Got it 👍 I’ll make you a detailed summary of Part-B: Advanced
Concepts of Modeling in AI from your PDF for your Class 10 CBSE AI
monthly test.
📘 Detailed Summary – Advanced Concepts of Modeling in AI
1. AI, ML, and DL
Artificial Intelligence (AI): Ability of machines to perform
cognitive tasks like thinking, learning, problem-solving, and
decision-making. Example – Virtual assistants, recommendation
systems.
Machine Learning (ML): Subset of AI where machines learn from
data to make predictions with less human intervention. Example –
Spam filtering, image recognition.
Deep Learning (DL): Subset of ML using Artificial Neural Networks
(ANNs) with multiple layers for complex tasks. Requires large
labeled datasets. Example – Speech recognition, autonomous
vehicles.
2. Data Concepts
Data: Information in raw form (e.g., table of fruits).
Features: Columns/attributes of data (e.g., name, color, size).
Labels: Special features acting as tags (e.g., fruit type).
Labeled Data: Data with tags (apple = fruit).
Unlabeled Data: Raw data without tags.
Training Dataset: Used to teach AI model patterns (like teacher’s
examples).
Testing Dataset: Used to check accuracy (like exams for students).
3. AI Modeling Approaches
1. Rule-Based Modeling
o Rules defined by developer.
o Machine follows fixed instructions.
o Example: FAQ chatbots.
2. Learning-Based Modeling
o Machine learns from data itself.
o Adaptive to changes.
o Uses Supervised, Unsupervised, and Reinforcement
Learning.
4. Types of Learning in AI
(a) Supervised Learning
Trained on labeled data.
Two types:
o Classification: Predicts categories. Example: Apple vs.
Pineapple.
o Regression: Predicts continuous values. Example: Weather
forecasting, house prices.
(b) Unsupervised Learning
Works on unlabeled data.
Machine finds hidden patterns.
Two techniques:
o Clustering: Groups similar data. Example: Market
segmentation.
o Association: Finds relationships. Example: “If a customer
buys bread, likely buys butter.”
(c) Reinforcement Learning
Learns by trial and error with feedback (reward/penalty).
Example: Vending machine – correct coin gives drink (reward),
wrong coin gives nothing (penalty).
Used in robotics, gaming, autonomous driving.
5. Deep Learning (DL) Subcategories
1. Artificial Neural Networks (ANN)
Inspired by the human brain.
Layers:
o Input Layer: Accepts raw data.
o Hidden Layer(s): Processes and extracts features.
o Output Layer: Gives final prediction.
Example: Handwritten digit recognition.
2. Convolutional Neural Networks (CNN)
Specialized for image data.
Identifies features (edges, shapes, objects) in images.
Example: Face recognition, medical image analysis.
6. Key Differences
Rule-Based vs. Learning-Based:
o Rule-based → manual rules, limited adaptability.
o Learning-based → automatic learning, adaptive.
Classification vs. Regression:
o Classification → Predicts categories (yes/no, spam/not spam).
o Regression → Predicts numerical values (temperature, prices).
Clustering vs. Classification:
o Clustering → Groups unlabeled data.
o Classification → Works on labeled data.
7. Applications / Examples
Classification: Loan approval (default/non-default).
Regression: Predict house price or weather.
Clustering: Grouping customers based on shopping behavior.
Association: Recommending items (bread → butter).
Reinforcement Learning: Self-driving cars, robots.
8. Exam-Oriented MCQ Pointers
1. Predict loan eligibility → Classification
2. Predict weather → Regression
3. Data labeled with output → Supervised Learning
4. Spam filter → Supervised Learning
5. Grouping users without labels → Unsupervised Learning
6. Autonomous vehicle learning with rewards/penalties →
Reinforcement Learning
✅ With this summary, you’ll be able to answer theory, case studies,
and MCQs confidently in your test.
Do you also want me to make a mind map + 15 expected MCQs with
answers from this chapter for fast last-minute revision?
