paper, we present a method of online training a DRL agent to drive autonomously on a physical vehicle by using a model-based safety supervisor. Our solution uses a supervisory system to check if the action selected by the agent is safe or unsafe and ensure that a safe action is always implemented on the vehicle. With this, we can bypass the sim-to-real problem while training the DRL algorithm safely, quickly, and efficiently. We compare our method with conventional learning in simulation and on a physical vehicle. We provide a variety of real-world experiments where we train online a small-scale vehicle to drive autonomously with no prior simulation training. The evaluation results show that our method trains agents with improved sample efficiency while never crashing, and the trained agents demonstrate better driving performance than those trained in simulation.
相關內容
Model predictive control (MPC) is a promising technique for motion cueing in driving simulators, but its high computation time limits widespread real-time application. This paper proposes a hybrid algorithm that combines filter-based and MPC-based techniques to improve specific force tracking while reducing computation time. The proposed algorithm divides the reference acceleration into low-frequency and high-frequency components. The high-frequency component serves as a reference for translational motion to avoid workspace limit violations, while the low-frequency component is for tilt coordination. The total acceleration serves as a reference for combined specific force with the highest priority to enable compensation of deviations from its reference values. The algorithm uses constraints in the MPC formulation to account for workspace limits and workspace management is applied. The investigated scenarios were a step signal, a multi-sine wave and a recorded real-drive slalom maneuver. Based on the conducted simulations, the algorithm produces approximately 15% smaller root means squared error (RMSE) for the step signal compared to the state-of-the-art. Around 16% improvement is observed when the real-drive scenario is used as the simulation scenario, and for the multi-sine wave, 90% improvement is observed. At higher prediction horizons the algorithm matches the performance of a state-of-the-art MPC-based motion cueing algorithm. Finally, for all prediction horizons, the frequency-splitting algorithm produced faster results. The pre-generated references reduce the required prediction horizon and computational complexity while improving tracking performance. Hence, the proposed frequency-splitting algorithm outperforms state-of-the-art MPC-based algorithm and offers promise for real-time application in driving simulators.
The vulnerabilities to backdoor attacks have recently threatened the trustworthiness of machine learning models in practical applications. Conventional wisdom suggests that not everyone can be an attacker since the process of designing the trigger generation algorithm often involves significant effort and extensive experimentation to ensure the attack's stealthiness and effectiveness. Alternatively, this paper shows that there exists a more severe backdoor threat: anyone can exploit an easily-accessible algorithm for silent backdoor attacks. Specifically, this attacker can employ the widely-used lossy image compression from a plethora of compression tools to effortlessly inject a trigger pattern into an image without leaving any noticeable trace; i.e., the generated triggers are natural artifacts. One does not require extensive knowledge to click on the "convert" or "save as" button while using tools for lossy image compression. Via this attack, the adversary does not need to design a trigger generator as seen in prior works and only requires poisoning the data. Empirically, the proposed attack consistently achieves 100% attack success rate in several benchmark datasets such as MNIST, CIFAR-10, GTSRB and CelebA. More significantly, the proposed attack can still achieve almost 100% attack success rate with very small (approximately 10%) poisoning rates in the clean label setting. The generated trigger of the proposed attack using one lossy compression algorithm is also transferable across other related compression algorithms, exacerbating the severity of this backdoor threat. This work takes another crucial step toward understanding the extensive risks of backdoor attacks in practice, urging practitioners to investigate similar attacks and relevant backdoor mitigation methods.
