Offline reinforcement learning (RL) aims to learn an effective policy from a pre-collected dataset. Most existing works are to develop sophisticated learning algorithms, with less emphasis on improving the data collection process. Moreover, it is even challenging to extend the single-task setting and collect a task-agnostic dataset that allows an agent to perform multiple downstream tasks. In this paper, we propose a Curiosity-driven Unsupervised Data Collection (CUDC) method to expand feature space using adaptive temporal distances for task-agnostic data collection and ultimately improve learning efficiency and capabilities for multi-task offline RL. To achieve this, CUDC estimates the probability of the k-step future states being reachable from the current states, and adapts how many steps into the future that the dynamics model should predict. With this adaptive reachability mechanism in place, the feature representation can be diversified, and the agent can navigate itself to collect higher-quality data with curiosity. Empirically, CUDC surpasses existing unsupervised methods in efficiency and learning performance in various downstream offline RL tasks of the DeepMind control suite.
相關內容
Many imitation learning (IL) algorithms employ inverse reinforcement learning (IRL) to infer the intrinsic reward function that an expert is implicitly optimizing for based on their demonstrated behaviors. However, in practice, IRL-based IL can fail to accomplish the underlying task due to a misalignment between the inferred reward and the objective of the task. In this paper, we address the susceptibility of IL to such misalignment by introducing a semi-supervised reward design paradigm called Protagonist Antagonist Guided Adversarial Reward (PAGAR). PAGAR-based IL trains a policy to perform well under mixed reward functions instead of a single reward function as in IRL-based IL. We identify the theoretical conditions under which PAGAR-based IL can avoid the task failures caused by reward misalignment. We also present a practical on-and-off policy approach to implementing PAGAR-based IL. Experimental results show that our algorithm outperforms standard IL baselines in complex tasks and challenging transfer settings.
Offline reinforcement learning (RL) aims to learn a policy that maximizes the expected cumulative reward using a pre-collected dataset. Offline RL with low-rank MDPs or general function approximation has been widely studied recently, but existing algorithms with sample complexity $O(\epsilon^{-2})$ for finding an $\epsilon$-optimal policy either require a uniform data coverage assumptions or are computationally inefficient. In this paper, we propose a primal dual algorithm for offline RL with low-rank MDPs in the discounted infinite-horizon setting. Our algorithm is the first computationally efficient algorithm in this setting that achieves sample complexity of $O(\epsilon^{-2})$ with partial data coverage assumption. This improves upon a recent work that requires $O(\epsilon^{-4})$ samples. Moreover, our algorithm extends the previous work to the offline constrained RL setting by supporting constraints on additional reward signals.
Offline reinforcement learning (RL) has attracted much attention due to its ability in learning from static offline datasets and eliminating the need of interacting with the environment. Nevertheless, the success of offline RL relies heavily on the offline transitions annotated with reward labels. In practice, we often need to hand-craft the reward function, which is sometimes difficult, labor-intensive, or inefficient. To tackle this challenge, we set our focus on the offline imitation learning (IL) setting, and aim at getting a reward function based on the expert data and unlabeled data. To that end, we propose a simple yet effective search-based offline IL method, tagged SEABO. SEABO allocates a larger reward to the transition that is close to its closest neighbor in the expert demonstration, and a smaller reward otherwise, all in an unsupervised learning manner. Experimental results on a variety of D4RL datasets indicate that SEABO can achieve competitive performance to offline RL algorithms with ground-truth rewards, given only a single expert trajectory, and can outperform prior reward learning and offline IL methods across many tasks. Moreover, we demonstrate that SEABO also works well if the expert demonstrations contain only observations. Our code is publicly available at //github.com/dmksjfl/SEABO.
In this work we present FreDSNet, a deep learning solution which obtains semantic 3D understanding of indoor environments from single panoramas. Omnidirectional images reveal task-specific advantages when addressing scene understanding problems due to the 360-degree contextual information about the entire environment they provide. However, the inherent characteristics of the omnidirectional images add additional problems to obtain an accurate detection and segmentation of objects or a good depth estimation. To overcome these problems, we exploit convolutions in the frequential domain obtaining a wider receptive field in each convolutional layer. These convolutions allow to leverage the whole context information from omnidirectional images. FreDSNet is the first network that jointly provides monocular depth estimation and semantic segmentation from a single panoramic image exploiting fast Fourier convolutions. Our experiments show that FreDSNet has similar performance as specific state of the art methods for semantic segmentation and depth estimation. FreDSNet code is publicly available in //github.com/Sbrunoberenguel/FreDSNet
The development of machine learning models requires a large amount of training data. Data marketplaces are essential for trading high-quality, private-domain data not publicly available online. However, due to growing data privacy concerns, direct data exchange is inappropriate. Federated Learning (FL) is a distributed machine learning paradigm that exchanges data utilities (in form of local models or gradients) among multiple parties without directly sharing the raw data. However, several challenges exist when applying existing FL architectures to construct a data marketplace: (i) In existing FL architectures, Data Acquirers (DAs) cannot privately evaluate local models from Data Providers (DPs) prior to trading; (ii) Model aggregation protocols in existing FL designs struggle to exclude malicious DPs without "overfitting" to the DA's (possibly biased) root dataset; (iii) Prior FL designs lack a proper billing mechanism to enforce the DA to fairly allocate the reward according to contributions made by different DPs. To address above challenges, we propose martFL, the first federated learning architecture that is specifically designed to enable a secure utility-driven data marketplace. At a high level, martFL is powered by two innovative designs: (i) a quality-aware model aggregation protocol that achieves robust local model aggregation even when the DA's root dataset is biased; (ii) a verifiable data transaction protocol that enables the DA to prove, both succinctly and in zero-knowledge, that it has faithfully aggregates the local models submitted by different DPs according to the committed aggregation weights, based on which the DPs can unambiguously claim the corresponding reward. We implement a prototype of martFL and evaluate it extensively over various tasks. The results show that martFL can improve the model accuracy by up to 25% while saving up to 64% data acquisition cost.
