Machine Learning systems for fast adaptive brands logos recognition in the sport - by Andrei Boiarov
Engagement with live audience in sports is considered as one of the most powerful advertisement strategy for brands. Sports fans are loyal, the audience is huge, and now it gets even bigger with the growth of streaming platforms. However, evaluating the effectiveness of brand placement is a challenging task. There is no google analytics to measure the performance of each campaign or ad manager which allows to pay for the real leads. This is one of the challenges our team works on. In my talk, I'll elaborate on how the combination of modern machine learning approaches for computer vision such as object detection, few-shot learning, events detection, and a universal computing platform automates this process and saves time for thousands of marketers.
Speaker:
Andrey Boiarov received his master's degree from Saint Petersburg State University with specialization in applied statistics. The breakthrough in deep learning that happened in 2012 inspired him to dive into these areas. Now he has 8+ years of experience in Machine Learning, Deep Learning, Computer Vision, Data Science, Robotics. During his career he has made applied research and developed ML systems for real-world applications both in industry and in academia.
His main passion in ML is developing fast adapting algorithms. He believes it will move AI closer to natural intelligence, since people can understand new concepts only by several examples. He has publications in scientific journals, patents. Also, Andrei is an active participant and speaker at scientific and tech conferences.
________________________________________________
_________________________________________________
Job Opportunities: https://dataart.team/vacancies
Facebook: / dataart.dev
Instagram: / dataart
Telegram: https://t.me/dataart
Twitter: / dataart
/ dataart_career
