Machine Learning for System Diagnosis

Dear Colleagues,

Machine learning (ML) is making a strong resurgence in tune with the massive generation of unstructured data, which in turn requires massive computational resources. From edge devices to the cloud, there are many applications that rely on the significant efficiency of ML techniques. However, with technology node scaling, such systems become more susceptible to faults, which in turn can impact the NN accuracy. Faults can be caused by different reasons such as hardware defects, voltage fluctuations, and radiation effects, as well as software issues such as stack overflow, and compiler uncertainty, among others. Dealing with such faults and designing resilient and fault-tolerant ML systems in edge devices and/or the cloud requires novel and efficient hardware-level techniques.

The goal of this Special Issue is to present a collection of fault-tolerant ML system designs to support efficient ML systems that are resilient against hardware faults. This Special Issue’s scope ranges from analyzing and understanding the fault tolerance of state-of-the-art ML systems, designing resilient ML systems, and exploiting such mechanisms in real-world environments. 

Potential topics include, but are not limited to:

• Experimentally or theoretically analyzing the fault tolerance of ML systems;
• Designing efficient resilient ML systems on hardware accelerators (e.g., ASIC, FPGA) or general-purpose processors;
• Understanding and improving the fault tolerance of ML systems on the cloud and HPC systems;
• Understanding and improving the fault tolerance of ML systems on edge devices;
• Understanding how ML systems can improve fault tolerance techniques;
• Improving the reliability of hardware and architecture targeted for machine learning;
• Hardware techniques for trading off the reliability of machine learning-based systems with energy, power, security, and performance;
• The reliability of machine learning-based applications on heterogeneous systems.

Dr. Leonardo Bautista-Gomez
Dr. Behzad Salami
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Micromachines is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2100 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.