Structural Testing Methodology for Deep Neural Networks based on RRAM

Conference Paper (2026)
Author(s)

Emmanouil Anastasios Serlis (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Emmanouil Arapidis (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Theofilos Spyrou (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Anteneh Gebregiorgis (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Mottaqiallah Taouil (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Said Hamdioui (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Moritz Fieback (TU Delft - Electrical Engineering, Mathematics and Computer Science)

Research Group
Computer Engineering
DOI related publication
https://doi.org/10.1109/ETS69887.2026.11591712 Final published version
More Info
expand_more
Publication Year
2026
Language
English
Research Group
Computer Engineering
Publisher
IEEE
ISBN (electronic)
9798319517630
Event
31st IEEE European Test Symposium, ETS 2026 (2026-05-25 - 2026-05-29), Chania, Greece
Page Views
101
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

RRAM-based compute-in-memory (CIM) AI accelerators integrate memory cells with mixed-signal peripherals to enable energy-efficient, low-latency edge computing. These architectures require novel test techniques for unique electrical defects. Current system-level functional tests cannot guarantee defect-free circuits, while accurate circuit-level techniques remain non-scalable to Deep Neural Networks (DNNs). To bridge this gap, we present the first structural test methodology for CIM DNNs, unifying system-level analysis with circuit-level precision. We accurately model RRAM crossbar defects to formulate test patterns, and then propagate results across layers for comprehensive structural testing. Generated test patterns are optimized either for 96% defect coverage or test time reduction by a factor of three while covering all unique defects and maintaining over 73% defect strength coverage.

Files

– Personal use only – Dutch Copyright Act (Article 25fa)
warning

File under embargo until 11-01-2027