Search results

This paper aims to study the agent learning from expert demonstration data while incorporating reinforcement learning (RL), which enables the agent to break through the limitations of expert demonstration data and reduces the dimensionality of the agent’s exploration space to speed up the training convergence rate.

Firstly, the decay weight function is set in the objective function of the agent’s training to combine both types of methods, and both RL and imitation learning (IL) are considered to guide the agent's behavior when updating the policy. Second, this study designs a coupling utilization method between the demonstration trajectory and the training experience, so that samples from both aspects can be combined during the agent’s learning process, and the utilization rate of the data and the agent’s learning speed can be improved.

The method is superior to other algorithms in terms of convergence speed and decision stability, avoiding training from scratch for reward values, and breaking through the restrictions brought by demonstration data.

The agent can adapt to dynamic scenes through exploration and trial-and-error mechanisms based on the experience of demonstrating trajectories. The demonstration data set used in IL and the experience samples obtained in the process of RL are coupled and used to improve the data utilization efficiency and the generalization ability of the agent.

This study aims to construct a sentiment series generation method for danmu comments based on deep learning, and explore the features of sentiment series after clustering.

This study consisted of two main parts: danmu comment sentiment series generation and clustering. In the first part, the authors proposed a sentiment classification model based on BERT fine-tuning to quantify danmu comment sentiment polarity. To smooth the sentiment series, they used methods, such as comprehensive weights. In the second part, the shaped-based distance (SBD)-K-shape method was used to cluster the actual collected data.

The filtered sentiment series or curves of the microfilms on the Bilibili website could be divided into four major categories. There is an apparently stable time interval for the first three types of sentiment curves, while the fourth type of sentiment curve shows a clear trend of fluctuation in general. In addition, it was found that “disputed points” or “highlights” are likely to appear at the beginning and the climax of films, resulting in significant changes in the sentiment curves. The clustering results show a significant difference in user participation, with the second type prevailing over others.

Their sentiment classification model based on BERT fine-tuning outperformed the traditional sentiment lexicon method, which provides a reference for using deep learning as well as transfer learning for danmu comment sentiment analysis. The BERT fine-tuning–SBD-K-shape algorithm can weaken the effect of non-regular noise and temporal phase shift of danmu text.

This study establishes a method for predicting ground vibrations caused by railway tunnels in unsaturated soils with spatial variability.

First, an improved 2.5D finite-element-method-perfect-matching-layer (FEM-PML) model is proposed. The Galerkin method is used to derive the finite element expression in the ub-pl-pg format for unsaturated soil. Unlike the ub-v-w format, which has nine degrees of freedom per node, the ub-pl-pg format has only five degrees of freedom per node; this significantly enhances the calculation efficiency. The stretching function of the PML is adopted to handle the unlimited boundary domain. Additionally, the 2.5D FEM-PML model couples the tunnel, vehicle and track structures. Next, the spatial variability of the soil parameters is simulated by random fields using the Monte Carlo method. By incorporating random fields of soil parameters into the 2.5D FEM-PML model, the effect of soil spatial variability on ground vibrations is demonstrated using a case study.

The spatial variability of the soil parameters primarily affected the vibration acceleration amplitude but had a minor effect on its spatial distribution and attenuation over time. In addition, ground vibration acceleration was more affected by the spatial variability of the soil bulk modulus of compressibility than by that of saturation.

Using the 2.5D FEM-PML model in the ub-pl-pg format of unsaturated soil enhances the computational efficiency. On this basis, with the random fields established by Monte Carlo simulation, the model can calculate the reliability of soil dynamics, which was rarely considered by previous models.