In pervasive machine learning, especially in Human Behavior Analysis (HBA), RGB has been the primary modality due to its accessibility and richness of information. However, linked with its benefits are challenges, including sensitivity to lighting conditions and privacy concerns. One possibility to overcome these vulnerabilities is to resort to different modalities. For instance, thermal is particularly adept at accentuating human forms, while depth adds crucial contextual layers. Despite their known benefits, only a few HBA-specific datasets that integrate these modalities exist. To address this shortage, our research introduces a novel generative technique for creating trimodal, i.e., RGB, thermal, and depth, human-focused datasets. This technique capitalizes on human segmentation masks derived from RGB images, combined with thermal and depth backgrounds that are sourced automatically. With these two ingredients, we synthesize depth and thermal counterparts from existing RGB data utilizing conditional image-to-image translation. By employing this approach, we generate trimodal data that can be leveraged to train models for settings with limited data, bad lightning conditions, or privacy-sensitive areas.
Deep learning has shown great potential for modeling the physical dynamics of complex particle systems such as fluids (in Lagrangian descriptions). Existing approaches, however, require the supervision of consecutive particle properties, including positions and velocities. In this paper, we consider a partially observable scenario known as fluid dynamics grounding, that is, inferring the state transitions and interactions within the fluid particle systems from sequential visual observations of the fluid surface. We propose a differentiable two-stage network named NeuroFluid. Our approach consists of (i) a particle-driven neural renderer, which involves fluid physical properties into the volume rendering function, and (ii) a particle transition model optimized to reduce the differences between the rendered and the observed images. NeuroFluid provides the first solution to unsupervised learning of particle-based fluid dynamics by training these two models jointly. It is shown to reasonably estimate the underlying physics of fluids with different initial shapes, viscosity, and densities. It is a potential alternative approach to understanding complex fluid mechanics, such as turbulence, that are difficult to model using traditional methods of mathematical physics.
Causality knowledge is vital to building robust AI systems. Deep learning models often perform poorly on tasks that require causal reasoning, which is often derived using some form of commonsense knowledge not immediately available in the input but implicitly inferred by humans. Prior work has unraveled spurious observational biases that models fall prey to in the absence of causality. While language representation models preserve contextual knowledge within learned embeddings, they do not factor in causal relationships during training. By blending causal relationships with the input features to an existing model that performs visual cognition tasks (such as scene understanding, video captioning, video question-answering, etc.), better performance can be achieved owing to the insight causal relationships bring about. Recently, several models have been proposed that have tackled the task of mining causal data from either the visual or textual modality. However, there does not exist widespread research that mines causal relationships by juxtaposing the visual and language modalities. While images offer a rich and easy-to-process resource for us to mine causality knowledge from, videos are denser and consist of naturally time-ordered events. Also, textual information offers details that could be implicit in videos. We propose iReason, a framework that infers visual-semantic commonsense knowledge using both videos and natural language captions. Furthermore, iReason's architecture integrates a causal rationalization module to aid the process of interpretability, error analysis and bias detection. We demonstrate the effectiveness of iReason using a two-pronged comparative analysis with language representation learning models (BERT, GPT-2) as well as current state-of-the-art multimodal causality models.
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.
We study the problem of learning to reason in large scale knowledge graphs (KGs). More specifically, we describe a novel reinforcement learning framework for learning multi-hop relational paths: we use a policy-based agent with continuous states based on knowledge graph embeddings, which reasons in a KG vector space by sampling the most promising relation to extend its path. In contrast to prior work, our approach includes a reward function that takes the accuracy, diversity, and efficiency into consideration. Experimentally, we show that our proposed method outperforms a path-ranking based algorithm and knowledge graph embedding methods on Freebase and Never-Ending Language Learning datasets.
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