TinyML and Reliability: Does Quantization affect the Reliability of Machine Learning Models?

Authors

Deriech Cumming II, Computer Engineering, College of Engineering, North Carolina Agricultural and Technical State University

Files

Description

Machine Learning (ML) ability to handle and automate complicated tasks allows for its wide application (facial recognition, predictive text, ChatGPT, etc.). TinyML is a field that aids in ML’s shortcomings (ML is complex and memory expensive) finding ways to compress these large models, which widens the scope of application that ML can aid. One such scope involves mission-critical tasks (self-navigation, healthcare, manufacturing, etc.) which have catastrophic consequences for failure in these tasks. Ensuring we can safely implement these machine learning applications such that they run consistently is important.

Publication Date

4-1-2025

Keywords

Machine Learning, Reliability, Model Compression, Quantization

Recommended Citation

Cumming, Deriech II, "TinyML and Reliability: Does Quantization affect the Reliability of Machine Learning Models?" (2025). 2025 Graduate Student Research Symposium. 172.
https://digital.library.ncat.edu/gradresearchsymposium25/172