Serverless computing (FaaS) has been extensively utilized for deep learning (DL) inference due to the ease of deployment and pay-per-use benefits. However, existing FaaS platforms utilize GPUs in a coarse manner for DL inferences, without taking into account spatio-temporal resource multiplexing and isolation, which results in severe GPU under-utilization, high usage expenses, and SLO (Service Level Objectives) violation. There is an imperative need to enable an efficient and SLO-aware GPU-sharing mechanism in serverless computing to facilitate cost-effective DL inferences. In this paper, we propose \textbf{FaST-GShare}, an efficient \textit{\textbf{Fa}aS-oriented \textbf{S}patio-\textbf{T}emporal \textbf{G}PU \textbf{Sharing}} architecture for deep learning inferences. In the architecture, we introduce the FaST-Manager to limit and isolate spatio-temporal resources for GPU multiplexing. In order to realize function performance, the automatic and flexible FaST-Profiler is proposed to profile function throughput under various resource allocations. Based on the profiling data and the isolation mechanism, we introduce the FaST-Scheduler with heuristic auto-scaling and efficient resource allocation to guarantee function SLOs. Meanwhile, FaST-Scheduler schedules function with efficient GPU node selection to maximize GPU usage. Furthermore, model sharing is exploited to mitigate memory contention. Our prototype implementation on the OpenFaaS platform and experiments on MLPerf-based benchmark prove that FaST-GShare can ensure resource isolation and function SLOs. Compared to the time sharing mechanism, FaST-GShare can improve throughput by 3.15x, GPU utilization by 1.34x, and SM (Streaming Multiprocessor) occupancy by 3.13x on average.
Dialogue policy learning (DPL) is a crucial component of dialogue modelling. Its primary role is to determine the appropriate abstract response, commonly referred to as the "dialogue action". Traditional DPL methodologies have treated this as a sequential decision problem, using pre-defined action candidates extracted from a corpus. However, these incomplete candidates can significantly limit the diversity of responses and pose challenges when dealing with edge cases, which are scenarios that occur only at extreme operating parameters. To address these limitations, we introduce a novel framework, JoTR. This framework is unique as it leverages a text-to-text Transformer-based model to generate flexible dialogue actions. Unlike traditional methods, JoTR formulates a word-level policy that allows for a more dynamic and adaptable dialogue action generation, without the need for any action templates. This setting enhances the diversity of responses and improves the system's ability to handle edge cases effectively. In addition, JoTR employs reinforcement learning with a reward-shaping mechanism to efficiently finetune the word-level dialogue policy, which allows the model to learn from its interactions, improving its performance over time. We conducted an extensive evaluation of JoTR to assess its effectiveness. Our extensive evaluation shows that JoTR achieves state-of-the-art performance on two benchmark dialogue modelling tasks, as assessed by both user simulators and human evaluators.
The existence of representative datasets is a prerequisite of many successful artificial intelligence and machine learning models. However, the subsequent application of these models often involves scenarios that are inadequately represented in the data used for training. The reasons for this are manifold and range from time and cost constraints to ethical considerations. As a consequence, the reliable use of these models, especially in safety-critical applications, is a huge challenge. Leveraging additional, already existing sources of knowledge is key to overcome the limitations of purely data-driven approaches, and eventually to increase the generalization capability of these models. Furthermore, predictions that conform with knowledge are crucial for making trustworthy and safe decisions even in underrepresented scenarios. This work provides an overview of existing techniques and methods in the literature that combine data-based models with existing knowledge. The identified approaches are structured according to the categories integration, extraction and conformity. Special attention is given to applications in the field of autonomous driving.
TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classication
Multiple instance learning (MIL) is a powerful tool to solve the weakly supervised classification in whole slide image (WSI) based pathology diagnosis. However, the current MIL methods are usually based on independent and identical distribution hypothesis, thus neglect the correlation among different instances. To address this problem, we proposed a new framework, called correlated MIL, and provided a proof for convergence. Based on this framework, we devised a Transformer based MIL (TransMIL), which explored both morphological and spatial information. The proposed TransMIL can effectively deal with unbalanced/balanced and binary/multiple classification with great visualization and interpretability. We conducted various experiments for three different computational pathology problems and achieved better performance and faster convergence compared with state-of-the-art methods. The test AUC for the binary tumor classification can be up to 93.09% over CAMELYON16 dataset. And the AUC over the cancer subtypes classification can be up to 96.03% and 98.82% over TCGA-NSCLC dataset and TCGA-RCC dataset, respectively.
