Advanced Concepts of Modeling in AI
Subjective
Assertion and Reason Questions – With Solutions
Directions: Choose the correct option:
- (a) Both A and R are true, and R is the correct explanation of A.
- (b) Both A and R are true, but R is not the correct explanation of A.
- (c) A is true, but R is false.
- (d) A is false, but R is true.
1. Assertion (A): Supervised learning models require labeled data for training.
Reason (R): Labeled data provides a clear mapping between inputs and outputs.
2. Assertion (A): Predicting house price based on size is a Regression model.
Reason (R): Regression models predict discrete values.
3. Assertion (A): Anomaly detection is an application of Unsupervised Learning.
Reason (R): It detects unusual patterns in unlabeled data.
4. Assertion (A): Rule-based AI models are adaptive.
Reason (R): They operate using fixed predefined instructions.
5. Assertion (A): Deep Learning is a subset of Machine Learning.
Reason (R): It is based on neural networks and large datasets.
6. Assertion (A): Reinforcement Learning uses trial and error.
Reason (R): It is useful when predefined data is insufficient.
7. Assertion (A): Spam detection is a Classification task.
Reason (R): Classification predicts continuous values.
8. Assertion (A): Hidden layers perform core processing in ANN.
Reason (R): Input and output layers perform all computations.
9. Assertion (A): Unsupervised learning resembles teacher-student learning.
Reason (R): It works without labeled guidance.
10. Assertion (A): Testing dataset trains the model.
Reason (R): Training dataset evaluates accuracy.
11. Assertion (A): Weights are adjusted during learning.
Reason (R): Adjusting weights reduces prediction error.
12. Assertion (A): Reinforcement learning is supervised learning.
Reason (R): It uses reward feedback.
13. Assertion (A): Clustering is unsupervised learning.
Reason (R): It groups similar data points without labels.
14. Assertion (A): Predicting hospital stay type is regression.
Reason (R): Output is continuous.
15. Assertion (A): Input layer does not process data.
Reason (R): It only forwards input to next layer.
16. Assertion (A): Association models find relationships between variables.
Reason (R): Example: recommending butter with bread.
17. Assertion (A): Deep Learning requires small data.
Reason (R): It uses vast amounts of data.
18. Assertion (A): Perceptron is a simplified decision model.
Reason (R): It sums weighted inputs and applies threshold.
19. Assertion (A): Spam filter is learning-based.
Reason (R): It improves with new data.
20. Assertion (A): Attaching tags to data is Data Labeling.
Reason (R): It is mainly required for supervised learning.
Difference Between Rule-Based and Learning-Based AI Models
| Basis of Comparison | Rule-Based AI Model | Learning-Based AI Model |
|---|---|---|
| Definition | Works using predefined rules and logic explicitly written by a developer. | Learns patterns from data and improves performance through experience. |
| Decision Making | Decisions are made strictly according to programmed rules. | Decisions are made based on patterns identified from training data. |
| Adaptability | Does not adapt automatically to new data or situations. | Can adapt and improve when exposed to new data. |
| Learning Ability | No self-learning capability; requires manual updates. | Has self-learning capability using algorithms and training processes. |
| Data Requirement | Does not require large datasets for functioning. | Requires data (often large datasets) for training and improvement. |
| Complexity Handling | Suitable for simple and well-defined problems. | Effective for complex and dynamic problems. |
| Example | A basic chatbot with fixed responses. | A spam email filter that improves over time. |
| Main Limitation | Static and inflexible; cannot handle unseen situations easily. | Requires computational resources and quality data for accurate performance. |
Three Main Types of Learning-Based Models
1. Supervised Learning
Supervised Learning is a method where the model is trained using labeled data. Each input example is paired with a correct output, allowing the model to learn the relationship between features and outcomes. After training, the model can predict results for new, unseen data.
Common tasks include classification (predicting categories) and regression (predicting numerical values).
2. Unsupervised Learning
Unsupervised Learning works with data that does not contain predefined labels. The model independently explores the dataset to discover hidden patterns, structures, or relationships among the data points.
Typical applications include clustering (grouping similar items) and association analysis (finding relationships between variables).
3. Reinforcement Learning
Reinforcement Learning is based on a reward-driven approach. The model learns by interacting with an environment and receives feedback in the form of rewards or penalties. Its objective is to make a sequence of decisions that maximizes the overall reward over time.
This approach is commonly used in robotics, game playing, and autonomous systems.
Primary Difference Between Labeled and Unlabeled Data
| Basis | Labeled Data | Unlabeled Data |
|---|---|---|
| Definition | Data that includes input features along with the correct output or target value. | Data that contains only input features without any predefined output or category. |
| Guidance | Provides clear supervision to the model during training. | No direct supervision; the model must discover patterns on its own. |
| Used In | Supervised Learning | Unsupervised Learning |
| Example | Email marked as “Spam” or “Not Spam”. | A collection of emails without any spam labels. |
| Purpose | Helps the model learn the relationship between inputs and outputs. | Helps the model identify hidden patterns or group similar data. |
Labeled data contains both input information and the correct answer, allowing the model to learn with guidance. Unlabeled data contains only input information, requiring the model to find patterns or structures independently.
Clustering and Classification
What is Clustering?
Clustering is a technique used in unsupervised learning where data points are grouped based on similarities. The model does not receive predefined labels. Instead, it analyzes patterns within the dataset and organizes similar items into clusters. The goal is to uncover hidden structures in the data.
Example: Grouping customers based on their purchasing behavior without knowing their categories in advance.
What is Classification?
Classification is a supervised learning method where data is assigned to predefined categories or classes. The model is trained using labeled data, meaning it learns from examples that already have correct outputs.
Example: Predicting whether an email is “spam” or “not spam” based on labeled training data.
Difference Between Clustering and Classification
| Basis | Clustering | Classification |
|---|---|---|
| Type of Learning | Unsupervised Learning | Supervised Learning |
| Data Used | Unlabeled Data | Labeled Data |
| Purpose | Finds patterns and groups similar data points. | Assigns data to predefined categories. |
| Output | Clusters formed based on similarity. | Specific class label assigned to each data point. |
| Example | Customer segmentation. | Email spam detection. |
Clustering discovers natural groupings within data without prior knowledge of categories, whereas classification assigns data into already defined classes using labeled examples.
Neural Networks and Their Layers
What is a Neural Network?
A Neural Network is a computational model inspired by the working structure of the human brain. It consists of interconnected nodes (also called neurons) that process information in layers. Each connection between nodes has a weight, and these weights are adjusted during training to improve the accuracy of predictions.
Neural networks are widely used in tasks such as image recognition, speech processing, text analysis, and pattern detection. Their main strength lies in their ability to automatically learn complex patterns from large amounts of data without being explicitly programmed.
Functions of the Three Layers of a Neural Network
| Layer | Function |
|---|---|
| 1. Input Layer | The input layer receives raw data and passes it to the next layer. It does not perform calculations. Each node in this layer represents a feature of the input data. |
| 2. Hidden Layer(s) | Hidden layers perform the main processing tasks. Each neuron calculates a weighted sum of inputs, adds a bias, and applies an activation function. This layer extracts patterns and relationships from the data. |
| 3. Output Layer | The output layer produces the final result of the network. It converts the processed information into a meaningful prediction, such as a class label or a numerical value. |
The input layer receives data, the hidden layers analyze and transform it, and the output layer delivers the final prediction.
