Artificial Intelligence Landscape
Artificial Intelligence (AI):
The overarching concept that encompasses all technologies
mimicking human intelligence.
AI Landscape: The overall map of AI’s subfields, techniques,
and applications.
Within this landscape, Machine Learning, Deep Learning,
and Generative AI are positioned as nested or overlapping
fields.
Machine Learning (ML) is a subfield of Artificial Intelligence
that enables computers to learn from data to make
decisions and predictions without being explicitly
programmed.
In traditional programming, humans write the rules and
logic manually, while in ML, humans provide data and
example outputs, and algorithms discover the rules
themselves.
In short: ML enables making predictions about the future
by learning from past data.
Example:
A system that predicts whether a student will “pass” or
“fail” based on attendance, homework, and exam scores
uses machine learning.
Machine learning is widely used across all engineering fields
today:
• Computer Engineering: Image recognition, natural language
processing, cybersecurity
• Electrical and Electronics Engineering: Energy consumption
prediction, fault detection, signal processing
• Industrial Engineering: Demand forecasting, production
planning, quality control
• Mechanical Engineering: Robotic control, vibration analysis,
maintenance planning
• Civil Engineering: Traffic congestion prediction, structural
lifespan estimation, concrete durability analysis
These examples show why engineering students should learn
ML:
Future engineers will be expected to design systems that think
and decide based on data.
Machine learning works in five basic stages:
1. Data Collection – Gather many example data points about
the target problem.
2. Data Preprocessing – Clean and convert data into
numerical form.
3. Model Training – Train an algorithm on most of the data
to learn patterns.
4. Model Testing – Measure the accuracy of the model on
unseen data.
5. Prediction – The model can now make decisions on new
data.
Logic:
• Traditional programming: Data + Rules → Output
• Machine learning: Data + Outputs → Rules (Model)
Supervised learning is training on labeled or
defined data.
• Uses labeled data (input + correct output)
• Goal: Predict the correct output for new
data
Unsupervised learning is presenting and
training with unlabeled unknown data.
• Uses unlabeled data
• Goal: Discover hidden patterns and
groups in the data
Reinforcement learning is a branch of
machine learning that focuses on training
computers to make optimal decisions by
interacting with their environment.
In this form of learning, computers learn
through trial and error, exploring their
environment and receiving rewards or
penalties for their actions, rather than
explicit instructions.
• An agent interacts with an environment
and learns through trial and error
• Uses rewards and penalties
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
RapidMiner
• Type: Visual (GUI) platform
• Features: Data cleaning, model building, graphical
results
• How to Use:
o Drag your dataset → add a decision tree block →
click “Run”
• Example: Predicting defective products based on
production data
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
Weka
• Type: Java-based desktop application
• Features: Naive Bayes, decision trees, k-means
• How to Use:
o Upload CSV data → choose an algorithm → click
“Start”
• Example: Classifying emails as spam / not spam
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
Google Cloud Vertex AI (AutoML)
• Type: Cloud-based end-to-end ML platform
• Features: Automatically trains models with
image/text data
• How to Use:
o Upload data through the web interface → click
“Train”
• Example: Classifying product photos into categories
(e.g., shoes, bags)
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
Microsoft Azure Machine Learning Studio
(Designer)
• Type: Cloud-based drag-and-drop platform
• Features: Data prep, model building, deploy as web
service
• Example: Predicting credit approval from customer
demographics
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
IBM Watson Studio (AutoAI)
• Type: Cloud-based visual ML platform
• Features: Automatically selects and evaluates models
• Example: Predicting student success from past exam
data
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
DataRobot
• Type: AutoML platform
• Features: Automatically tries the best algorithms after
data upload
• Example: Predicting machine failures from sensor
data
There are many tools that allow engineering students
to build ML models without coding. These platforms
use drag-and-drop interfaces.
H2O Driverless AI
• Type: AutoML platform
• Features: Automated feature selection, model
building, performance metrics
• Example: Building a model to predict diseases from
medical data
Industrial Engineering
• Demand forecasting, inventory optimization
• Can be done with RapidMiner or DataRobot
Computer Engineering
• Face recognition, spam filtering
• Can be done with Weka or Vertex AI
Electrical and Electronics Engineering
• Energy consumption prediction, smart meter analysis
• Can be done with Azure ML Studio
Mechanical Engineering
• Bearing fault prediction from vibration data
• Can be done with H2O Driverless AI
Civil Engineering
• Concrete durability, structural lifespan prediction
• Can be done with IBM Watson Studio
Scenario: Predicting Software System Failures from Log
DataGoal: Predict potential crashes or failures of software
components running on servers by using previously collected
system log data.
Steps:
1. Collecting System Log Data
2. Uploading the Data to RapidMiner or Azure Machine
Learning Studio
3. Selecting a “Decision Tree” or “Logistic Regression” Model
4. Automatically Splitting the Data into Training and Test Sets
5. Training the Model and Checking Its Accuracy
6. Predicting Failures on New System Log Data
Goal: Enable predictive maintenance to reduce production
downtime.
• Data Quality: Missing or incorrect data leads to poor
models.
• Overfitting: If the model memorizes training data, it fails
on new data.
• Ethics and Privacy: Personal data must be protected.
• Explainability: Engineering decisions should be
transparent and justifiable.
Machine learning is no longer just an optional skill for
engineering students — it has become a core engineering
competency.
With no-code tools, students can:
• Work with real datasets
• Build prediction models
• Develop data-driven thinking skills
Students who start learning ML today will become the
engineers who design intelligent systems in the future.
Tool Type Use Case Requires Coding
RapidMiner GUI platform Production, quality control No
Weka GUI tool Classification, clustering No
Vertex AI (AutoML) Cloud platform Image, text classification No
Azure ML Studio (Designer) Cloud platform Finance, energy prediction No
IBM Watson Studio (AutoAI) Cloud platform Education, healthcare data No
DataRobot AutoML platform Production, industrial applications No
H2O Driverless AI AutoML platform Healthcare, engineering data No
[Link] KESİM
0 533 7693205
mkesim@[Link]
[Link]@[Link]
[Link]
[Link]
UIeCaQ
[Link]
[Link]