08-13, 15:30–16:10 (Asia/Yerevan), 214W PAB
Cyberattacks are continuously growing in volume and entanglement. They
target organizations' systems, networks, and private data, causing financial loss,
customer loss, and data leakage. As technology improves nowadays, Artificial
Intelligence (AI) based solutions help boost Cybersecurity. Attend this talk to
discover how AI-powered algorithms are used to stay ahead of Cyberattacks such as
Phishing, Lookalike domains, or Name Spoofing.
During the last years, EasyDmarc gathered massive experience in the
domain of AI-powered Cybersecurity (CS) and is happy to share it with you for the
first time. This will be an informative talk for the general audience interested in
learning the best practices in CS and state-of-the-art Artificial Intelligence (AI)
solutions. One of the core takeaway messages is that not only AI but also data
quality must be addressed when developing AI solutions. Current trends in CS will be
discussed following the introduction of AI in different kinds of Cyberattacks. Namely,
any company is a target for attacks like Email Display Name Spoofing, Domain
Lookalike attack, and Phishing. At EasyDmarc, we use Machine Learning(ML) based
techniques and construct sophisticated algorithms to address those attacks.
The key to successful AI algorithm performance is the data. AI algorithms use the
data to find hidden patterns and characteristics of Cyberattacks, so thoroughly
processed data should be present. An important property of the algorithm is the
ability to improve its performance and continuously adjust to the new environment
via receiving up-to-date data from Cyberattacks.
There are many examples when different research groups train AI algorithms on
available public data and report more than 95% precision score in the case of
balanced datasets. Those examples, however, are limited to the train data
distribution and show pure performance in production. One such example is phishing
URL detection. In our talk, we will present a number of AI projects at EasyDmarc that
boost CS, as well as focus on URL phishing detection. You will learn the main
lessons when handling such complex real-world problems.
No previous knowledge expected
Elina Israelyan is a Data Scientist at EasyDMARC, where they develop AI solutions for email security and deliverability. She also studies at the American University of Armenia, majoring in Data Science. Her experience is mainly related to Natural Language Processing and Anomaly Detection. Elina Israelyan has been working on various projects such as URL and Website Phishing Detection, Email traffic Anomaly Detection, Spam Detection, etc., where she has developed end-to-end projects starting from the raw data preprocessing, exploratory data analysis, and finishing with the model deployment processes.