Digital Library[ Search Result ]
Risk Scheduling-based Optimistic Exploration for Distributional Reinforcement Learning
Jihwan Oh, Joonkee Kim, Se-Young Yun
http://doi.org/10.5626/JOK.2023.50.2.172
Distributional reinforcement learning demonstrates state-of-the-art performance in continuous and discrete control systems with the features of variance and risk, which can be used to explore action space. However, the exploration method employing the risk property is hard to find, although numerous exploration methods in Distributional RL employ the variance of the return distribution for an action. This paper presents risk scheduling approaches that explore risk levels and optimistic behaviors from a risk perspective in Distributional reinforcement learning. We demonstrate the performance (win-rate) enhancement of the DMIX, DDN, and DIQL algorithms, which integrate Distributional reinforcement learning into a multi-agent system using risk scheduling in a multi-agent setting with comprehensive experiments.
Distribution of Malicious Apps Considering App Categories and Development Tools in Major Android Markets
Jihwan Oh, Myeonggeon Lee, SeongJe Cho, Sangchul Han
http://doi.org/10.5626/JOK.2019.46.2.109
According to recent cyber security analysis reports, there are numerous malicious apps available in online markets. In this paper, we analyzed the portion of malicious apps by market, main category, and cross-platform development tools for apps distributed on Android"s official market (Google Play) and a third-party market (Amazon Appstore). The apps were collected from the 13 main categories of the markets and examined using the VirusTotal service. We classified them into benign app, malware and potentially-unwanted applications (PUA). The percentage of each category and development tool used was then quantified. The distribution of malicious apps created with primary cross-platform development tools was also measured. Out of the total 22,615 apps collected, 4,741 of them were found to be malicious apps. The percentage of malicious apps was found to be 14.39% and 24.85% in Google play and Amazon Appstore respectively. The categories with the highest percentage of malicious apps were Utilities (19.8%) and Weather (19.1%) in Google Play, and Social (40.2%), Travel&Local (36.3%) and Weather (34.9%) in Amazon Appstore. Caution should be exercised when users install apps from these categories. Additionally, the percentage of malicious apps written using cross-platform development tools was 17.8%, a dramatic increase in comparison to previous statistics.
Search
Journal of KIISE
- ISSN : 2383-630X(Print)
- ISSN : 2383-6296(Electronic)
- KCI Accredited Journal
Editorial Office
- Tel. +82-2-588-9240
- Fax. +82-2-521-1352
- E-mail. chwoo@kiise.or.kr