scholarly journals Enhanced DQN Framework for Selecting Actions and Updating Replay Memory Considering Massive Non-Executable Actions

2021 ◽  
Vol 11 (23) ◽  
pp. 11162
Author(s):  
Bonwoo Gu ◽  
Yunsick Sung

A Deep-Q-Network (DQN) controls a virtual agent as the level of a player using only screenshots as inputs. Replay memory selects a limited number of experience replays according to an arbitrary batch size and updates them using the associated Q-function. Hence, relatively fewer experience replays of different states are utilized when the number of states is fixed and the state of the randomly selected transitions becomes identical or similar. The DQN may not be applicable in some environments where it is necessary to perform the learning process using more experience replays than is required by the limited batch size. In addition, because it is unknown whether each action can be executed, a problem of an increasing amount of repetitive learning occurs as more non-executable actions are selected. In this study, an enhanced DQN framework is proposed to resolve the batch size problem and reduce the learning time of a DQN in an environment with numerous non-executable actions. In the proposed framework, non-executable actions are filtered to reduce the number of selectable actions to identify the optimal action for the current state. The proposed method was validated in Gomoku, a strategy board game, in which the application of a traditional DQN would be difficult.

2019 ◽  
Vol 34 ◽  
Author(s):  
Mao Li ◽  
Yi Wei ◽  
Daniel Kudenko

Abstract One way to address this low sample efficiency of reinforcement learning (RL) is to employ human expert demonstrations to speed up the RL process (RL from demonstration or RLfD). The research so far has focused on demonstrations from a single expert. However, little attention has been given to the case where demonstrations are collected from multiple experts, whose expertise may vary on different aspects of the task. In such scenarios, it is likely that the demonstrations will contain conflicting advice in many parts of the state space. We propose a two-level Q-learning algorithm, in which the RL agent not only learns the policy of deciding on the optimal action but also learns to select the most trustworthy expert according to the current state. Thus, our approach removes the traditional assumption that demonstrations come from one single source and are mostly conflict-free. We evaluate our technique on three different domains and the results show that the state-of-the-art RLfD baseline fails to converge or performs similarly to conventional Q-learning. In contrast, the performance level of our novel algorithm increases with more experts being involved in the learning process and the proposed approach has the capability to handle demonstration conflicts well.


2016 ◽  
Vol 1 (1) ◽  
pp. 1
Author(s):  
Suraida Suraida

Abstrak Penelitian ini dilakukan karena proses pembelajaran di laboratorium Biologi IAIN STS Jambi yang masih minim sarana prasarana yang ada di laboratorium, sehingga menghambat proses pembelajaran khususnya untuk mata kuliah Morfologi Tumbuhan. Penelitian ini bertujuan mengembangkan buku ajar praktikum dan mengetahui praktikalitasnya. Jenis Penelitian ini adalah penelitian pengembangan (Research and Development) dengan menggunakan 4-D Models yang terdiri dari 4 tahap yaitu Define, Design, Develop dan Disseminate. Karena adanya keterbatasan waktu dan biaya maka tahap disseminate tidak dilakukan. Produk yang dikembangkan berupa buku ajar praktikum yang divalidasi oleh validator. Produk yang telah divalidasi dan dinyatakan valid oleh validator, kemudian diujicobakan pada proses pembelajaran yang bertujuan untuk melihat nilai praktis buku ajar praktikum di laboratorium Biologi. Analisis data yang digunakan adalah data deskriptif untuk memvalidasi perangkat pembelajaran oleh pakar pendidikan. Selain itu juga diteliti data praktikalitas penggunaan perangkat pembelajaran ini yang diperoleh dari observasi dosen dan respon siswa. Nilai validitas produk 83,31% yang dikategorikan valid. Sementara nilai kepraktisan berdasarkan data observasi keterlaksanaan SAP, angket respon dosen dan siswa dikategorikan sangat baik atau sangat praktis. Penelitian menyimpulkan bahwa perangkat pembelajaran di Laboratorium Biologi yang dikembangkan adalah valid dan sangat praktis digunakan baik dosen maupun siswa. Kata Kunci : Pengembangan, buku ajar praktikum, laboratorium biologi Abstract [The development of a course book for plant morphology at biology laboratory] This research was triggered by the limited facilities of the biology laboratory at the State Institute of Islamic Studies Sulthan Thaha Saifuddin Jambi which became a constrain in the teaching and learning process of Plant Morphology classroom sessions. The objective of this research was to develop a course book as well as to reveal its practicality. The researcher did a research and development using 4-D Models consisting of four stages namely; define, design, develop, and disseminate. Considering the limitation of time and finance, the disseminate stage was not executed. The test revealed the validity score of the product was 83,31% which categorized as good. For its practicality, the product was considered as very good based on observation of lesson plan execution and lecturers’ and students’ response. In summary, the course book developed for the course at Biology Laboratory was categorized as valid and practical to be used by both students and lecturers. Keywords: development, a course book, biology laboratory


