Optimization in Machine Learning:a Distribution-Space Approach

We present the viewpoint that optimization problems encountered in machine learning can often be interpreted as minimizing a convex functional over a function space,but with a non-convex constraint set introduced by model parameterization.This observation allows us to repose such problems via a suitable relaxation as convex optimization problems in the space of distributions over the training parameters.We derive some simple relationships between the distribution-space problem and the original problem,e.g.,a distribution-space solution is at least as good as a solution in the original space.Moreover,we develop a numerical algorithm based on mixture distributions to perform approximate optimization directly in the distribution space.Consistency of this approximation is established and the numerical efficacy of the proposed algorithm is illustrated in simple examples.In both theory and practice,this formulation provides an alternative approach to large-scale optimization in machine learning.

Convergence analysis for complementary-label learning with kernel ridge regression

Complementary-label learning(CLL)aims at finding a classifier via samples with complementary labels.Such data is considered to contain less information than ordinary-label samples.The transition matrix between the true label and the complementary label,and some loss functions have been developed to handle this problem.In this paper,we show that CLL can be transformed into ordinary classification under some mild conditions,which indicates that the complementary labels can supply enough information in most cases.As an example,an extensive misclassification error analysis was performed for the Kernel Ridge Regression(KRR)method applied to multiple complementary-label learning(MCLL),which demonstrates its superior performance compared to existing approaches.

Recent advances in protein conformation sampling by combining machine learning with molecular simulation

The rapid advancement and broad application of machine learning(ML)have driven a groundbreaking revolution in computational biology.One of the most cutting-edge and important applications of ML is its integration with molecular simulations to improve the sampling efficiency of the vast conformational space of large biomolecules.This review focuses on recent studies that utilize ML-based techniques in the exploration of protein conformational landscape.We first highlight the recent development of ML-aided enhanced sampling methods,including heuristic algorithms and neural networks that are designed to refine the selection of reaction coordinates for the construction of bias potential,or facilitate the exploration of the unsampled region of the energy landscape.Further,we review the development of autoencoder based methods that combine molecular simulations and deep learning to expand the search for protein conformations.Lastly,we discuss the cutting-edge methodologies for the one-shot generation of protein conformations with precise Boltzmann weights.Collectively,this review demonstrates the promising potential of machine learning in revolutionizing our insight into the complex conformational ensembles of proteins.

Offline Reinforcement Learning with Constrained Hybrid Action Implicit Representation Towards Wargaming Decision-Making

Reinforcement Learning(RL)has emerged as a promising data-driven solution for wargaming decision-making.However,two domain challenges still exist:(1)dealing with discrete-continuous hybrid wargaming control and(2)accelerating RL deployment with rich offline data.Existing RL methods fail to handle these two issues simultaneously,thereby we propose a novel offline RL method targeting hybrid action space.A new constrained action representation technique is developed to build a bidirectional mapping between the original hybrid action space and a latent space in a semantically consistent way.This allows learning a continuous latent policy with offline RL with better exploration feasibility and scalability and reconstructing it back to a needed hybrid policy.Critically,a novel offline RL optimization objective with adaptively adjusted constraints is designed to balance the alleviation and generalization of out-of-distribution actions.Our method demonstrates superior performance and generality across different tasks,particularly in typical realistic wargaming scenarios.

Sim-to-Real: A Performance Comparison of PPO, TD3, and SAC Reinforcement Learning Algorithms for Quadruped Walking Gait Generation

