A cross-border community for researchers with openness, equality and inclusion
Join USS Create Account Sign In
Join USS Create Account Sign In
Presentation details
  1. Home
  2. Presentation list
  3. Presentation details
FaultXAI: A Controlled Fault-Injection Study of Explanation Drift in ECG Classifiers
ID：86 View protection：Participant Only Updated time：2026-07-22 16:09:48 Views：16 Online

Start Time：2026-07-31 15:25

Duration：15min

Session：[S4] Computer Vision and Pattern Recognition [S4-5] Computer Vision and Pattern Recognition

Video No right to play Slide
Abstract
Explainable AI (XAI) methods are increasingly deployed to interpret deep learning models in clinical decision support systems. However, the stability of these explanations under realistic signal degradation remains poorly understood. We introduce FaultXAI, a systematic framework for quantifying explanation drift the divergence between explanations of clean and corrupted inputs under controlled fault injection. We evaluate four clinically relevant fault types (Gaussian noise, baseline wander, lead dropout, segment dropout) across two benchmark ECG datasets: PTB-XL (12-lead, 21,837 records; 2,163 test records) and MIT-BIH (2-lead, 10,452 beats). Using Integrated Gradients and multiple stability metrics, our 5-seed experiments reveal: (1) additive faults exhibit strong cross-dataset agreement (r > 0.98), suggesting model intrinsic behaviour; (2) structural faults show dataset dependent effects; and (3) explanation drift can occur substantially even when predictive performance remains high (ρ < 0.7 at accuracy > 0.85). Our framework establishes reproducible benchmarks for
Keywords
Explainable AI, ECG classification, explanation drift, fault injection, robustness, deep learning, clinical decision support
Speaker
Deepak Chamarthi
Nagarjuna University

Post comments
Verification Code Change Another
All comments
Important Dates

  • Conference date

    07-30

    2026

    -

    08-01

    2026

  • 07-28 2026

    Draft paper submission deadline

  • 07-28 2026

    Registration deadline

Sponsored By

The United Societies of Science

Organized By

Kongunadu College of Engineering and Technology

Contact info
Previous Conferences
请输入有效的邮箱地址
© Copyright 2024 USS - the United Societies of Science - All rights reserved. The USS website is developed and maintained by BoomSCI limited.
×

USS WeChat Official Account

USSsociety

Please scan the QR code to follow
the wechat official account.