The LNM Institute of Information Technology

Menu
CMLBDA Logo - 1

National Workshop on Mathematics for Deep Learning (NWMDL-2026)

09-13 March 2026

jointly organized by
Department of Mathematics, and Centre for Machine Learning and Big Data Analytics (CMLBDA)

  • nwmdl2026@lnmiit.ac.in

Registration
Committee
Schedule
Contact

About LNMIIT, Jaipur

The LNM Institute of Information Technology is a deemed-to-be-university established in 2002 as a joint venture between Govt. of Rajasthan and Lakshmi & Usha Mittal (LUM) Foundation as a philanthropic initiative. It is accredited by the National Assessment & Accreditation Council (NAAC) as an ‘A’ grade institution and all the engineering programmes of the institute are approved by the AICTE. It offers programmes such as B.Tech., B.Tech.-M.Tech. (5-year integrated degree), M.Tech., M.S. by Research, M.Sc. and Ph.D. For more details, please visit www.lnmiit.ac.in

About the Workshop

This workshop aims to bring together advanced mathematical ideas and the rapidly evolving field of deep learning. The workshop focuses on the role of core mathematical tools, particularly from linear algebra, operator theory, and graph theory, in the analysis, design, and interpretation of neural network models. The program highlights how these rigorous mathematical structures underpin the behaviour, stability, and performance of state-of-the-art neural architectures, and how these ideas are essential in cutting-edge research worldwide.

The workshop is structured to progress systematically from mathematical foundations to deep learning applications. Participants will be introduced to the mathematical principles underlying modern neural network architectures, including convolutional neural networks and graph neural networks, and to how concepts from linear algebra, operator theory, and spectral graph methods naturally arise in the analysis and design of deep learning models. The program combines theoretical lectures with hands-on sessions, enabling participants to directly experiment with these ideas through the implementation and analysis of deep learning models.

This workshop is intended for faculty members, postdoctoral researchers, and research scholars seeking a deeper, principle-driven understanding of deep learning beyond surface-level implementation. By the end of the workshop, participants will gain a clearer perspective on how mathematical foundations shape modern deep learning models, how contemporary architectures are developed from these principles, and how this understanding can be applied across domains such as scientific computing, computer vision, molecular sciences, and physics-based learning. Overall, the workshop aims to strengthen the mathematical foundation of the research community and foster more robust, interpretable, and theoretically grounded advances in deep learning.

Topics to be covered:

  1. Operator norms, condition numbers, singular value decomposition (SVD), and dimensionality reduction
  2. Universal approximation theorem and theoretical foundations of modern neural networks
  3. Mathematical principles of convolutional neural networks (CNNs)
  4. Few-shot learning and self-supervised learning
  5. Graph theory and spectral methods for learning on structured data
  6. Graph neural networks: core architectures and message-passing frameworks

List of Resource Persons

  1. Prof.  Sivakumar K. C., IIT Madras
  2. Prof. C. S. Sastry, IIT Hyderabad
  3. Prof. R. Balasubramanian, IIT Roorkee  (yet to be confirmed)
  4. Dr. Ankit Jha, LNMIIT, Jaipur
  5. Prof. Sivaramakrishnan Sivasubramanian, IIT Bombay
  6. Prof. B. S. Panda, IIT Delhi
  7. Dr. Subrat Dash, LNMIIT Jaipur
  8. Dr. Akansha, LNMIIT Jaipur

Target Audience

Research Scholars, Post-Doctoral Fellows, Faculty Members and other interested participants from universities, institutions, industries and R&D Organizations.