Centralised training and decentralised execution (CTDE) is one of the most effective approaches in multiagent reinforcement learning (MARL). However, these CTDE methods still require large amounts of interaction with the environment, even to reach the same performance as very simple heuristic-based algorithms. Although modelbased RL is a prominent approach to improve sample efficiency, its adaptation to a multi-agent setting combining existing CTDE methods has not been well studied in the literature. The few existing studies only consider settings with relaxed restrictions on the number of agents and observable range. In this paper, we consider CTDE settings where some information about each agent’s observations (e.g. each agent’s visibility, number of agents) are changed dynamically. In such a setting, the fundamental challenge is how to train models that accurately generate each agent’s observations with complex transitions in addition to the central state, and how to use it for sample efficient policy learning. We propose a multi-agent model based RL algorithm based on the novel model architecture consisting of global and local prediction models with model mediator. We evaluate our model-based RL approach applied to an existing CTDE method on challenging StarCraft II micromanagement tasks and show that it can learn an effective policy with fewer interactions with the environment.

Pose generation plays an essential role in computer graphics, such as game character design, and 3D modeling. Rather than inverse Kinematics solvers using deterministic heuristic methods suffering from poor diversity, sample-based methods promise to generate a wider variety of poses satisfying the given constraints. In order to obtain generative models from sample data, Generative Adversarial Networks (GANs) are widely used in many problems including pose generation. However, GANs are known to be suffering from mode collapse which causes the generation of specific patterns. Therefore, we propose a novel generative model for pose generation using Implicit Maximum Likelihood Estimation (IMLE), which is a training method for avoiding mode collapse by adaptive sampling of the input-output pairs. The proposed model accepts not only the latent variable, but also the condition of the pose such as a position of the kinematic model’s joint. We trained the proposed model by the IMLE’s optimization method using the dataset consisting of the pair of the pose condition and the corresponding joint angles. In the experiment of a simulated 3-DoF arm simulation, the proposed model successfully avoided mode collapse, thus better diversity rather than the GAN variants while satisfying the given conditional input. Furthermore, we report that the proposed model performs lower prediction error and higher variance than the GAN variants through the experiments on 30-DoF human pose using CMU Mocap Dataset.

In order to design materials efficiently and to avoid quality defects, it is important to understand design rationales, which are usually implicit behind numerical data. This paper proposes an ontological knowledge modeling framework for design rationale of steel materials. The framework integrates tree-like structural models of design actions, manufacturing processes and machining processes with an ontological classification of nodes in these models. Based on the framework, general steel design knowledge models had been described using ontological dictionaries extended for steel materials. Using the knowledge models, a prototype system summarizes crucial points for realizing a given target property and for avoiding quality defects. For deployment of the framework in the steel company, a browser-based client software and a knowledge server system had been developed. The software with a real-scale general design knowledge model for a specific kind of steel had been used by researchers in the company and evaluated through questionnaires. The result shows it is useful for understanding design rationale and for inheritance of steel design knowledge.

Estimating counterfactual outcomes of time-varying treatments is important for medication, vaccination, advertisement and maintenance of equipment. However, it is more difficult than prediction of one-time treatment effect because of lasting treatment effect and time-varying confounders. The current method CRN predicts counterfactual outcomes of time-varying treatments with high accuracy by using LSTM-based Encoder-Decoder to consider lasting treatment effect and learning representations to reduce the effect of time-varying confounders. However, CRN needs a lot of training data and cannot calculate the reliability of the prediction when forecasting counterfactual data. In this study, we introduce a new method Deep CMGP that is a combination of CRN and CMGP. CMGP is a method for estimating one-time treatment effect using multi-task Gaussian process that can be trained on small amounts of training data and calculate the reliability of the prediction. We extend CMGP to estimation of counterfactual outcomes of time-varying treatments by replacing the deep learning part of CRN with deep multi-task Gaussian process. The experimental results show that when the number of training data is limited (Ntrain=1000, 2000), the prediction accuracy is improved compared to baseline methods, and the reliability of the prediction of the counterfactual data is also obtained. In the case of Ntrain=5000, 10000, the accuracy was not as good as those of the baseline methods based on deep learning, but it was found that Deep CMGP can be trained using stochastic variational inference method on a large amount of data (Ntrain=10000), which Gaussian process is not good at.

Cryptocurrencies are highly anonymous, poorly regulated in many countries, and can issue tokens at nearzero cost using existing platforms. As a result, there is no shortage of fraudulent cryptocurrencies that raise large sums of money through hype, then disappear and do little actual project development. The prevalence of fraudulent cryptocurrencies not only harms investors but can also prevent sound companies from raising funds. To remedy this situation, it would be useful to develop a method to determine whether a particular cryptocurrency is fraudulent or not. The information in cryptocurrency whitepapers could be useful in detecting fraudulent cryptocurrency, but there are no clear criteria to evaluate the reliability and feasibility of their content. Besides, most studies analyzing whitepapers focus on the success or failure of ICO ”fundraising” and fail to adequately consider the ongoing development and operation of the project. On the other hand, a few studies have attempted to detect fraudulent cryptocurrencies from whitepapers, but their results suggest the possibility of identifying fraud with high accuracy. The objective of this paper is to build a model to detect fraudulent cryptocurrencies from whitepapers using natural language processing and machine learning techniques, and to verify whether the model has sufficient predictive accuracy in detecting fraud, after solving the problems of previous studies. We collected 250 cryptocurrency whitepapers consisting of 150 frauds and 100 controls, extracted features, and applied multiple machine learning methods to classify frauds and controls. Then analyzed the feature differences between the fraud and control groups, and examined the tendency of fraudulent cryptocurrency whitepapers. We observed 0.841 F1 Score for the best prediction model, which outperforms previous studies. Furthermore, the performance of K-Means, which is unsupervised learning, was not significantly lower than that of other machine learning methods, and a certain level of accuracy was confirmed. Therefore, there is a possibility that K-Means can be used in cases where fraud criteria cannot be clearly defined. We also found that fraudulent cryptocurrency whitepapers used relatively more business and finance-related words. On the other hand, whitepapers in the control group tended to use more blockchain-related technical terms.