Thesis supervisor: Balázs Renczes
Location of studies (in Hungarian): Department of Measurement and Information Systems Abbreviation of location of studies: MIT
Description of the research topic:
"During digital signal processing, the operations can only be performed with finite accuracy. After the calculation, the result is rounded to the nearest representable number. Theoretically, we know that the result is inaccurate. Notwithstanding, we generally assume that it is in the order of the representing system. However, this assumption is not even approximately true. If many operations are performed, the result's error may be by several orders of magnitude higher than that of the number representation.
When floating-point arithmetic (especially double and quadruple precision arithmetics) became widely used, the problem became temporarily insignificant. Namely, these representation methods exceed the general accuracy requirement of the user. However, with the expansive growth of high-speed 4G and 5G networks, and big data applications where the number of operations is enormous, the investigation of roundoff errors is brought into focus, again. On smaller precision arithmetic, the operations can be performed faster. The question is whether the accuracy of the result is satisfactory for the user. Such a current research area is the analysis of the roundoff errors in neural networks, but the framework can be applied generally. The investigation of the topic involves thorough considerations, analytical thinking, and a strong interest in probability theory.
"
Required language skills: English Number of students who can be accepted: 1
Deadline for application: 2021-09-01
2024. IV. 17. ODT ülés Az ODT következő ülésére 2024. június 14-én, pénteken 10.00 órakor kerül sor a Semmelweis Egyetem Szenátusi termében (Bp. Üllői út 26. I. emelet).
Login