The performance of the state-of-the-art Deep Reinforcement algorithms such as Proximal Policy Optimization, Twin Delayed Deep Deterministic Policy Gradient, and Soft Actor-Critic for generating a quadruped walking gait in a virtual environment was presented in previous research work titled “A Comparison of PPO, TD3, and SAC Reinforcement Algorithms for Quadruped Walking Gait Generation”. We demonstrated that the Soft Actor-Critic Reinforcement algorithm had the best performance generating the walking gait for a quadruped in certain instances of sensor configurations in the virtual environment. In this work, we present the performance analysis of the state-of-the-art Deep Reinforcement algorithms above for quadruped walking gait generation in a physical environment. The performance is determined in the physical environment by transfer learning augmented by real-time reinforcement learning for gait generation on a physical quadruped. The performance is analyzed on a quadruped equipped with a range of sensors such as position tracking using a stereo camera, contact sensing of each of the robot legs through force resistive sensors, and proprioceptive information of the robot body and legs using nine inertial measurement units. The performance comparison is presented using the metrics associated with the walking gait: average forward velocity (m/s), average forward velocity variance, average lateral velocity (m/s), average lateral velocity variance, and quaternion root mean square deviation. The strengths and weaknesses of each algorithm for the given task on the physical quadruped are discussed.

Design space exploration of neural network accelerator based on transfer learning

With the increasing demand of computational power in artificial intelligence(AI)algorithms,dedicated accelerators have become a necessity.However,the complexity of hardware architectures,vast design search space,and complex tasks of accelerators have posed significant challenges.Tra-ditional search methods can become prohibitively slow if the search space continues to be expanded.A design space exploration(DSE)method is proposed based on transfer learning,which reduces the time for repeated training and uses multi-task models for different tasks on the same processor.The proposed method accurately predicts the latency and energy consumption associated with neural net-work accelerator design parameters,enabling faster identification of optimal outcomes compared with traditional methods.And compared with other DSE methods by using multilayer perceptron(MLP),the required training time is shorter.Comparative experiments with other methods demonstrate that the proposed method improves the efficiency of DSE without compromising the accuracy of the re-sults.

Lotus LearningSpace在高等特殊教育教学领域的应用

Lotus LearningSpace是Lotus在知识管理策略基础上推出的最新教学解决方案,旨在提供一个可协作、可按计划、辅助指导、分布式的虚拟学习环境。Lotus LearningSpace简体中文版,为国内有关高等特殊教育的研究机构提供了一个很好的参考。斯塔福德郡大学的实验表明,相关软件开发必须要了解各类残疾人群的特点,考虑残疾学生的主要缺陷,开发专门适合残疾学生的、适用领域广泛的虚拟学习系统。

应用Learning Space4实现生理学的远程教育

实现远程教育的关键是有机地组织各类教育资源和高效率的双向通信。 L earning Space4是可以有效地解决高效率有机组织教育资源、跟踪、评估学生的学习状况、非实时和实时教学等关键问题的一个优秀的网络远程教学和管理平台系统。介绍应用 L earning Space4创建远程教育教程的基本方法 ,并以《生理学》第四版为例介绍应用 L

Combination of effective color information and machine learning for rapid prediction of soil water content

Soil water content(SWC)is one of the critical indicators in various fields such as geotechnical engineering and agriculture.To avoid the time-consuming,destructive,and laborious drawbacks of conventional SWC measurements,the image-based SWC prediction is considered based on recent advances in quantitative soil color analysis.In this study,a promising method based on the Gaussian-fitting gray histogram is proposed for extracting characteristic parameters by analyzing soil images,aiming to alleviate the interference of complex surface conditions with color information extraction.In addition,an identity matrix consisting of 32 characteristic parameters from eight color spaces is constituted to describe the multi-dimensional information of the soil images.Meanwhile,a subset of 10 parameters is identified through three variable analytical methods.Then,four machine learning models for SWC prediction based on partial least squares regression(PLSR),random forest(RF),support vector machines regression(SVMR),and Gaussian process regression(GPR),are established using 32 and 10 characteristic parameters,and their performance is compared.The results show that the characteristic parameters obtained by Gaussian-fitting can effectively reduce the interference from soil surface conditions.The RGB,CIEXYZ,and CIELCH color spaces and lightness parameters,as the inputs,are more suitable for the SWC prediction models.Furthermore,it is found that 10 parameters could also serve as optimal and generalizable predictors without considerably reducing prediction accuracy,and the GPR model has the best prediction performance(R^(2)≥0.95,RMSE≤2.01%,RPD≥4.95,and RPIQ≥6.37).The proposed image-based SWC predictive models combined with effective color information and machine learning can achieve a transient and highly precise SWC prediction,providing valuable insights for mapping soil moisture fields.

