Learning Objectives
By the end of this chapter, you will be able to:
- Define Data Science and Machine Learning, and relate them to AI.
- Describe the data science process.
- Distinguish supervised, unsupervised, and reinforcement learning.
- Understand the basic concept and structure of a Neural Network.
- Identify business applications and limitations of machine learning.
Key Concepts in Artificial Intelligence
The fields below build on one another: data science provides the methods, machine learning finds patterns in data, and neural networks are a powerful type of ML model used in modern AI applications.
flowchart TB
AI[Artificial Intelligence: machines performing tasks that need human-like intelligence]
AI --> ML[Machine Learning: learning patterns from data]
ML --> DL[Deep Learning: many-layered neural networks]
DS[Data Science: extracting insight from data] -. uses .-> ML
Data Science
Data Science is an interdisciplinary field that uses scientific methods, processes, algorithms, and systems to extract knowledge and insights from structured and unstructured data. It combines elements of statistics, computer science, and domain expertise.
The data science process
- Define the problem — e.g., “Which customers are likely to stop using our service?”
- Collect data — from databases, apps, surveys, or sensors.
- Clean and prepare — remove errors, duplicates, and missing values (often the most time-consuming step).
- Explore and analyze — use statistics and charts to find patterns.
- Build a model — often with machine learning.
- Communicate and act — present findings to managers and put the model to use.
Machine Learning (ML)
Machine Learning is a subset of AI that focuses on building systems that can learn from data, identify patterns, and make decisions with minimal human intervention. Instead of being explicitly programmed, an ML model is “trained” on a large dataset.
Traditional programming: a person writes rules → the computer applies them.
Machine learning: the computer is given examples (data and answers) → it learns the rules itself.
Types of machine learning
| Type | How it learns | Business example |
|---|---|---|
| Supervised learning | From labeled examples (inputs with known correct answers) | Predicting loan default from past borrowers’ records |
| Unsupervised learning | Finds hidden groups or patterns in unlabeled data | Grouping customers into segments for marketing |
| Reinforcement learning | Learns by trial and error, receiving rewards or penalties | Optimizing delivery routes or game-playing agents |
Worked example (supervised): a bank gives a model 10,000 past loan records, each labeled “repaid” or “defaulted,” with details such as income, existing debt, and repayment history. The model learns which patterns are linked to default. When a new applicant applies, it estimates the risk, helping the loan officer decide.
Neural Networks
An Artificial Neural Network (ANN) is a type of machine learning model that is inspired by the structure and function of the human brain. It consists of interconnected nodes, or “neurons,” organized in layers. Neural networks are particularly effective at finding complex patterns in data and are the foundation of deep learning.
flowchart LR
I1((Input 1)) --> H1((Hidden))
I1 --> H2((Hidden))
I2((Input 2)) --> H1
I2 --> H2
I3((Input 3)) --> H1
I3 --> H2
H1 --> O((Output))
H2 --> O
- Input layer — receives the data (e.g., pixel values of an image).
- Hidden layers — each neuron combines its inputs using weights and passes the result on; training adjusts the weights to reduce errors.
- Output layer — gives the result (e.g., “cat” or “dog,” or a probability).
Deep learning uses many hidden layers and powers face recognition, speech assistants, machine translation, and generative AI chatbots.
Business Applications
- Recommendations — products on e-commerce sites, videos on YouTube.
- Fraud detection — flagging unusual card or wallet transactions.
- Customer service chatbots — answering common questions 24/7.
- Demand forecasting — predicting sales to plan stock.
- Credit scoring — assessing loan risk.
- Image and speech recognition — document scanning, voice search.
In Nepal, banks and digital wallets use automated rules and models to detect suspicious transactions, and businesses increasingly use AI chatbots on Facebook Messenger and websites to handle customer queries, including in Nepali.
Limitations and Ethical Concerns
- Data quality — a model trained on poor or incomplete data gives poor results (“garbage in, garbage out”).
- Bias — if past data reflects unfair decisions, the model can repeat them.
- Lack of explainability — deep neural networks are often “black boxes” that cannot easily explain a decision.
- Privacy — training data may contain personal information.
- Cost and skills — requires large datasets, computing power, and skilled people.
Summary
Data Science is the broad field of extracting insights from data through a process of defining, collecting, cleaning, analyzing, modeling, and communicating. A key part of data science is Machine Learning, where systems learn from data to make predictions—through supervised, unsupervised, or reinforcement learning. Neural Networks, which are modeled after the human brain with input, hidden, and output layers, are a powerful type of machine learning model that has led to many of the recent breakthroughs in Artificial Intelligence. These tools offer large business benefits but must be used with attention to data quality, bias, and privacy.
Key Takeaways
- Data Science is about extracting knowledge from data.
- Machine Learning is a subset of AI where systems learn from data instead of being explicitly programmed.
- Supervised learning uses labeled data; unsupervised finds patterns; reinforcement learns from rewards.
- Neural Networks are a type of ML model inspired by the brain, with input, hidden, and output layers.
- ML raises concerns about bias, explainability, and privacy.
Discussion Questions
- What is the relationship between Data Science and Machine Learning?
- What are some of the ethical concerns related to the use of machine learning?
- Why are neural networks so powerful for certain types of tasks?
Exam-Style Questions
Short answer
- Define machine learning. How does it differ from traditional programming?
- Differentiate between supervised and unsupervised learning with examples.
- What is a neural network? Name its three types of layers.
Long answer
- Explain the relationship among AI, machine learning, deep learning, and data science with a diagram.
- Discuss the business applications and limitations of machine learning with examples.
