Large discrete action spaces remain a central challenge for reinforcement learning methods. Such spaces are encountered in many real-world applications, e.g., recommender systems, multi-step planning, and inventory replenishment. The mapping of continuous proxies to discrete actions is a promising paradigm for handling large discrete action spaces. Existing continuous-to-discrete mapping approaches involve searching for discrete neighboring actions in a static pre-defined neighborhood, which requires discrete neighbor lookups across the entire action space. Hence, scalability issues persist. To mitigate this drawback, we propose a novel Dynamic Neighborhood Construction (DNC) method, which dynamically constructs a discrete neighborhood to map the continuous proxy, thus efficiently exploiting the underlying action space. We demonstrate the robustness of our method by benchmarking it against three state-of-the-art approaches designed for large discrete action spaces across three different environments. Our results show that DNC matches or outperforms state-of-the-art approaches while being more computationally efficient. Furthermore, our method scales to action spaces that so far remained computationally intractable for existing methodologies.
相關內容
The mechanisms behind the success of multi-view self-supervised learning (MVSSL) are not yet fully understood. Contrastive MVSSL methods have been studied through the lens of InfoNCE, a lower bound of the Mutual Information (MI). However, the relation between other MVSSL methods and MI remains unclear. We consider a different lower bound on the MI consisting of an entropy and a reconstruction term (ER), and analyze the main MVSSL families through its lens. Through this ER bound, we show that clustering-based methods such as DeepCluster and SwAV maximize the MI. We also re-interpret the mechanisms of distillation-based approaches such as BYOL and DINO, showing that they explicitly maximize the reconstruction term and implicitly encourage a stable entropy, and we confirm this empirically. We show that replacing the objectives of common MVSSL methods with this ER bound achieves competitive performance, while making them stable when training with smaller batch sizes or smaller exponential moving average (EMA) coefficients. Github repo: //github.com/apple/ml-entropy-reconstruction.
This work proposes novel techniques for the efficient numerical simulation of parameterized, unsteady partial differential equations. Projection-based reduced order models (ROMs) such as the reduced basis method employ a (Petrov-)Galerkin projection onto a linear low-dimensional subspace. In unsteady applications, space-time reduced basis (ST-RB) methods have been developed to achieve a dimension reduction both in space and time, eliminating the computational burden of time marching schemes. However, nonaffine parameterizations dilute any computational speedup achievable by traditional ROMs. Computational efficiency can be recovered by linearizing the nonaffine operators via hyper-reduction, such as the empirical interpolation method in matrix form. In this work, we implement new hyper-reduction techniques explicitly tailored to deal with unsteady problems and embed them in a ST-RB framework. For each of the proposed methods, we develop a posteriori error bounds. We run numerical tests to compare the performance of the proposed ROMs against high-fidelity simulations, in which we combine the finite element method for space discretization on 3D geometries and the Backward Euler time integrator. In particular, we consider a heat equation and an unsteady Stokes equation. The numerical experiments demonstrate the accuracy and computational efficiency our methods retain with respect to the high-fidelity simulations.
Motivated by the recent trend towards the usage of larger receptive fields for more context-aware neural networks in vision applications, we aim to investigate how large these receptive fields really need to be. To facilitate such study, several challenges need to be addressed, most importantly: (i) We need to provide an effective way for models to learn large filters (potentially as large as the input data) without increasing their memory consumption during training or inference, (ii) the study of filter sizes has to be decoupled from other effects such as the network width or number of learnable parameters, and (iii) the employed convolution operation should be a plug-and-play module that can replace any conventional convolution in a Convolutional Neural Network (CNN) and allow for an efficient implementation in current frameworks. To facilitate such models, we propose to learn not spatial but frequency representations of filter weights as neural implicit functions, such that even infinitely large filters can be parameterized by only a few learnable weights. The resulting neural implicit frequency CNNs are the first models to achieve results on par with the state-of-the-art on large image classification benchmarks while executing convolutions solely in the frequency domain and can be employed within any CNN architecture. They allow us to provide an extensive analysis of the learned receptive fields. Interestingly, our analysis shows that, although the proposed networks could learn very large convolution kernels, the learned filters practically translate into well-localized and relatively small convolution kernels in the spatial domain.
In this paper, we present a controller that combines motion generation and control in one loop, to endow robots with reactivity and safety. In particular, we propose a control approach that enables to follow the motion plan of a first order Dynamical System (DS) with a variable stiffness profile, in a closed loop configuration where the controller is always aware of the current robot state. This allows the robot to follow a desired path with an interactive behavior dictated by the desired stiffness. We also present two solutions to enable a robot to follow the desired velocity profile, in a manner similar to trajectory tracking controllers, while maintaining the closed-loop configuration. Additionally, we exploit the concept of energy tanks in order to guarantee the passivity during interactions with the environment, as well as the asymptotic stability in free motion, of our closed-loop system. The developed approach is evaluated extensively in simulation, as well as in real robot experiments, in terms of performance and safety both in free motion and during the execution of physical interaction tasks.
The practical utility of causality in decision-making is widespread and brought about by the intertwining of causal discovery and causal inference. Nevertheless, a notable gap exists in the evaluation of causal discovery methods, where insufficient emphasis is placed on downstream inference. To address this gap, we evaluate seven established baseline causal discovery methods including a newly proposed method based on GFlowNets, on the downstream task of treatment effect estimation. Through the implementation of a distribution-level evaluation, we offer valuable and unique insights into the efficacy of these causal discovery methods for treatment effect estimation, considering both synthetic and real-world scenarios, as well as low-data scenarios. The results of our study demonstrate that some of the algorithms studied are able to effectively capture a wide range of useful and diverse ATE modes, while some tend to learn many low-probability modes which impacts the (unrelaxed) recall and precision.
