Introduction to
AI & Machine
Learning Models
This slide provides an overview of the core concepts and differences
between Artificial Intelligence (AI) and Machine Learning (ML).
What is AI?
COMPUTER SCIENCE FIELD INTELLIGENT TASKS EXAMPLES
AI is a broad field of computer science These tasks include problem-solving, Examples of AI include calculators, voice
focused on creating systems that can natural language understanding, pattern assistants, self-driving cars, object
perform tasks typically requiring human recognition, and decision-making. recognition, and chatbots.
intelligence.
A S A I AND M AC H I N E L EA R N I N G A LG O R IT H M S CONTINUE TO EVOLVE, WE CAN
E X P ECT TO S E E E V E N M O R E I M P R E S S I V E A I APPLICATIONS IN THE FUTURE.
Examples of AI
CALCUL ATORS RECOGNIZING OBJECTS BY
Devices that perform mathematical calculations, a PICTURE N AME
classic example of narrow AI that excels at a AI models trained on large datasets of images that
specific task. can accurately identify and classify objects,
animals, or scenes depicted in photographs,
VOICE ASSISTANTS LIKE SIRI OR demonstrating advancements in computer vision.
ALEXA
AI-powered virtual assistants that can understand CH ATBOTS TH AT PROVIDE
and respond to natural language, enabling hands- CUSTOMER SERVICE
free interaction and voice-based control of various AI-powered conversational agents that can engage
tasks. in natural language interactions to assist
customers with queries, complaints, or other
SELF-DRIVING CARS support requests, automating and improving
customer service experiences.
Vehicles equipped with AI-powered systems that
can navigate roads, detect obstacles, and make
driving decisions without human intervention,
showcasing the potential of AI in transportation.
The Approaches of AI
RULE-BASED SYSTEMS M ACHINE LEARNING
An earlier form of AI that uses manually A more advanced approach to AI that learns from
programmed 'if-then' rules derived from expert data, identifying patterns and relationships to
knowledge to make decisions. make predictions or decisions, without relying on
predefined rules.
Strengths and Limitations of Rule-Based Systems
Ease of Understanding and Adaptability
80% 85%
30%
20%
EASE OF UNDERSTANDING TRANSPARENCY ABILIT Y TO LEARN AND PERFORM ANCE IN COMPLEX
ADAPT ENVIRONMENTS
What is Machine
Learning?
Machine Learning is a specific approach to
achieve Artificial Intelligence (AI) by giving
machines access to data and allowing them to
learn for themselves. Instead of being explicitly
programmed to perform a task, a machine is
trained on data, and it learns patterns and
relationships from this data to make decisions
or predictions.
Examples of Machine Learning
SPAM FILTERS IN EM AIL RECOMMENDATION ENGINES PREDICTIVE M AINTEN ANCE IN
Machine learning models are trained on Netflix, YouTube, and other platforms use INDUSTRI AL SE T TINGS
large datasets of spam and non-spam machine learning algorithms to analyze Machine learning models can be trained
emails to learn the patterns and user preferences, browsing history, and on sensor data from industrial equipment
characteristics of spam messages. These other data to provide personalized to predict when maintenance or repairs
models can then automatically detect and recommendations for movies, videos, and will be needed. This helps optimize
filter out spam emails in a user's inbox. other content the user is likely to enjoy. maintenance schedules, reduce downtime,
and prevent costly breakdowns.
A S M ACHINE L E A R N I N G A LG O R IT H M S C O NT I NUE TO EVOLVE AND BECOME MORE SOPHIST ICATED,
WE C AN EXP E CT TO S E E E V E N M O R E I M P R E SSIVE AI APPLICATIONS IN THE FUTURE.
Examples of Machine Learning
IM AGE SPAM FILTERING FRAUD DE TECTION RECOMMENDATION N ATURAL L ANGUAGE
CL ASSIFICATION SYSTEMS PROCESSING
Machine learning Financial institutions use
Machine learning models algorithms can analyze machine learning to E-commerce and media Machine learning models
can be trained to email content and header analyze transaction data platforms leverage can be trained to
recognize and classify data to accurately and detect anomalies or machine learning to understand, interpret, and
different objects, animals, identify and filter out suspicious activities that analyze user preferences generate human
or scenes in images with spam messages based on may indicate fraudulent and behaviors to provide language, enabling
high accuracy. patterns learned from behavior. personalized product or applications such as
previous examples. content chatbots, language
recommendations. translation, and text
summarization.
Strengths and Limitations of Machine Learning
Comparison of machine learning capabilities (0-100 scale)
90%
85%
80%
60%
FLEXIBILIT Y SCAL ABILIT Y INTERPRE TABILIT Y DATASE T SIZE REQUIREMENT
The Evolution of AI
2000S-2010S
Deep learning, a more
advanced form of
1950S-1970S
machine learning,
Rule-based systems becomes widely
and expert systems adopted, allowing AI
emerge, using to tackle complex
predefined if-then problems like image
rules to make recognition and
decisions. natural language
processing.
2010S-PRESENT
Hybrid approaches
1980S-1990S combining rule-based
Machine learning systems and machine
algorithms like neural learning emerge,
networks and leveraging the
decision trees are strengths of both to
developed, enabling create more robust
systems to learn from and adaptable AI
data. systems.
Types of Machine Learning
SUPERVISED LEARNING UNSUPERVISED REINFORCEMENT
LEARNING LEARNING
The model is trained on
labeled data, meaning the The model is trained on The model learns by
input comes with the correct unlabeled data and must find interacting with an
output. Examples include patterns and relationships in environment and receiving
spam filtering, image the data. Examples include rewards or penalties based on
classification, and predictive customer segmentation, its actions. Examples include
analytics. anomaly detection, and game-playing algorithms,
recommendation systems. robotics, and resource
management.