Automatic Generation of Artificial Space Weather Forecast Product Based on Sequence-to-sequence Model

Both analyzing a large amount of space weather observed data and alleviating personal experience bias are significant challenges in generating artificial space weather forecast products.With the use of natural language generation methods based on the sequence-to-sequence model,space weather forecast texts can be automatically generated.To conduct our generation tasks at a fine-grained level,a taxonomy of space weather phenomena based on descriptions is presented.Then,our MDH(Multi-Domain Hybrid)model is proposed for generating space weather summaries in two stages.This model is composed of three sequence-to-sequence-based deep neural network sub-models(one Bidirectional Auto-Regressive Transformers pre-trained model and two Transformer models).Then,to evaluate how well MDH performs,quality evaluation metrics based on two prevalent automatic metrics and our innovative human metric are presented.The comprehensive scores of the three summaries generating tasks on testing datasets are 70.87,93.50,and 92.69,respectively.The results suggest that MDH can generate space weather summaries with high accuracy and coherence,as well as suitable length,which can assist forecasters in generating high-quality space weather forecast products,despite the data being starved.

Development of New Capabilities Using Machine Learning for Space Weather Prediction

With the development of space exploration and space environment measurements,the numerous observations of solar,solar wind,and near Earth space environment have been obtained in last 20 years.The accumulation of multiple data makes it possible to better use machine learning technique,which has achieved unforeseen results in industrial applications in last decades,for developing new approaches and models in space weather investigation and prediction.In this paper,the efforts on the forecasting methods for space weather indices,events,and parameters using machine learning are briefly introduced based on the study works in recent years.These investigations indicate that machine learning,especially deep learning technique can be used in automatic characteristic identification,solar eruption prediction,space weather forecasting for solar and geomagnetic indices,and modeling of space environment parameters.

Reinforcement learning method for machining deformation control based on meta-invariant feature space

Precise control of machining deformation is crucial for improving the manufacturing quality of structural aerospace components.In the machining process,different batches of blanks have different residual stress distributions,which pose a significant challenge to machining deformation control.In this study,a reinforcement learning method for machining deformation control based on a meta-invariant feature space was developed.The proposed method uses a reinforcement-learning model to dynamically control the machining process by monitoring the deformation force.Moreover,combined with a meta-invariant feature space,the proposed method learns the internal relationship of the deformation control approaches under different stress distributions to achieve the machining deformation control of different batches of blanks.Finally,the experimental results show that the proposed method achieves better deformation control than the two existing benchmarking methods.

Day-ahead scheduling based on reinforcement learning with hybrid action space

Driven by the improvement of the smart grid,the active distribution network(ADN)has attracted much attention due to its characteristic of active management.By making full use of electricity price signals for optimal scheduling,the total cost of the ADN can be reduced.However,the optimal dayahead scheduling problem is challenging since the future electricity price is unknown.Moreover,in ADN,some schedulable variables are continuous while some schedulable variables are discrete,which increases the difficulty of determining the optimal scheduling scheme.In this paper,the day-ahead scheduling problem of the ADN is formulated as a Markov decision process(MDP)with continuous-discrete hybrid action space.Then,an algorithm based on multi-agent hybrid reinforcement learning(HRL)is proposed to obtain the optimal scheduling scheme.The proposed algorithm adopts the structure of centralized training and decentralized execution,and different methods are applied to determine the selection policy of continuous scheduling variables and discrete scheduling variables.The simulation experiment results demonstrate the effectiveness of the algorithm.

