Welcome to Royfactory

Transformer Architecture Basics: From Attention to Modern AI (Lecture 15) In this lecture, we’ll introduce the Transformer architecture, which has become the foundation of modern AI models like GPT and BERT. Unlike RNNs or LSTMs that process sequences step by step, Transformers rely entirely on attention mechanisms and allow parallel processing, making them both faster and more effective. Table of Contents {% toc %} 1) Why Transformers? Traditional sequence models like RNNs and LSTMs process data sequentially, making training slow and prone to long-term dependency issues. ...

Attention Mechanism Basics: Understanding Query, Key, and Value (Lecture 14) In this lecture, we’ll explore the Attention Mechanism, one of the most impactful innovations in deep learning and Natural Language Processing (NLP). The key idea is simple: instead of treating all words equally, the model focuses on the most relevant words to improve context understanding. Table of Contents {% toc %} 1) Why Attention Matters Traditional sequence models like RNN, LSTM, and GRU struggle with long sentences, often forgetting earlier information. Example: ...

GRU Basics: Simplifying Recurrent Neural Networks (Lecture 13) In this lecture, we introduce GRU (Gated Recurrent Unit) networks, a simpler and faster variant of LSTM. You will learn the theory behind GRU gates, compare it with RNN and LSTM, and implement a sentiment analysis model on the IMDB dataset using TensorFlow/Keras. Table of Contents {% toc %} 1) Why GRU? Traditional RNNs suffer from the vanishing gradient problem, making it difficult to learn long-term dependencies. LSTMs solve this with a more complex structure but at the cost of slower training. ...

LSTM Basics: Understanding Long Short-Term Memory Networks (Lecture 12) In this lecture, we will explore LSTM (Long Short-Term Memory) networks. Unlike simple RNNs that struggle with long-term dependencies, LSTMs use special gates to remember or forget information, making them powerful for NLP, speech recognition, and time-series prediction. Table of Contents {% toc %} 1) Why Do We Need LSTM? Traditional RNNs suffer from the vanishing gradient problem, making it difficult to capture long-term context. For example: ...

Neural Network Basics: Build a Simple Image Classification Model (Lecture 8) In this lecture, we’ll introduce Neural Networks, explain their core components, and build a simple image classification model using the MNIST handwritten digits dataset with TensorFlow/Keras. Table of Contents {% toc %} 1) What Is a Neural Network? A Neural Network is made up of interconnected units called neurons, organized in layers. Data flows through Input → Hidden Layers → Output, with weights and activations applied at each step. ...

Recurrent Neural Network (RNN) Basics: Theory and PyTorch Implementation (Lecture 11) In this lecture, we’ll explore Recurrent Neural Networks (RNNs), one of the fundamental architectures for handling sequential data. We’ll cover the theory behind RNNs, their mathematical formulation, limitations, and implement simple RNNs in PyTorch for both text and time-series prediction. Table of Contents {% toc %} 1) What is an RNN? Unlike feedforward networks that treat each input independently, RNNs are designed to remember previous states and use them in predicting future outputs. This makes RNNs highly effective for tasks where context and sequence order matter, such as: ...

Word Embeddings in NLP: From Word2Vec to Transformers (Lecture 10) In this lecture, we will explore Word Embeddings, a fundamental concept in Natural Language Processing (NLP) that allows machines to understand words in terms of vectors. Instead of treating words as discrete symbols, embeddings capture semantic meaning by placing similar words closer in a vector space. Table of Contents {% toc %} 1) What Are Word Embeddings? Word embeddings are numerical representations of words in a continuous vector space. ...

Data Preprocessing and Visualization for AI: A Complete Guide (Lecture 5) In this lecture, we’ll cover data preprocessing—the crucial step to ensure your AI models work with clean, structured, and meaningful data. We’ll also explore data visualization techniques to better understand your dataset. Table of Contents {% toc %} 1) Why Data Preprocessing Matters Model performance depends heavily on data quality. Even the most advanced algorithms can fail if the input data is noisy or inconsistent. ...

Supervised Learning Practice: Classification and Regression (Lecture 6) In this lecture, we’ll explore Supervised Learning, understand the difference between Classification and Regression, review popular algorithms, and implement both tasks using scikit-learn. Table of Contents {% toc %} 1) What Is Supervised Learning? Supervised Learning uses input data (X) paired with labels (y) to train a model that can predict the correct output for unseen inputs. 1.1 Classification vs. Regression Type Description Output Examples Use Cases Classification Predicts a category Spam/Not spam, species Spam detection, diagnosis Regression Predicts a continuous value Price, temperature House price, sales forecast 2) Classification 2.1 Concept Assigns each input to one of several classes. Example: “Is this email spam?” 2.2 Common Algorithms Logistic Regression Decision Tree Support Vector Machine (SVM) Random Forest 3) Regression 3.1 Concept Predicts a continuous value based on input features. Example: “Predict apartment price given size, location, and year built.” 3.2 Common Algorithms Linear Regression Ridge Regression Lasso Regression Decision Tree Regression 4) General Supervised Learning Workflow Prepare data: Separate features (X) and labels (y) Split into training and testing sets Choose a model and train it Predict on test data Evaluate model performance Improve results via hyperparameter tuning 5) Lab: Classification Example (Iris Species) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 from sklearn.datasets import load_iris from sklearn.model_selection import train_test_split from sklearn.linear_model import LogisticRegression from sklearn.metrics import accuracy_score, classification_report # Load data iris = load_iris() X, y = iris.data, iris.target # Train/test split X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.2, random_state=42, stratify=y ) # Train model model = LogisticRegression(max_iter=200) model.fit(X_train, y_train) # Predict y_pred = model.predict(X_test) # Evaluate print("Accuracy:", f"{accuracy_score(y_test, y_pred)*100:.2f}%") print(classification_report(y_test, y_pred, target_names=iris.target_names)) 6) Lab: Regression Example (California Housing Prices) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 from sklearn.datasets import fetch_california_housing from sklearn.model_selection import train_test_split from sklearn.linear_model import LinearRegression from sklearn.metrics import mean_squared_error, r2_score import numpy as np # Load data housing = fetch_california_housing() X, y = housing.data, housing.target # Train/test split X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.2, random_state=42 ) # Train model reg = LinearRegression() reg.fit(X_train, y_train) # Predict y_pred = reg.predict(X_test) # Evaluate mse = mean_squared_error(y_test, y_pred) rmse = np.sqrt(mse) r2 = r2_score(y_test, y_pred) print(f"RMSE: {rmse:.3f}") print(f"R² Score: {r2:.3f}") 7) Evaluation Metrics Classification ...

Unsupervised Learning Practice: Clustering and Dimensionality Reduction (Lecture 7) In this lecture, we’ll explore Unsupervised Learning, understand the concepts of Clustering and Dimensionality Reduction, and implement both techniques using scikit-learn. Table of Contents {% toc %} 1) What Is Unsupervised Learning? Unsupervised Learning finds patterns, structures, or relationships in data without labels. Unlike supervised learning, there is no “answer key”—the model discovers hidden rules on its own. 1.1 Common Applications Customer Segmentation: Grouping customers based on purchase history Anomaly Detection: Fraud detection, early fault detection in systems Data Visualization: Reducing high-dimensional data to 2D/3D for interpretation 2) Clustering Clustering groups similar data points together. Popular algorithms include: ...