CAINE 2022:Papers with Abstracts

Abstract. As the number of video streaming platforms is growing, the risk factor associated with illegal and inappropriate content streaming is increasing exponentially. Therefore, mon- itoring such content is essential. Many researches have been conducted on classifying encrypted videos. However, most existing techniques only pass raw traffic data into clas- sification models, which is an ineffective way of training a model. This research proposes a bucket-based data pre-processing technique for a video identification in network traffic. The bucketed traffic is then incorporated with a fine-tuned word2vec-based neural net- work to produce an effective encrypted video classifier. Experiments are carried out with different numbers and sizes of buckets to determine the best configuration. Furthermore, previous research has overlooked the phenomenon of concept drift, which reduces the effec- tiveness of a model. This paper also compares the severity of concept drift on the proposed and previous technique. The results indicate that the model can predict new samples of videos with an overall accuracy of 81% even after 20 days of training.
Abstract. Facing the current situation with the correct tools has made the difference between companies that have managed to adapt to current technological and social changes and those that are in serious trouble. Today more than ever it is evident that the digital transformation of companies is the best way to follow and these transformative processes have been streamlined. There has been talk for several years about the need for companies to go through a digital transformation. Today more than ever it is necessary to face this situation of global uncertainty, having the necessary tools so that businesses can adapt quickly to the new reality. Thus, a trend that cannot be ignored in any way is the rapid growth of ecommerce. Therefore, for any organization, it will be of great help to have tools that allow evaluating the pros and cons of carrying out this digital transformation process. Therefore, in this paper we present a strategy for measuring business processes in their digital transformation process.
Abstract. A graph database system is a type of NoSQL databases that is based on the graph data model using nodes and arcs. Existing graph databases are passive and can only handle crisp data. Information in the real world can be imprecise and vague rather than crisp. Integrating fuzzy logic into database systems allows users to use uncertain data, which presents the degree to which something is true. Compared with passive databases, active databases support event handling by monitoring and reacting to specific circumstances automatically. This paper describes our approach of incorporating fuzzy concepts and active rules into a graph database system. Our recent publication has described our work of temporal event processing. This paper focuses on mutation events handling and active rule processing in a fuzzy system, covering the language model, the execution model, and architecture design. The language model defines the rule structure and contains the metadata for rule processing. Architecture design identifies the system's architectural components and user interfaces including rule specification interface and query interface. The execution model handles rule processing and execution at run time. A supply chain application is used to demonstrate the examples of active rule specification and execution.
Abstract. This paper presents a prototype for an information exchange system, which allows information exchange between companies without actually sharing data. First, the need for such an intercompany exchange platform is explained and the value for supply chains resulting from such a platform is described. A literature review presents the existing concepts and techniques contributing to the development of an architecture. Finally. The information exchange concept and the prototype implementation are explained in detail.*
Abstract. This paper discussed how to build deep reinforcement learning (DRL) agents to determine the allocation of money for assets in a portfolio so that the maximum return can be gained. The policy gradient method from reinforcement learning and convolutional neural network/recurrent neural network/convolutional neural network concatenated with the recurrent neural network from deep learning are combined together to build the agents. With the proposed models, three types of portfolios are tested: stocks portfolio which has a positive influence due to the Covid-19, stocks portfolio which has a negative influence due to the Covid-19, and portfolio of stocks combined with cryptocurrency which are randomly selected. The performance of our DRL agents was compared with that of equal-weighted agent and all the money fully invested on one stock agents. All of our DRL agents showed the best performance on the randomly selected portfolio, which has an overall stable up-ticking trend. In addition, the performance of linear regression model was also tested with the random selected portfolio, and it shows a poor result compared to other agents.
Abstract. In this paper, we show the effectiveness of pipeline implementations of Dynamic Pro- gramming (DP) on Graphics Processing Unit (GPU). We deal with a simplified DP problem where each element of its solution table is calculated in order by semi-group operations among several of already computed elements in the table. We implement the DP program on GPU in a pipeline fashion, i.e., we use GPU cores for supporting pipeline-stages so that several elements of the solution tables are partially computed at one time. Further, to accelerate the pipeline implementation, we propose a p-fold pipeline technique, which enables larger parallelism more than the number of pipeline-stages.
Abstract. Because peer-to-peer networks are inherently insecure, they provide a special chal- lenge in terms of network security. In this study, we have considered a recently described non-DHT-based 2-level structured P2P network. It is an architecture built on interests. Utilizing RC, a modular arithmetic-based residue class, the overlay topology has been achieved. This design was chosen because it allows for minimal latency in both intra and inter-group communications. In the present study, we provide efficient schemes for public- key cryptographic security of existing communication protocols. We have also extended these approaches to include anonymity.
Abstract. The introspection of virtual machines is an important aspect of protecting against the threat of malware that can hide from traditional automated malware-detection systems. A distributed system for live virtual machine introspection is presented utilizing the Xen Project hypervisor and LibVMI for introspection. The system incorporates the importing of VMs through the OVF specification, VM management through libvirt, and the streaming of various kernel data structures and system calls into data stores with no delay between introspection and storage.
Abstract. This study considers the convex formation of polygons via a two-phase procedure. The approach infuses features from the behavioral and virtual structure methodologies. In the first phase, the agents form a circle. In the second phase, the agents are reconfigured into a polygon formation. Since the reconfiguration of virtual structures are often faced with challenges, the circle formation is adopted as the regrouping feature of agents before reconfiguration into a different polygon formation. No distinction is made among the agents, which simplifies formations. In addition, the agents can avoid collision during the formation process. Simulation results show precise formation of agents into different polygons. The results further indicate that the proposed approach has the potential to maintain the formation whilst the formation is rotating or changes location.