Lecture of SASA Corresponding Member Miodrag Mihaljević ’AI for BlockChain Technology and Vice-versa’
As part of the lecture series ’Artificial Intelligence’, SASA Corresponding Member Miodrag Mihaljević will give a lecture titled ’ ’AI for BlockChain Technology and Vice-versa’ on Tuesday, 31 October, at 11 a.m.
In addition to being among high-priority structures within information technologies, and opening the door to numerous new applications, Artificial intelligence (AI) and BlockChain (BC) have appeared as mutually supportive fields. This lecture provides certain highlights of AI usage for BC system analysis, as well as the importance of BC for the development of reliable applications based on a large amount of data and artificial intelligence.
The BC technology application in federated learning (FL) is an interesting illustration. FL is an approach to learning which enables distributed entities to mutually educate AI models by using their local data, but without data exchange, i.e. FL enables devices to train machine learning models locally and distribute only model parameters instead of the distribution of unprocessed data to the central server.
FL provides a solution to privacy and safety issues in centralized machine learning and reduction of communicative network load due to the exchange of large amounts of data to the central server.
In the traditional FL, coordinated by the central FL server, various devices cooperate to solve ML issues by means of updating the local model with their local data.
The features of BC technology facilitate the use of this approach without the need for the existence of a centralized FL server. By BC technology integration into the FL framework, aggregations are conducted without the central server by various hubs and in a decentralized manner.
Miodrag J. Mihaljević is a research professor and deputy director at the Mathematical Institute, Serbian Academy of Sciences and Arts, Belgrade. His main research fields include cryptology, information security, BlockChain technology and elements of coding theory. He has published more than 100 papers in the leading international journals and conference proceedings, as well as over 200 publications. He co-invented eight accepted patents in Japan, the USA and China.
He organized and was a leading researcher in more than ten national projects and participated in over ten international projects.
Since 1997, he has been a visiting scholar at universities and research institutions in Japan, including the University of Tokyo, SONY Computer Science Laboratories, Tokyo, National Institute of Advanced Industrial Sciences and Technology (AIST) and Chuo University in Tokyo. He received the National Award of the Serbian Academy of Sciences and Arts in 2013 for ten-year achievements. The following year saw him becoming a member of Academia Europea. In three consecutive years (2020, 2021 and 2022) he was Included in the ranked list colloquially known as “World’s Top 2% Scientists by Stanford University for his overall achievements in his career. He was a visiting editor of several journals and is one of Springer SN Computer Science journal editors, In 2021, he was elected a SASA corresponding member. More information is available at https://www.sanu.ac.rs/en/member/mihaljevic-miodrag/
A live stream is available at LIVE STREAMING
The footage of previously held lectures within the Artificial Intelligence lecture series is available at Снимци предавања из циклуса „Вештачка интелигенција”.