Why is it that, despite repeated claims that digital-content firms and internet-based businesses can internationalize everywhere almost instantly, many seem unable to profitably expand outside their home markets? Why have emerging market firms (EMNEs) caught up with established developed-country multinationals (DMNEs) so much faster than expected? In this chapter, the author argues that the clue to these two puzzles lies in the realization that, contrary to the dominant view in the international business (IB) literature that focuses only on the intangibles exploited by DMNEs and assumes that these firms are free to unilaterally decide on their mode of entry and operation, doing business in a foreign country is only possible if intangibles are bundled with complementary local resources, usually held by local firms. Taking into account these complementary local resources and their owners makes it clear that DMNEs are not always free to choose their entry mode but must enlist the cooperation of local resource owners. The need of digital-content and internet-based firms for local complementary resources also explains why they sometimes experience problems when expanding abroad. Lastly, control of complementary local resources provides EMNEs with a home advantage against DMNEs competing with them in their home market. The author shows how EMNEs can capitalize on this advantage to obtain the intangibles they lack and need. The fact that these advantages are available on efficient global markets, while complementary local resources are not, explains the surprising speed of EMNE catch-up.

The purpose of this work is to optimize the monitoring of vibrations on dynamometric test rigs for railway brakes. This is a quite demanding application considering the continuous increase of performances of high-speed trains that involve higher testing specifications for brake pads and disks.

In this work, authors propose a mixed approach in which relatively simple finite element models are used to support the optimization of a diagnostic system that is used to monitor vibration levels and rotor-dynamical behavior of the machine. The model is calibrated with experimental data recorded on the same rig that must be identified and monitored. The whole process is optimized to not interfere with normal operations of the rig, using common inertial sensor and tools and are available as standard instrumentation for this kind of applications. So at the end all the calibration activities can be performed normally without interrupting the activities of the rig introducing additional costs due to system unavailability.

Proposed approach was able to identify in a very simple and fast way the vibrational behavior of the investigated rig, also giving precious information concerning the anisotropic behavior of supports and their damping. All these data are quite difficult to be found in technical literature because they are quite sensitive to assembly tolerances and to many other factors. Dynamometric test rigs are an important application widely diffused for both road and rail vehicles. Also proposed procedure can be easily extended and generalized to a wide value of machine with horizontal rotors.

Most of the studies in literature are referred to electrical motors or turbomachines operating with relatively slow transients and constant inertial properties. For investigated machines both these conditions are not verified, making the proposed application quite unusual and original with respect to current application. At the same time, there is a wide variety of special machines that are usually marginally covered by standard testing methodologies to which the proposed approach can be successfully extended.

Because of the various advantages of reinforcement learning (RL) mentioned above, this study uses RL to train unmanned aerial vehicles to perform two tasks: target search and cooperative obstacle avoidance.

This study draws inspiration from the recurrent state-space model and recurrent models (RPM) to propose a simpler yet highly effective model called the unmanned aerial vehicles prediction model (UAVPM). The main objective is to assist in training the UAV representation model with a recurrent neural network, using the soft actor-critic algorithm.

This study proposes a generalized actor-critic framework consisting of three modules: representation, policy and value. This architecture serves as the foundation for training UAVPM. This study proposes the UAVPM, which is designed to aid in training the recurrent representation using the transition model, reward recovery model and observation recovery model. Unlike traditional approaches reliant solely on reward signals, RPM incorporates temporal information. In addition, it allows the inclusion of extra knowledge or information from virtual training environments. This study designs UAV target search and UAV cooperative obstacle avoidance tasks. The algorithm outperforms baselines in these two environments.

It is important to note that UAVPM does not play a role in the inference phase. This means that the representation model and policy remain independent of UAVPM. Consequently, this study can introduce additional “cheating” information from virtual training environments to guide the UAV representation without concerns about its real-world existence. By leveraging historical information more effectively, this study enhances UAVs’ decision-making abilities, thus improving the performance of both tasks at hand.

This study aims to identify an automated machine learning algorithm with high accuracy that sport practitioners can use to identify the specific factors for predicting Major League Baseball (MLB) attendance. Furthermore, by predicting spectators for each league (American League and National League) and division in MLB, the authors will identify the specific factors that increase accuracy, discuss them and provide implications for marketing strategies for academics and practitioners in sport.

