Modern Cryptography and AI/ML Security

DSC4110/DSC6111  [L-T-P-C: 3-0-0-3]  |  BS-MS Program (Session: 2026-27, Varsha)

 Information |  Announcements |  Mid-Sem Classes |  End-Sem Classes |  Assignments |  References 

Course offered at Instructor
iiser-tvm-logo The School of Data Science,
Indian Institute of Science Education and Research Thiruvananthapuram,
Thiruvananthapuram, Kerala, India 		👤 Dr. Laltu Sardar ✉

Course Objective
This course builds a rigorous understanding of modern cryptography — from symmetric-key primitives, public-key systems and key infrastructure to the emerging challenges posed by quantum computing. Students will study the mathematical foundations of lattice-based post-quantum cryptography (including NIST-standardized schemes such as CRYSTALS-Kyber and CRYSTALS-Dilithium), and learn to apply cryptographic tools (liboqs, OpenSSL) to build secure, quantum-resistant systems. The course further examines security and privacy challenges specific to AI/ML systems — differential privacy, federated learning, secure aggregation and privacy-preserving inference — and how post-quantum cryptography can be integrated into modern AI/ML pipelines.

Prerequisites: DSC 314 or MAT 2012

Learning Outcomes:

  • Explain the fundamental concepts, security definitions, and threat models of modern cryptography, including symmetric-key cryptography, public-key cryptography, hash functions, and MACs.
  • Analyze the security challenges posed by quantum computing and the principles of post-quantum cryptography, including lattice-based assumptions and NIST-standardized schemes.
  • Apply post-quantum cryptographic primitives and tools such as liboqs, OpenSSL, Kyber, and Dilithium for secure communication and hybrid cryptographic systems.
  • Analyze security and privacy threats in AI/ML systems and compare privacy-preserving techniques used in secure AI/ML systems.
  • Evaluate classical, post-quantum, and hybrid cryptographic solutions for secure system design with respect to security, performance, and deployment feasibility.

📃 Information

✦ Class Timing: To Be Announced ✦ Place: To Be Announced
✦ Full Syllabus (PDF)  ✦ IISER-TVM Complete Syllabus 
✦ Marks Distribution: Time-wise → [Mid-Sem: 40% + End-Sem: 60%]; Type-wise → [Written Exam: 80% + Assignments: 20%]
✦ Pass Marks: According to IISER TVM policy
# Total Marks 100 Written Exam Assignments Total
1 Mid-Semester 30 10 40
2 End-Semester 50 10 60
✦ Total 80 20 100

🗣 Announcements

☑ Course page created — details will be updated as the semester schedule is finalized.

Updates

  1. Course page initialized from syllabus.

📅 Class Schedule: Mid-Sem

# Module Topics Hours
1 Foundations of Cryptography Security goals and threat models, symmetric cryptography, block ciphers and stream ciphers, encryption and decryption, hash functions, MACs, authenticated encryption, PRG and PRP, one-way functions, security definitions and reductions. 4
2 Public-Key Cryptography and Key Infrastructure Integer factorization problem, discrete logarithm problem, elliptic curve discrete logarithm problem, RSA and ElGamal encryption, ECDH, digital signatures (RSA, ECDSA), Diffie-Hellman key exchange, PKI and key management, secret sharing, IBE, ABE, authenticated key exchange, TLS 1.3 overview. 3
3 Introduction to Post-Quantum Cryptography Motivation for post-quantum cryptography, quantum computing and cryptographic threats, Shor's algorithm, Grover's algorithm, NIST PQC standardization process, code-based cryptography, hash-based signatures, multivariate cryptography, isogeny-based cryptography, hybrid classical/post-quantum systems. 3
4 Lattice-Based Cryptography Introduction to lattices, properties of lattices, Shortest Vector Problem (SVP), Closest Vector Problem (CVP), hardness assumptions (GapSVP, SIVP), Learning With Errors (LWE), Ring-LWE, Module-LWE, encryption and KEM from lattices, CRYSTALS-Kyber, CRYSTALS-Dilithium, security parameter selection and performance trade-offs. 3

📅 Class Schedule: End-Sem

# Module Topics Hours
5 Privacy-Preserving AI/ML Systems Security and privacy in AI/ML systems, differential privacy, federated learning, secure aggregation, privacy-preserving inference, secure model sharing and watermarking, access control and authentication. 3
6 PQC Integration for AI/ML Systems Post-quantum secure communication for AI/ML systems, hybrid cryptographic systems, Open Quantum Safe (OQS), liboqs, PQC integration with OpenSSL, performance benchmarking and implementation considerations. 3
7 Advanced Topics in Modern Cryptography Zero-knowledge proofs, fully homomorphic encryption, cryptographic agility and PQC migration strategies, selected papers and recent advances in secure AI/ML systems. 3

📝 Assignments

# Assignment Topic Status
1 To Be Announced Symmetric-key cryptography & hash functions Pending
2 To Be Announced Public-key cryptography & PKI Pending
3 To Be Announced Lattice-based / post-quantum cryptography (Kyber, Dilithium, liboqs) Pending
4 To Be Announced Privacy-preserving AI/ML (differential privacy, federated learning) Pending

📚 References

# Title Author(s) Publisher / Edition
1 Introduction to Modern Cryptography Jonathan Katz, Yehuda Lindell CRC Press, 3rd edition, 2020
2 Cryptography: Theory and Practice Douglas R. Stinson, Maura Paterson CRC Press, 4th edition, 2018
3 Foundations of Cryptography: Basic Applications Oded Goldreich Cambridge University Press, 1st edition, 2004
4 A Graduate Course in Applied Cryptography David Boneh, Victor Shoup Draft edition, 2023
5 Post-Quantum Cryptography D. J. Bernstein, Johannes Buchmann, Erik Dahmen Springer, 1st edition, 2009
6 Privacy-Preserving Machine Learning J. Morris Chang, Di Zhuang, Dumindu Samaraweera CRC Press, 1st edition, 2022

Helpful Tools & Resources

# Resource Useful for
1 Open Quantum Safe (liboqs) Post-quantum algorithm implementations
2 OpenSSL PQC / hybrid cryptography integration
3 NIST PQC Project Standardization documents (Kyber, Dilithium, etc.)
4 CRYSTALS: Kyber & Dilithium Lattice-based KEM and signature schemes
Top