1999 ◽  
Vol 16 (1) ◽  
Author(s):  
Murad Wilfried Hofmann

This article examines the state of Islamic jurisprudence with regard to many sensitive issues, such as the status of women and minorities in Islam, Islam and Democracy, hudud punishments. The author explores the current state of Islamic discourse on jurisprudence and identifies three approaches-traditional, secular and reformist. The paper explores the positions of the traditional ulama and the reformist muj­tahids on the mentioned topics and finds the reformist position more sensible and closer to the position of ihe Qur'an and Sunnah. This paper while advocating neo-ijtihad, makes an impressive case for the merit???? and Islamic credibility of the reformist jurisprudence.


Energies ◽  
2021 ◽  
Vol 14 (13) ◽  
pp. 3765
Author(s):  
Jarosław Brodny ◽  
Magdalena Tutak ◽  
Peter Bindzár

The global economic development is, to a great extent, dependent on access to large amounts of cheap energy sources. The growing social awareness of ecology and the enormous damage to the Earth’s ecosystem due to the production of energy from conventional sources have forced fundamental changes in the energy sector. Renewable energy is considered to be an opportunity for such changes. The current state of the art allows such changes to be made without restricting economic development. Therefore, activities related to the energy transition are being taken all over the world. The European Union has definitely managed to achieve the most tangible effects in this regard. This article presents the findings of the research aimed at presenting the current state of renewable energy in the European Union and analyzing the changes reported in this sector in the last decade. The research was carried out using a selected set of 11 indicators characterizing renewable energy in individual countries. These indicators were selected on the basis of literature review and own studies of the state of renewable energy and its development prospects. Based on these indicators, changes in the energy structure of individual European Union countries between 2008–2018 were determined. The study is divided into two main stages. The principal components analysis (PCA) was used for the first analysis. In turn, the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) was adopted to assess the level of renewable energy development in the European Union countries. Both these methods and the extended statistical analysis were applied to determine the state of renewable energy development in the European Union countries in the studied period and to divide the Member States into classes with different levels of development. The results of the study showed that the EU countries are characterized by significant differences in the development of RES during the period in question. The unquestionable leaders in this respect are Sweden, Austria, Finland, and Latvia. Based on the findings, it is possible to evaluate the effects of activities related to renewable energy development and to prepare assumptions for future activities. Additionally, both the research and its findings broaden the knowledge of the directions of renewable energy development in individual European Union countries. This is particularly important in the context of changes related to the need to reduce harmful substance emissions and the implementation of the European Green Deal idea.


2021 ◽  
pp. 001458582110225
Author(s):  
Thomas E Peterson

A central question facing the reader of the Paradiso terrestre (Pg 28–33) concerns the selfhood of the protagonist, the character Dante. While the state of Dante’s soul was critical to the poem’s beginning in the dark wood, and remained implicit through the intervening cantos, it is only in the Paradiso terrestre that it becomes the poem’s central focus. This question is explored in cognitive and theological terms in a sequential reading of the six cantos that elucidates the learning process occurring in the character before and after his confession in Pg 31: in his encounter with Matelda, his sensory and perceptual experience of the procession, his dialogues with Beatrice, and his witnessing of her divine beauty as the analogia entis reflecting the beauty of God. The analysis acknowledges the changes in Dante’s style in this interval, which serves as a fulcrum of the entire Commedia, a spatio-temporal threshold in which the transition of one soul, from confession to redemption to instruction on the divine word, is linked to the destiny of humankind and the prospect of universal salvation. Throughout this process of becoming, the character’s cognitive limitations are exposed, not simply as flaws but as signs of his intrinsic humanity.


2020 ◽  
pp. 48-59
Author(s):  
A.M. Agapkin

The state of the problem of processing agricultural waste as a newly forming industry for the disposal of such waste in conjunction with the development of the emerging industry of organic production and the market of organic fertilizers is considered. The issue is regarded in the interrelation of the regulatory, economic and technological components in their dynamics from the current state to the target (desired).


