Artificial Intelligence and Edge Computing

Credits 3 Credit
Total Contact Hours 45 Hours

Course learning objectives

1. Introduce composite relational model of edge computing along with AI, Machine Learning
and IoT
2. Impart the ML and IoT frameworks suitable for Edge Computing
3. Elevate the learners with a knowledge and practice on tools supporting for Edge Computing
solutions
4 Offer pilot training on modular development boards of Tiny ML

Pre-requisites • Fundamentals of Engineering Mathematics

for the course • Some basic programming knowledge

Unit – I: Artificial Intelligence (AI) and its Next Wave - Edge Computing, Programming
Fundamentals - A backbone for AI and Edge Computing
AI & Edge Computing - Course Description, Relational Model of AI and Edge Computing, ARTIFICIAL
INTELLIGENCE(AI) PRINCIPLES AND PRODUCTS, Machine Learning (ML) Fundamentals and
Principles, Application of ML in Manufacturing and Production Industries, Diesel Generators with IoT
Framework - A Model IoT Architecture, Quick Tour on Edge Devices in IoT, Edge AI and Cloud AI - An
Overview, ‘TinyML’ – A Cutting Edge Field, Case Study on ' Edge AI Practices in Industrial
Applications', Introduction to Python, Python Internals, Programming Essentials in python Part I and
II, DEMO PROGRAMS IN PYTHON - Beginners Perspective, DEMO PROGRAMS IN PYTHON - Beginners
Perspective, Machine Learning Libraries and Functions in Python, House Price Prediction with
Sklearn Framework, Working with Data for ML - Demo with Kaggle Data set, Working with
Micropython - An Overview, Python in " Deep Learning for Self Driving Cars " - A case study approach

Demos on:
• Demo programs in Python and House price prediction

©2023 L&T EduTech. All Rights Reserved The content of this file is confidential and intended for the recipient only. It is strictly
forbidden to share any part of this file with any third party, without a written consent of the sender.
 Unit – II: Python Demos and Case-Studies on Machine Learning (ML) Algorithm Fundamentals,
Demonstrating Unsupervised & Reinforcement Machine Learning Algorithms with Python
demos, Principles and successful demonstrations of Neural Networks (Text Analytics),
Advanced Applications with Deep Learning Networks
MACHINE LEARNING ALGORITHMS ARCHITECTURE, Big Data in the context of Machine Learning,
Machine Learning Types & Algorithm Selection Strategy, Bias and Variance - Trade-off, Machine
Learning Strategies for Business Improvement – An Overview, Preparing Data for Optimization in
Production Manhours, Supervised Machine Learning Algorithm- Principle and types, Regression
algorithm, Classification algorithm, Implementation framework of ML algorithms, Implementation
framework of ML algorithms, PRINCIPLES OF UNSUPERVISED MACHINE LEARNING ALGORITHM,
Clustering algorithm, DBSCAN clustering algorithm, DIMENSIONALITY REDUCTION ALGORITHM –
PRINCIPLE & IMPLEMENTATION OF PCA, Linear Discriminant Analysis - A Quantitative Approach,
Autonomous vehicle embedded with Dimensionality Reduction Algorithm, REINFORCEMENT
MACHINE LEARNING ALGORITHM – WITH A PRACTICE APPROCH IN HVAC SYSTEM, MODEL BASED RL
ALGORITHMS – PRINCIPLE AND EXAMPLE WITH DYNA Q ALGORITHM, Paradigm shift in health care
diagnosis with reinforcement learning - a review exercise, Model Free Reinforcement learning –
exploring policy based methods, DEPLOYMENT OF DEEP Q-LEARNING IN PICK AND PLACE COBOT -
An Industrial application of ML, FUNDAMENTALS OF NEURAL NETWORK, DIGIT RECOGNITION USING
MLP MODEL – HANDS-ON PRACTICE, GRADIENT DESCENT ALGORITHM- WORKING PRINCIPLE,
BACKPROPAGATION ALGORITHM – WORKING PRINCIPLE, Cross-Entropy cost function and its
implementation using MLP, Overfitting and Regularization principles with a hands-on approach,
DIGIT RECOGNITION SYSTEM FOR VISUALLY IMPARED – CNN BASED ML ALGORITHM, STOCHASTIC
GRADIENT DESCENT ALGORITHM PRINCIPLE AND ANALYSIS USING IRIS DATASET, Simulation of
Neural Networks - Weka tool based exercise, Strategic deployment of shallow neural network for
enhancing agriculture - a review exercise, VANISHING GRADIENT PRINCIPLES AND ITS
MEASUREMENTS IN SIGMOID ACTIVATION FUNCTION, Unstable Gradient in Complex networks,
Unstable Gradient in Complex networks, Introduction to convolutional neural networks,
Introduction to convolutional neural networks, Applications of CNN, Generative Network Principles,
Introduction to RNN, Properties and Construction of RNN, Implementation of RNN for NLP, Case-
Study on Recommendation Systems

