MSE 153 – INTRODUCTION TO

ARTIFICIAL INTELLIGENCE
Learning Objectives
By the end of this lesson, students will be able to:
• Understand fundamental concepts, techniques and application of AI
• Understand the purpose and basic structure of CNNs.
• Recognize how CNNs are used for image recognition and other tasks
• Evaluate social and ethical impact of AI
• Understand the importance of AI in Robotics and Perception
What is Artificial Intelligence?
• Artificial Intelligence (AI) is the ability of a computer or a machine to
mimic the cognitive functions of humans to perform tasks.

• Cognitive functions: These include abilities like:

 Learning: Acquiring information and improving performance over time.
 Problem-solving: Finding solutions to complex challenges.
 Decision-making: Choosing between different options based on available
information.
 Understanding Language: Processing and interpreting human language.
 Reasoning: Draw logical conclusions.
 Perception: Interpreting sensory information (like sight or sound).
Key Areas of AI
1. Machine Learning: A subset of AI that involves training algorithms to learn
from data and improve their performance over time without being explicitly
programmed. There are three main types of learning:
• Supervised Learning (learning from labeled data)
• Unsupervised Learning (finding patterns in unlabeled data)
• Reinforcement Learning (learning from trail and error)

2. Deep Learning: A subset of machine learning that uses neural networks with
many layers (bring about the term “deep”) to analyze large amounts of data.
• It is useful in tasks like image recognition, speech processing and natural language
understanding.
Key Areas of AI
3. Neural Networks: A model inspired by the human brain, made up of
layers of interconnected nodes (neurons) that work together to process
information.
4. Natural Language Processing (NLP): This area of AI focuses on the
interaction between computers and human languages. It involves
understanding, interpreting and generating human language, enabling
systems to understand speech and text.
5. Computer Vision: In this field, AI techniques allow machines to
interpret and make decisions based on visual data, such as identifying
objects in images or videos.
Key Areas of AI
6. Decision Tree: is a flowchart-like structure used for decision-making
and classification. It splits data into branches based on feature
values, leading to outcomes or class labels at the leaf nodes.
A Summary of How AI Works
• AI systems typically go through the following stages:
1. Data Collection: AI systems require large amount of data to learn from. This data
comes from various sources, such as images, text or sensors.
2. Model Selection: Choosing the right algorithm to solve the problem
3. Model training: The system uses algorithm to identify patterns in the data. Feeding
the data so it can learn from the pattern is called “training”, and it involves
adjusting the models parameters to make accurate predictions or decisions.
4. Testing and Evaluating: After training, the AI model is tested with new, unseen
data to evaluate its performance and ensure it generalizes well to real-world
situations.
5. Deployment: Once the model performs well, it is deployed into a real-world
system, where it can make predictions or automate tasks.
Types of Machine Learning
1. Supervised Learning:
• How it works: The algorithm is trained on a labeled dataset, meaning the data
has known outcomes (labels). The goal is for the model to learn the relationship
between the input and output so it can predict the correct output for new, unseen
data.
• Applications: Spam email detection, image classification and medical
diagnosis.

2. Unsupervised Learning:
• How it works: In this case, the data does not come with labeled outcomes. The
algorithm tries to find patterns or groupings in the data by itself.
• Applications: Anomaly detection and customer segmentation.
Types of Machine Learning
3. Reinforcement Learning:
• How it works: An agent learns by interacting with an environment, receiving
rewards or penalties based on its actions. It aims to maximize cumulative
reward over time.
• Applications: Robotics, game AI and self-driving cars.
Introduction to Convolutional Neural Networks
• Convolutional Neural Network (CNN) is a type of deep learning model
designed to process data with a grid-like topology, such as images.

• CNN are especially powerful for:

1. Image classification
2. Object detection
3. Medical imaging
4. Facial recognition
5. Self-driving car vision system
Basic Architecture of a CNN
• Typical CNN Architecture
Basic Architecture of a CNN
1. Input Layer
• Raw image
2. Convolutional Layer
• Applies filters (kernels) to extract features like edges or textures
3. Activation Function (ReLU)
• Introduces non-linearity: ReLU(x) = max(0, x)
4. Pooling Layer (e.g. Max Pooling)
• Reduces spatial dimension and computation
5. Fully Connected Layer (FC)
• Standard neural network layer for final classification
6. Output Layer (Softmax)
• Outputs class probabilities
Advantages and Limitation of CNN
Advantages:
• Requires fewer parameters than fully connected networks
• Captures spatial and hierarchical patterns in input data
• Translational invariance (detects features regardless of position )

Limitation:
• Less effective on non-grid data (e.g. graphs)
Popular CNN Architectures
• AlexNet
• VGGNet
• ResNet
• MobileNet
Ethical Considerations and Challenges of AI
1. Bias and Fairness: AI systems can inherit biases present in the data used to
train them, leading to unfair outcomes, especially in areas like hiring,
criminal justice and lending.
2. Job Displacement: Automation driven by AI may replace some jobs,
especially in manufacturing, transportation and customer service.
However, it could also create new job opportunities in AI-related fields.
3. Privacy and Security: AI systems that handle sensitive data (health or
financial records) raise concerns about data privacy and security.
4. Accountability: Determining responsibility when AI systems make
decisions (especially in critical areas like healthcare or self-driving cars)
presents legal and ethical challenges.
Applications of AI
1. Healthcare: AI is used in diagnosing diseases, analyzing medical images
and even personalizing treatment plans. AI-powered chatbots and virtual
assistants are also becoming common in patient care.
2. Finance: AI algorithms are employed for fraud detection, risk management
and customer service in banking.
3. Autonomous Vehicles: AI is at the core of self-driving technology, enabling
cars to navigate streets, detect obstacles and making driving decisions.
4. Customer Service: AI-driven chatbots and virtual assistants provide
automated support, improving efficiency and customer experience.
5. Entertainment: Streaming services use AI to recommend content based on
user preferences.
Robotics and Perception
• A robot is a programmable machine capable of carrying out a series of
actions automatically, especially by being responsive to its environment.

• Basic Elements of a Robot

1. Sensors – to perceive the environment
2. Actuators – to take action
3. Controllers – to make decisions
4. Power source – for energy
5. Body – mechanical structure
Perception
• Perception is how a robot interprets and understands data from its
environment, primarily through sensors
• It is the “sense” in the robot’s “sense-think-act” cycle.

• Key goals of Robotic Perception:

 Obstacle detection and avoidance
 Localization and mapping
 Object recognition
 Tracking and navigation
Perception Cycle
1. Sensing – Collect raw data from environment
2. Processing – Filter, clean and interpret data
3. Understanding – Use algorithms to generate insights
4. Decision-making – Choose appropriate action
Applications of Robotics and Perception