Adversarial Attacks & Explainable AI — BDAIO Session
Guest session for BDAIO candidates — adversarial attacks on ML models and explainable AI (XAI).
Instructor: Istiaq Ahmed Fahad
Term: June 2025
Location: Institute of Information Technology (IIT), University of Dhaka
Session Overview
A guest session for BDAIO candidates covering two complementary topics in trustworthy AI: adversarial attacks — how small, crafted perturbations can fool ML models — and explainable AI (XAI) — how to interpret and communicate model decisions.
Audience
- BDAIO candidates; no prior background in adversarial ML or XAI required.
Schedule
| Topic | Materials |
|---|---|
| Part 1 — Adversarial Attacks Adversarial attacks on machine learning models — attack types, examples and implications for trustworthy AI. | |
| Part 2 — Explainable AI Explainable AI (XAI) — interpreting model decisions and building trust for BDAIO candidates. |