Demos on:
• Bias and Variance, EDA, Classification on Iris dataset using SVM, Salary Prediction, EV Vehicle
Purchasing, Lung Cancer Prediction, K-Means Clustering, DBSCAN Clustering, PCA,
Reinforcement Learning, Digit Recognition Using MLP, Backpropagation Algorithm, Cross
Entropy, Overfitting and Regularization, Digit Recognition for Visually Impaired, Stochastic
Gradient Descent, WEKA Simulation, Vanishing Gradient, Unstable Gradient, Pneumonia
Detection and NLP

©2023 L&T EduTech. All Rights Reserved The content of this file is confidential and intended for the recipient only. It is strictly
forbidden to share any part of this file with any third party, without a written consent of the sender.
 Unit – III: IoT Architecture and Development model of Edge Computing, Software platforms and
tools for Edge Computing, Hardware tools and resources for Edge Computing
IoT General Architecture, IoT Architecture with AI, Case-Study Analysis, Issues and Challenges in IOT
with Cloud AI, Revised IOT Architecture with Fog AI and Edge AI, Tiny ML working principle, Tiny ML
as SaaS model, High Computing Machine based Edge Architecture, Distributed Training,
Compression technique, Case Study-1 on flow sensing with totalizer, Case-Study-2 on Water
management in public utilities, Software tools and their scope for AI and ML, Additional software
requirements for Tiny ML, Example IDE for ML – Colab, Libraries required for ML, Tensor Flow Library
– Principles, Keras Library – Principles, Arduino IDE for Edge Computing, Tensorflow Lite – Principles,
Example – ML, Example - ML with ANN, Example - Tiny ML AI based Software development
methodologies, Quick survey on the hardware resources, Basics of Arduino Nano BLE Board,
Programming with Arduino Nano BLE(ANB), Programming with built-in Sensors in ANB, Design
framework for Edge computing in ANB, Sinewave prediction model analysis, Demonstration using
ANB, Word Detection model using ANB, Person Detection Model using ANB, Magic Wand model using
ANB
Demos on:
• Diesel Level Monitoring, Crop Analysis, Edge Impulse, WEKA, Blink, Air Quality Monitoring,
Garbage Classification, Gesture Initial Model, Evolution of Software Development
Methodologies, Colors Blink, Microphone & IMU (Gesture Recognition), Sine wave prediction,
Wake Word and Magic Wand

Unit – IV: ML algorithms and scope for Edge Computing in Electrical Engineering Applications,
ML algorithms and scope for Edge Computing in Mechanical Engineering Applications, ML
algorithms and scope for Edge Computing in Civil Engineering Applications
Basics of Substation, ANN Architecture for Substations, Load Prediction in Substations - A case Study,
Geo-spatial database for power infrastructure, Feature Extraction of substation using Deep Learning-
A Case study, Power grid stability and secondary substation model, Estimation of unknown
secondary substation profile - A case study, Characterization of substation site features-Practitioner
approach, CNN based preliminary siting of substation - A Case Study, Substation Device Diagnosis
using unsupervised ML algorithm, CNN based Infrared Fault Image Diagnosis - A Case Study, Impact
of ML in O&G Industry - A Review, Seismic Data Processing Techniques, Geomodeling Process, ML in
Reservoir Engineering, Optimal Production Engineering in O&G Industry, AI in upstream sector of
O&G Industry, Advances in AI Technology for O&G Industry, Fundamentals of Data handling in O&G
Industry, SOA of big data for O&G Industry ML for Civil Engineering, Cloud Data collection about the
Construction Site, Generic ML modelling framework for Civil Engineering Applications, Deep Learning
Techniques in Construction Industry, ML approach for Construction Management, CNN based
Planetary Lego Brick, AI in Transport Engineering - A Survey, Road Traffic Prediction - Bayesian
Approach, ML for Naval Architecture, AI based Wave Height Forecasting - A Case Study