Research on Learning Space Management Transformation

Learning space management transformation is an inevitable guidance for learners’increasingly abundant learning needs and technological innovation.Learning space management should be transformed for students,teachers,and schools to form a new pattern that centers on learners,which is led by professional teachers,and breaks the inherent shape of schools.The development of learning space management transformation needs top level design from top to bottom and basic level exploration from bottom to top,meantime combining the overall construction with key breakthroughs.The learning space sharing mechanism proposed in this research will provide references for the learning space management transformation.

L2 Teaching and Learning in the Classroom Space

L2 teaching and learning is a way of using language,but it happens in a particular space—the classroom space,which,to some extent,has a restriction to language using.This paper provides a valuable sight into L2 teaching and learning in the classroom space,and discusses the viewpoint of how to make an actual learning of L2 under the way of teaching.

Explanatory Multi-Scale Adversarial Semantic Embedding Space Learning for Zero-Shot Recognition

The goal of zero-shot recognition is to classify classes it has never seen before, which needs to build a bridge between seen and unseen classes through semantic embedding space. Therefore, semantic embedding space learning plays an important role in zero-shot recognition. Among existing works, semantic embedding space is mainly taken by user-defined attribute vectors. However, the discriminative information included in the user-defined attribute vector is limited. In this paper, we propose to learn an extra latent attribute space automatically to produce a more generalized and discriminative semantic embedded space. To prevent the bias problem, both user-defined attribute vector and latent attribute space are optimized by adversarial learning with auto-encoders. We also propose to reconstruct semantic patterns produced by explanatory graphs, which can make semantic embedding space more sensitive to usefully semantic information and less sensitive to useless information. The proposed method is evaluated on the AwA2 and CUB dataset. These results show that our proposed method achieves superior performance.

E-learning学习共享空间构建分析

介绍了E-Learning、虚拟学习共享空间内容,并从虚拟资源体系、虚拟社交环境、虚拟服务的建设以及网络技术应用几个方面阐述了如何构建E-learning学习共享空间。

基于Cite Space的我国护理专业案例教学研究可视化分析

目的探讨我国护理专业案例教学研究现状与发展趋势,为护理专业开展案例教学提供借鉴与指导。方法以万方和中国知网为数据源,运用Cite Space软件可视化分析1993年1月—2022年3月护理专业案例教学研究现状及其趋势。结果共提取到2878篇有效文献,发文量处于快速上升阶段,《卫生职业教育》载文量最多(215篇),尚未形成核心期刊,基金支持占10.67%,形成了以魏志名、吴伟、伊晓峰和郭大英为代表的高产作者群体,核心作者群和机构群尚未形成,研究热点主要集中在教学方法、问题式学习(PBL)、护理带教、教学效果、应用效果、情景模拟、临床带教和教学改革,高频关键词形成了11个聚类,研究前沿主要集中在教学方法、PBL、护理带教、教学效果、应用效果、情景模拟、临床带教和教学改革。结论近年来护理专业案例教学受到了护理教育者的重视,核心作者和核心机构群尚未形成,研究质量有待进一步提升。

Structure learning on Bayesian networks by finding the optimal ordering with and without priors

Ordering based search methods have advantages over graph based search methods for structure learning of Bayesian networks in terms on the efficiency. With the aim of further increasing the accuracy of ordering based search methods, we first propose to increase the search space, which can facilitate escaping from the local optima. We present our search operators with majorizations, which are easy to implement. Experiments show that the proposed algorithm can obtain significantly more accurate results. With regard to the problem of the decrease on efficiency due to the increase of the search space, we then propose to add path priors as constraints into the swap process. We analyze the coefficient which may influence the performance of the proposed algorithm, the experiments show that the constraints can enhance the efficiency greatly, while has little effect on the accuracy. The final experiments show that, compared to other competitive methods, the proposed algorithm can find better solutions while holding high efficiency at the same time on both synthetic and real data sets.

Feature Extraction of Kernel Regress Reconstruction for Fault Diagnosis Based on Self-organizing Manifold Learning

The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.