This study used six years of daily MLB game data (2014–2019). All data were collected as predictors, such as game performance, weather and unemployment rate. Also, the attendance rate was obtained as an observation variable. The Random Forest, Lasso regression models and XGBoost were used to build the prediction model, and the analysis was conducted using Python 3.7.

The RMSE value was 0.14, and the R2 was 0.62 as a consequence of fine-tuning the tuning parameters of the XGBoost model, which had the best performance in forecasting the attendance rate. The most influential variables in the model are “Rank” of 0.247 and “Day of the week”, “Home team” and “Day/Night game” were shown as influential variables in order. The result was shown that the “Unemployment rate”, as a macroeconomic factor, has a value of 0.06 and weather factors were a total value of 0.147.

This research highlights unemployment rate as a determinant affecting MLB game attendance rates. Beyond contextual elements such as climate, the findings of this study underscore the significance of economic factors, particularly unemployment rates, necessitating further investigation into these factors to gain a more comprehensive understanding of game attendance.

This paper aims to estimate contact location from sparse and high-dimensional soft tactile array sensor data using the tactile image. The authors used three feature extraction methods: handcrafted features, convolutional features and autoencoder features. Subsequently, these features were mapped to contact locations through a contact location regression network. Finally, the network performance was evaluated using spherical fittings of three different radii to further determine the optimal feature extraction method.

This paper aims to estimate contact location from sparse and high-dimensional soft tactile array sensor data using the tactile image.

This research indicates that data collected by probes can be used for contact localization. Introducing a batch normalization layer after the feature extraction stage significantly enhances the model’s generalization performance. Through qualitative and quantitative analyses, the authors conclude that convolutional methods can more accurately estimate contact locations.

The paper provides both qualitative and quantitative analyses of the performance of three contact localization methods across different datasets. To address the challenge of obtaining accurate contact locations in quantitative analysis, an indirect measurement metric is proposed.

An in-depth analysis of how senior managers in a large multinational corporation interpret their social and environmental responsibilities revealed that, notwithstanding formal corporate interpretations, discrepancies persisted in their interpretation of what was expected of them and how to implement it. Two fault lines emerged: (1) an instrumental versus a normative interpretation of corporate societal responsibilities, and (2) a focus on ‘doing less/no harm’ versus ‘doing more good’. This chapter introduces a theoretical framework that combines these fault lines to form four quadrants that each represent a different set of challenges managers face as they commit to improving their organisation’s impact on society. Rather than adjudicate between them, a holistic interpretation of corporate social responsibility (CSR) takes all four types into account. But the four types of challenges differ considerably in nature and thus in the strategic approach that is necessary to deal with them. In this chapter, each quadrant is discussed in detail. What characterises the issues in this quadrant, what mindset, and what strategy are necessary to address them? The chapter concludes with the observation that the framework, and the taxonomy of types of CSR challenges that it brings to the fore, creates greater awareness of how industries are confronted with different sets of challenges and thus need different strategic approaches. A better understanding of these differences may lead to more support, in particular for those managers who work in industries that face a disproportionate share of one particular type of challenges, the ‘nasty trade-offs’.

The purpose of this paper is to investigate the nonsimultaneous impact of three impactors with spherical tip on the response of a low-velocity impact on a beam.

In this research, the third-order shear deformation theory of the beam with hyperbolic shear-strain function is used. Hamilton’s principle is applied to derive the motion equations. To simulate nonsimultaneous impacts, by using the Hertz nonlinear contact law, the contact of the impactors with different times is simulated. Comparisons with other articles are carried out in the one impactor form.

In the parametric study, the histories of the contact force and displacement of the beam are investigated in the presence of only one impactor in the center of the beam and also in the presence of three impactors, one in the center of the beam and the other two around the first impactor with a delay. One of the important and noteworthy points is that the presence of two impactors with a delay causes the maximum contact force and contact time to decrease and the maximum displacement of the beam center to increase.

The original point of this paper is what is the difference between the impact response of one projectile and three nonsimultaneous projectiles on the beam.