Legal Concept ◽  
2020 ◽  
pp. 13-20
Author(s):  
Svyatoslav Biryukov ◽  
Mikhail Bobovkin ◽  
Mikhail Shmatov

Introduction: the Constitution of the Russian Federation and other Federal laws in this country guarantee the protection of the population against crimes, including criminal attacks of extremist orientation. However, recently there has been a steady trend towards an increase in the number of committed crimes of extremist orientation, which determines the need to improve the quality of protection of individual rights, and along with them, the constitutional framework of the state, since demonstratively committed extremist crimes cause a great public response and contribute to the undermining of state power. The crime statistics show a significant increase in the number of extremist crimes; there is a natural tendency to spread the ideas of extremism among the population. Unfortunately, only some of the extremist crimes are counted as such in the official statistics. The crimes of this category are often registered without taking into account the qualifying feature – the motive of national, racial, religious hatred or enmity, and, as a result, are not considered in the group of crimes of extremism. Another reason for not fully accounting for these crimes is their latency: not all victims of such criminal actions declare this for various objective and subjective reasons. The public danger of crimes of the group in question is due, on the one hand, as usual, to their group character, and on the other hand, such illegal actions incite interethnic and other hatred, which is very harmful in the context of the efforts being made to build a civil society. Currently, the legislative bodies do not clearly pay enough attention to the organization of counteraction to extremism as an anti-social phenomenon. For example, over the past ten years, the problems of countering extremism have been resolved through the adoption of only four normative legal acts of a national nature. In this regard, the authors aim to give a general description of such a phenomenon as extremism and the state of the fight against such crimes. Methods: the methodological framework for this research is a set of methods of scientific knowledge, among which the main ones are the methods of information processing and logical analysis, synthesis, induction, deduction and generalization. Results: the authors’ content of the general characteristics of extremism and analysis of the current state of the fight against crimes of extremist orientation actualizes the problem of the need to improve the state of the theoretical base, prepare recommendations based on it, which would contribute to improving the efficiency of the state authorized bodies in the fight against various manifestations of extremism, and primarily in order to solve and investigate crimes of extremist orientation. Conclusions: the study has given the general characteristics of extremism and the analysis of the current state of the fight against extremist crimes in order to inform law students, and the teaching staff of law schools and practitioners to better understand the characteristics and dangers of this phenomenon.


2020 ◽  
pp. 10-14
Author(s):  
Kseniia KOVTUNENKO ◽  
Kateryna BONDARENKO

The purpose of the paper is to improve institutions and legislation on public finance management in Ukraine, to define the concept of "financial control", to consider the process of development and formation of financial control, to highlight the financial control`s features, to justify the need for long-term financial control. Financial control and financial management in the enterprise are very important for every state. The current state of economic improvement in Ukraine should increase the role of the state in regulating the economy in order to identify differences between the law, recognizing the standards and functioning of financial control. At the present stage in Ukraine, the issue of establishing unification of public finance management remains unresolved: there is a discrepancy in state legislation. The Ukrainian economy`s growth leads to an additional need for financial control in a moment. Thus, its role and importance in securing the assets of different types of members’ organizations, the equipment`s efficiency, labour and finance. The paper is devoted to the key issues of finding ways to ensure financial control in the organization and the regulatory framework in this area. The paper presents an overview of financial control by state (local) resource management and their use, as well as financial control of public administration by the state external financial regulator (audit), the Ministry of Finance on behalf of the Verkhovna Rada of Ukraine, and public finance management, including internal management and internal audit, which are provided by current legislation. In the paper was presented a study of the financial control`s concept. The main types of financial control, its goals and objectives are researched. The author emphasizes the importance of the organization’s internal financial control and the key stages of its development. The types of external control and features of their application were also researched. In conclusion the Ukrainian financial management`s current state is researched.


2021 ◽  
pp. 113-114
Author(s):  
Melise Maia Ribeiro

The objective of this research is to know new decisions about the teaching and learning process in the context of the pandemic in the state of Amazonas, Brazil. The pandemic suspended classroom classes at more than 200 schools, causing the reorganization of pedagogical practices in distance education. The result was the applicability of the Special Regime of Non-Attendance Classes adopted by the Government of Amazonas (Aula em Casa Project). It is concluded that new directions can be taken from formal education in view of this new reality.


2021 ◽  
Vol 35 ◽  
pp. 00006
Author(s):  
Ainur Biembetov ◽  
Nur Yanybayev ◽  
Ilnar Valiev

Environmental monitoring of specially protected natural reservations in Russia makes it necessary to analyze periodically the parameters of natural reservations to identify the state of components of nature. The Bashkir Nature Reserve is located in the Southern Urals. The availability of materials on forest management in 1956, 1969, 1979, and 2016 is one of the special features of the scientific fund of the Bashkir Nature Reserve. The analysis of these materials showed stable positive dynamics of the development of coniferous and small-leaved deciduous forestry and its current state.


Sign in / Sign up

Export Citation Format

Share Document