©2023 L&T EduTech. All Rights Reserved The content of this file is confidential and intended for the recipient only. It is strictly
forbidden to share any part of this file with any third party, without a written consent of the sender.
 Demos on:
• Energy Prediction, Solar Energy Power Generation Analysis, Feature Extraction Substation
using Deep Learning, Estimating the Power Grid Stability - A Hands on Approach, Practitioner
approach, Fault Detection in Power Line System, Diesel Price Prediction, Salt Dome
Detection, Reservoir Simulation, Well Bore Testing in Oil and Gas Industry, Stock Price
Prediction for Oil and Gas Industry, Concrete Compressive Strength Prediction, IIoT Data
Collection for ML Algorithms, Building Crack Detection, Classification of LEGO Bricks, Road
Traffic Prediction and Ship Classification

Unit – V: Real-Time Applications of ML - A Structured Approach and Demos, ML algorithms and

scope for Edge Computing in Future
Automated Vehicle support using ML, Fraud System Diagnosis using ML, Deep learning based shop
floor management, Neural networks based ground water quality distribution analysis, Potential
applications of AI in Healthcare – Discussion, Image Classification in IoT Devices - Case Study,
Remotely Piloted Aircraft - Case Study, AI Products - A survey, Education Quality updates in Design,
Development and Delivery using ML, AI impacts in Education, AI open source software libraries,
Computer Vision (Image and Video), Language and Language Reasoning, Speech Recognition,
Healthcare and Biology, Ethical Challenges in AI, Economy Implications due to AI, Policies and
Strategies for AI

Demos on:
• Email Spam Detection, Bosch Performance Line, Diabetic Retinopathy, Edge Impulse -
Flower, Forest Fire, Graduate admission prediction, Car object detection, Language
prediction, Speech recognition, Heart disease prediction and L&T Stock data analysis

Books
Reference Books:
1. Understanding Machine Learning: From Theory to Algorithms- Shai Shalev-Shwartz and Shai
Ben-David
2. Machine Learning Algorithms and Applications- Mohsen Mohammed Muhammad Badruddin
Khan Eihab Bashier Mohammed Bashier
3. TinyML Machine Learning with TensorFlow Lite on Arduino and Ultra-Low-Power
Microcontrollers- OREILLY, By
Pete Warden & Daniel Situnayake
E-resources
1. Machine Learning- by Judith Hurwitz and Daniel Kirsch

©2023 L&T EduTech. All Rights Reserved The content of this file is confidential and intended for the recipient only. It is strictly
forbidden to share any part of this file with any third party, without a written consent of the sender.
 Software
S.No Software Taught Versions available(Student/Paid/Free)
1 Python 3.8.5 - Free
2 Jupyter Notebook 6.1.4 - Free

S.No Software Required Versions available(Student/Paid/Free)

1 Python 3.8.5 - Free
2 Jupyter Notebook 6.1.4 - Free

Project
S.No Project objectives* Project outcomes*
1 To provide AI Fundamentals Learner will be able to code for cleaning and
optimizing the datasets
2 To offer Hands-on practice in ML/DL Learner will be able to analyze the features
algorithms and Python language correlation with solutions
3 To promote real-world solution Learners will be able to design ML/DL model and
thinking develop programming to evaluate a suitable
metric for the model

Course delivery methods Assessment methods

1. Self-paced learning 1. Assessment (MCQ)
2. Online expert sessions 2. Project work

©2023 L&T EduTech. All Rights Reserved The content of this file is confidential and intended for the recipient only. It is strictly
forbidden to share any part of this file with any third party, without a written consent of the sender.
 Course Outcome (COs)
At the end of the course, the student will be able to:
Learning Levels: Re - Remember; Un - Understand; Ap - Apply; An - Analysis; Ev -
Learning Level
Evaluate; Cr - Create
Understand the AI, ML and Python programming in the context of Edge
1. Un
Computing
2. Model Python based ML solutions for simple applications Ap
3. Develop ML and IoT frameworks for EC architecture Cr
Defend for various ML algorithms and CNNs for their compatibility with
4. An
EC
5. Estimate hardware and software tools appropriate for Tiny ML An
Adapt EC based architecture for various applications of :
1) Civil engineering like construction management, transport engineering,
naval architecture etc.
6. 2) Electrical engineering like load prediction, multi-channel Ap
characterization etc., for substations
3) Mechanical engineering domains like reservoir engineering, upstream
sector of O&G Industry etc., (K6,S5)

Scheme of Continuous Internal Evaluation (CIE):

Total
Components Assessments –I,II,III Project work Learning score
Marks

Marks MCQ – 10+10+50 15 15 100

Rubrics:
Target
Levels
<40 Participation
41-70 Completion
71 - 100 First class

©2023 L&T EduTech. All Rights Reserved The content of this file is confidential and intended for the recipient only. It is strictly
forbidden to share any part of this file with any third party, without a written consent of the sender.