Deep Learning in Material Science and Engineering: From Concepts to Applications

Overview

ABOUT THE COURSE:

Why Deep Learning in MSE? Deep Learning has widely been used in various sectors; health care, finance, agriculture, education, energy, entertainment etc have been using Deep Learning to improve their performances.

Tech giants have been pushing their boundaries to achieve more as Deep Learning is the driving force everywhere and similarly the demand for skilled professional has steeply been rising. The design and discovery of new materials, processing optimization and prediction of properties would require applications of Deep Learning techniques and even deployment of the same.

The present course is meant for students preparing for higher studies, researchers applying AI tools, the industry professionals wanting to upskill. This is a balanced course with conceptual learning and hands-on with ability to build intuition, apply concepts to data, and interpret model output.

Upon successfully completing this course, you will have a mini portfolio of Deep Learning projects, which can be showcased to the potential employers or use for research.INTENDED AUDIENCE:

Materials, Mechanical, Physics, ChemistryPREREQUISITES:

Knowledge of Artificial Intelligence and Machine Learning

Syllabus

• Introduction to Deep Learning in Material Science
Overview of Deep Learning
Importance for Material Science and Engineering (MSE)
Course objectives and expected outcomes
• Fundamentals of Deep Learning
Neural Networks basics
Activation functions and optimization techniques
Overfitting and regularization techniques
Deep learning frameworks (TensorFlow, PyTorch)
• Data Handling in MSE
Data types and sources in Material Science
Data preprocessing and normalization
Feature extraction techniques specific to MSE
• Architecture Design for Material Science
Convolutional Neural Networks (CNNs) for microstructure analysis
Recurrent Neural Networks (RNNs) for time series data
Autoencoders for material property prediction
• Application of Deep Learning in Material Design and Discovery
Case studies on material discovery using Deep Learning
Predictive modeling of material properties
Applications in nanomaterials and composites
• Process Optimization using Deep Learning
Simulation and optimization in manufacturing processes
Deep reinforcement learning for process control
• Interpretation and Visualization of Model Outputs
Techniques for model explainability
Visualization tools for interpreting deep learning results
• Hands-on Projects
Project 1: Building a CNN for Microstructure Analysis
Project 2: Predicting Material Properties with Autoencoders
Project 3: Process Optimization in Materials Manufacturing
• Deployment of Deep Learning Models
Introduction to deploying models in production environments
Best practices for maintaining and updating models
• Future Trends and Challenges in Deep Learning for MSE
Emerging technologies and applications
Ethical considerations and challenges
• Course Review and Portfolio Development
Review of key concepts
Guidance on building and presenting a project portfolio
• Assessment and Certification
Quizzes and practical exams
Final project presentation and evaluation

---

Taught by

Prof. Krishanu Biswas

---

Subjects

Computer Science