| Attribute | Detail |
|---|---|
| Format | Online (e-LMS) |
| Level | Advanced |
| Duration | 3 Weeks |
| Certification | e-Certification + e-Marksheet |
| Fee | ₹10749 / $124 |
| Tools | Foolbox ART CleverHans |
About the Adversarial ML & Security Threats Course
Adversarial ML & Security Threats is an advanced, research-driven training program that explores how malicious actors exploit weaknesses in machine learning systems.
As AI becomes central to decision-making in defense, finance, healthcare, and cybersecurity, understanding adversarial threats is essential. This course provides technical insights into how models can be tricked, poisoned, or reverse-engineered, and trains participants to build defenses against such attacks using robust ML practices, secure deployment methods, and adversarial training.
Program Highlights
• Comprehensive coverage of Adversarial ML from fundamentals to advanced applications
• Hands-on projects and real-world case studies in AI
• Expert-curated curriculum aligned with current industry standards
• Access to recorded lectures and e-LMS platform for flexible, self-paced learning
• e-Certification and e-Marksheet upon successful completion
• Dedicated mentor support and interactive doubt-clearing sessions
• Practical experience with tools: Foolbox, ART, CleverHans
• Career-oriented training for academic and professional growth in AI
Course Curriculum
Module 1: Foundations of Adversarial Machine Learning
- Define Adversarial ML and its importance.
- Outline historical context and emerging trends.
- Categorize types of adversarial threats (white-box, black-box, gray-box).
- Survey vulnerabilities within ML pipelines.
Module 2: Attacks Against ML Models
- Execute evasion attacks on diverse models (image, text, tabular).
- Implement poisoning attacks during model training.
- Perform model inversion and membership inference attacks.
- Utilize leading adversarial ML tools (Foolbox, ART, CleverHans).
Module 3: Defensive Strategies and Robust Model Design
- Apply adversarial training techniques to enhance resilience.
- Employ input preprocessing and gradient masking for defense.
- Explore certified defenses and formal security guarantees.
- Evaluate model robustness using specialized metrics.
Module 4: Security in the ML Lifecycle
- Design secure data pipelines and ensure label integrity.
- Identify and mitigate attack surfaces in model deployment.
- Conduct threat modeling for machine learning systems.
- Implement secure MLOps practices and monitoring pipelines.
Module 5: Real-World Applications and Future Challenges
- Analyze real-world case studies of attacks on AI systems.
- Investigate adversarial threats in federated learning and Edge AI.
- Address legal, ethical, and compliance risks in AI security.
- Practice AI red teaming and offensive security testing.
Module 6: Capstone and Emerging Trends
- Design and conceptualize an adversarial attack scenario.
- Simulate and evaluate robust defense mechanisms.
- Present a final capstone project showcasing applied skills.
- Examine future directions in AI security and regulation.
Tools, Techniques, or Platforms Covered
Foolbox ART CleverHans
Real-World Applications
- Apply Adversarial ML skills directly to academic research, thesis work, and publications
- Build a professional portfolio showcasing practical AI competencies
- Solve industry-relevant problems using Adversarial ML methodologies and tools
- Contribute to open-source projects and collaborative research in AI
- Prepare for competitive examinations, interviews, and professional certifications in AI
Who Should Attend & Prerequisites
- Industry-recognized e-Certification + e-Marksheet from NSTC
- Hands-on training with practical projects and industrial datasets
- Dedicated expert mentorship and doubt resolution
Certification
