Skip to main content

New Research by TSYS School Computer Scientist Riduan Abid Examines Use of Split Neural Networks in Edge Computing

Cloud computing is a critical component in the success of 5G and 6G networks, particularly given the computation-intensive nature of emerging applications. Despite all it advantages, cloud computing faces limitations in meeting the strict latency and bandwidth requirements of applications such as eHealth and automotive systems. To overcome these limitations, edge computing has emerged as a novel paradigm that bring computation closer to the user. Moreover, intelligent tasks related to deep learning demand more memory and processing power than edge devices can handle. To address these challenges, methods like quantization, pruning, and distributed inference have been proposed. A new study by TSYS School computer scientist Riduan Abid, Salmane Douch, Khalid Zine-Dine and Driss Bouzidi of Mohammed V University in Rabat, and Driss Benhaddou of Alfaisal University examines a promising approach for running deep learning models at the edge that employs split neural networks. Split neural networks feature a neural network architecture with multiple early exit points, allowing the model to make confident decisions at earlier layers without processing the entire network. This not only reduces memory and computational demands but it also makes split neural networks well-suited for edge computing applications. As the use of split neural networks expands, ensuring their safety—particularly their robustness to perturbations—becomes crucial for deployment in safety-critical scenarios. Abid et al.'s new study, set to appear in a future issue of IEEE Access, presents the first in-depth study on the robustness of split Edge Cloud neural networks. In doing so it reviews state-of-the-art robustness certification techniques and evaluate split neural networks robustness using the auto_LiRPA and Auto Attack libraries, comparing them to standard neural networks. The results demonstrate that split neural networks reduce average inference time by 75‘% and certify four to 10 times more images as robust, while improving overall robustness accuracy by 1% to 10%.

Comments

Popular posts from this blog

ABDC Releases 2025 Journal Review, Now Ranks Journal Edited by Phil Bryant

The long-awaited journal review being conducted by the Australian Business Deans Council (ABDC) has been released and there are a number of news items that relate to faculty in the Turner College. One of these is the ABDC's decision to now include  Compensation and Benefits Review in its journal rankings. This is big news for the Turner College as its editor, Phil Bryant , is a professor of management in the Turner College. The ABDC is proposing that the journal enter its system for the first time as a C-rated journal. Acting Turner College Dean Tesa Leonce sits on the journal's editorial board, while Turner College management professor Mark James has guest-edited an issue of the journal. Published by SAGE,  Compensation & Benefits Review is the leading journal for senior executives and professionals who design, implement, evaluate and communicate compensation and benefits policies and programs. The journal supports compensation and benefits specialists and academic ex...

The New Financial Times Top 50 Business Journals Just Released

The newest list of Top 50 journals in business has been released by The Financial Times . This list is used by many of the top business schools around the globe to evaluate business faculty. Containing only 50 journals, this list is generally too restrictive to use as an assessment tool at most business schools. The typical business school employs longer lists constructed by the Chartered Association of Business Schools, the Australian Business Deans Council and Scimago. The new FT Top 50 list is provided below. Congrats to Turner College faculty who have published from this list in the past, such as Charles Boster (accounting), Deborah Kidder (management), Laurence Marsh (management), Frank Mixon (economics) and Ed O'Donnell (marketing). Academy of Management Annals Academy of Management Journal Academy of Management Review Accounting Review Accounting, Organizations and Society Administrative Science Quarterly American Economic Review American Sociological Review Contemporar...

Fall 2026 Enrollment Around the USG

As we await official enrollment numbers to be released, we recently asked Google AI to provide some of the enrollment trends and changes for fall 2026 at the University System of Georgia's institutions. That query indicated that Georgia State University 's enrollment increased by 1.24% from last fall, reaching an institution record enrollment of 54,320 students. At the University of Georgia , more than 51,600 student applications were processed for entry in fall 2026, a record for the institution. UGA projected an incoming class of 6,250 students. Georgia College & State University received more than 10,000 student applications for the first time in its history. At Fort Valley State University , 3,310 new students were welcomed this fall, representing an enrollment increase of 27% overall. Next, South Georgia State College  reported strong early-semester numbers with 2,086 students, an 8.25% year-over-year increase fueled by an 18.64% surge in new student enrollment.   Ge...