A. Open-Source Intelligence (OSINT)
Open-source intelligence (OSINT) is a main methodology in this proposal for collecting, analyzing and making decisions. OSINT is nowadays used in national security and law enforcement to prevent and detect unwanted crimes. In this report, we demonstrate three proposals to detect and monitor concerned or illegal trades of dual-used goods using OSINT.
B. System Proposals
Proposal 1. Detecting the export of dual-use goods using search word and images
Figure 1. Strategic Good Search System (Proposal 1)
This system allows users to collect, analyze and monitor data related to dual-use goods transactionson the web through search term and image extraction. The Search-Pro provided by KOSTI gives us a list of controlled items for specific dual-use goods. From the list, we extract keywords and images to help specify the corresponding items.
By establishing a search term-based system, related web documents are crawled by combining keywords extracted above. After extracting the keywards, we collect and analyze desired information from well-known search engines like Google, Naver, and Zum. Image-based systems utilize collected images through dual-use item search systems or other sources. Image files can be transformed to amplify multiple images to build the flexible searching system. Also, the information for monitoring is presented through the visualization of various features such as categories and keyword types. It has a control function on the concerned transaction and becomes a reference to the performance of follow-up research. This helps establish counter-measures for illegal transactions through visual data. The system provides a function to periodically update web documentsand produce auto-reports containing analysed information.
Proposal 2. A monitoring system to list up the companies dealing with strategic goods and technologies
We propose to develop a system that automatically lists up entities dealing with strategic goods and technologies. This system is important and necessary to build due to the following reasons. First, the businesses trader may not be aware that the components that make up the product contain the dual-used items. Second, even if they have not previously handled strategic goods, businesses belonging to the same or similar industries are likely to handle strategic goods when expanding their investment areas. Therefore, we can reduce the risk of future occurrence by keeping a list of companies that may have the potential problems through the analysis of websites of companies.
Proposal 3. Detecting Illegal transaction with money laundering on the dark web
Figure 2. Frame work strategic illegal transaction detection systems
This system detects illegal transactions and tracks down the participants on the deals. This only works when transactions are completed using cryptocurrency. This is how to establish the system as follows. This system on the dark web is connected to proposal 1.
First, using the dark web search engine developed in first proposal, individuals and businesses functions handling strategic goods on the deep web are monitored. The monitored results are databaseized as “Linked companies”, which are updated in real time whenever users look up. Second, the system accesses the sites that belong to the database created in second proposal and crawls the seller's cryptocurrency address. Then it tracks information and transaction event occurrences for virtual asset handling establishments using dusting attack techniques with crawled addresses. Next, finger-printing technique is performed using deep learning-based convolutional neural networks (CNNs) to obtain location information of participants. After all, it tracts the IDs and addresses of transaction participants from the transaction information obtained through second process. Then, it updates the blacklist database in real time. Lastly, users in black list and transactions are likely to recur, so monitor them continuously.