There recently has been a surge of interest in developing a new class of deep learning (DL) architectures that integrate an explicit time dimension as a fundamental building block of learning and representation mechanisms. In turn, many recent results show that topological descriptors of the observed data, encoding information on the shape of the dataset in a topological space at different scales, that is, persistent homology of the data, may contain important complementary information, improving both performance and robustness of DL. As convergence of these two emerging ideas, we propose to enhance DL architectures with the most salient time-conditioned topological information of the data and introduce the concept of zigzag persistence into time-aware graph convolutional networks (GCNs). Zigzag persistence provides a systematic and mathematically rigorous framework to track the most important topological features of the observed data that tend to manifest themselves over time. To integrate the extracted time-conditioned topological descriptors into DL, we develop a new topological summary, zigzag persistence image, and derive its theoretical stability guarantees. We validate the new GCNs with a time-aware zigzag topological layer (Z-GCNETs), in application to traffic forecasting and Ethereum blockchain price prediction. Our results indicate that Z-GCNET outperforms 13 state-of-the-art methods on 4 time series datasets.
Few sample learning (FSL) is significant and challenging in the field of machine learning. The capability of learning and generalizing from very few samples successfully is a noticeable demarcation separating artificial intelligence and human intelligence since humans can readily establish their cognition to novelty from just a single or a handful of examples whereas machine learning algorithms typically entail hundreds or thousands of supervised samples to guarantee generalization ability. Despite the long history dated back to the early 2000s and the widespread attention in recent years with booming deep learning technologies, little surveys or reviews for FSL are available until now. In this context, we extensively review 200+ papers of FSL spanning from the 2000s to 2019 and provide a timely and comprehensive survey for FSL. In this survey, we review the evolution history as well as the current progress on FSL, categorize FSL approaches into the generative model based and discriminative model based kinds in principle, and emphasize particularly on the meta learning based FSL approaches. We also summarize several recently emerging extensional topics of FSL and review the latest advances on these topics. Furthermore, we highlight the important FSL applications covering many research hotspots in computer vision, natural language processing, audio and speech, reinforcement learning and robotic, data analysis, etc. Finally, we conclude the survey with a discussion on promising trends in the hope of providing guidance and insights to follow-up researches.
The notion of uncertainty is of major importance in machine learning and constitutes a key element of machine learning methodology. In line with the statistical tradition, uncertainty has long been perceived as almost synonymous with standard probability and probabilistic predictions. Yet, due to the steadily increasing relevance of machine learning for practical applications and related issues such as safety requirements, new problems and challenges have recently been identified by machine learning scholars, and these problems may call for new methodological developments. In particular, this includes the importance of distinguishing between (at least) two different types of uncertainty, often refereed to as aleatoric and epistemic. In this paper, we provide an introduction to the topic of uncertainty in machine learning as well as an overview of hitherto attempts at handling uncertainty in general and formalizing this distinction in particular.
Reinforcement learning (RL) is a popular paradigm for addressing sequential decision tasks in which the agent has only limited environmental feedback. Despite many advances over the past three decades, learning in many domains still requires a large amount of interaction with the environment, which can be prohibitively expensive in realistic scenarios. To address this problem, transfer learning has been applied to reinforcement learning such that experience gained in one task can be leveraged when starting to learn the next, harder task. More recently, several lines of research have explored how tasks, or data samples themselves, can be sequenced into a curriculum for the purpose of learning a problem that may otherwise be too difficult to learn from scratch. In this article, we present a framework for curriculum learning (CL) in reinforcement learning, and use it to survey and classify existing CL methods in terms of their assumptions, capabilities, and goals. Finally, we use our framework to find open problems and suggest directions for future RL curriculum learning research.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191