In the field of cybersecurity, the ability to compute similarity scores at the function level is import. Considering that a single binary file may contain an extensive amount of functions, an effective learning framework must exhibit both high accuracy and efficiency when handling substantial volumes of data. Nonetheless, conventional methods encounter several limitations. Firstly, accurately annotating different pairs of functions with appropriate labels poses a significant challenge, thereby making it difficult to employ supervised learning methods without risk of overtraining on erroneous labels. Secondly, while SOTA models often rely on pre-trained encoders or fine-grained graph comparison techniques, these approaches suffer from drawbacks related to time and memory consumption. Thirdly, the momentum update algorithm utilized in graph-based contrastive learning models can result in information leakage. Surprisingly, none of the existing articles address this issue. This research focuses on addressing the challenges associated with large-scale BCSD. To overcome the aforementioned problems, we propose GraphMoco: a graph momentum contrast model that leverages multimodal structural information for efficient binary function representation learning on a large scale. Our approach employs a CNN-based model and departs from the usage of memory-intensive pre-trained models. We adopt an unsupervised learning strategy that effectively use the intrinsic structural information present in the binary code. Our approach eliminates the need for manual labeling of similar or dissimilar information.Importantly, GraphMoco demonstrates exceptional performance in terms of both efficiency and accuracy when operating on extensive datasets. Our experimental results indicate that our method surpasses the current SOTA approaches in terms of accuracy.
In many complex sequential decision-making tasks, online planning is crucial for high performance. For efficient online planning, Monte Carlo Tree Search (MCTS) employs a principled mechanism for trading off exploration for exploitation. MCTS outperforms comparison methods in many discrete decision-making domains such as Go, Chess, and Shogi. Following, extensions of MCTS to continuous domains have been proposed. However, the inherent high branching factor and the resulting explosion of search tree size are limiting existing methods. To address this problem, we propose Continuous Monte Carlo Graph Search (CMCGS), a novel extension of MCTS to online planning in environments with continuous state and action spaces. CMCGS takes advantage of the insight that, during planning, sharing the same action policy between several states can yield high performance. To implement this idea, at each time step, CMCGS clusters similar states into a limited number of stochastic action bandit nodes, which produce a layered directed graph instead of an MCTS search tree. Experimental evaluation shows that CMCGS outperforms comparable planning methods in several complex continuous DeepMind Control Suite benchmarks and a 2D navigation task with limited sample budgets. Furthermore, CMCGS can be parallelized to scale up and it outperforms the Cross-Entropy Method (CEM) in continuous control with learned dynamics models.
Graph Neural Networks (GNNs) have recently become increasingly popular due to their ability to learn complex systems of relations or interactions arising in a broad spectrum of problems ranging from biology and particle physics to social networks and recommendation systems. Despite the plethora of different models for deep learning on graphs, few approaches have been proposed thus far for dealing with graphs that present some sort of dynamic nature (e.g. evolving features or connectivity over time). In this paper, we present Temporal Graph Networks (TGNs), a generic, efficient framework for deep learning on dynamic graphs represented as sequences of timed events. Thanks to a novel combination of memory modules and graph-based operators, TGNs are able to significantly outperform previous approaches being at the same time more computationally efficient. We furthermore show that several previous models for learning on dynamic graphs can be cast as specific instances of our framework. We perform a detailed ablation study of different components of our framework and devise the best configuration that achieves state-of-the-art performance on several transductive and inductive prediction tasks for dynamic graphs.
Graph Neural Networks (GNNs) have been shown to be effective models for different predictive tasks on graph-structured data. Recent work on their expressive power has focused on isomorphism tasks and countable feature spaces. We extend this theoretical framework to include continuous features - which occur regularly in real-world input domains and within the hidden layers of GNNs - and we demonstrate the requirement for multiple aggregation functions in this context. Accordingly, we propose Principal Neighbourhood Aggregation (PNA), a novel architecture combining multiple aggregators with degree-scalers (which generalize the sum aggregator). Finally, we compare the capacity of different models to capture and exploit the graph structure via a novel benchmark containing multiple tasks taken from classical graph theory, alongside existing benchmarks from real-world domains, all of which demonstrate the strength of our model. With this work, we hope to steer some of the GNN research towards new aggregation methods which we believe are essential in the search for powerful and robust models.
Recently, graph neural networks (GNNs) have revolutionized the field of graph representation learning through effectively learned node embeddings, and achieved state-of-the-art results in tasks such as node classification and link prediction. However, current GNN methods are inherently flat and do not learn hierarchical representations of graphs---a limitation that is especially problematic for the task of graph classification, where the goal is to predict the label associated with an entire graph. Here we propose DiffPool, a differentiable graph pooling module that can generate hierarchical representations of graphs and can be combined with various graph neural network architectures in an end-to-end fashion. DiffPool learns a differentiable soft cluster assignment for nodes at each layer of a deep GNN, mapping nodes to a set of clusters, which then form the coarsened input for the next GNN layer. Our experimental results show that combining existing GNN methods with DiffPool yields an average improvement of 5-10% accuracy on graph classification benchmarks, compared to all existing pooling approaches, achieving a new state-of-the-art on four out of five benchmark data sets.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191