IF3270 Pembelajaran Mesin

Back to index

UTS

• Intro

  • Foundational Definitions and Scope of Machine Learning
  • Machine Learning Paradigms - Supervised Unsupervised and Reinforcement Learning
  • Supervised Learning Techniques - Regression and Classification
  • Neural Network Architectures for Specialized Tasks
  • Machine Learning System Design Workflow

• Ensemble Methods

  • Foundations of Ensemble Learning - Concepts, Terminology, and the Wisdom of Crowds
  • Bagging and Random Forests - Bootstrap Sampling, Model Diversity, and Majority Voting
  • Heterogeneous Parallel Ensembles, Weighting Schemes, and Stacking Meta‑Learning
  • Adaptive Boosting (AdaBoost) - Sequential Weak Learner Fusion via Instance Weighting
  • Gradient Boosting and XGBoost - Residual‑Based Additive Modeling and Gradient‑Descent Optimization

• Perceptron

  • Biological Inspiration and Formal Definition of the Perceptron
  • Logical Function Representation with Perceptrons (m‑of‑n, AND, OR, XOR)
  • Batch Gradient Descent and the Perceptron Learning (Delta) Rule
  • Stochastic and Mini‑Batch Gradient Descent for Perceptron Training

• Feed-Forward Neural Network

  • Fundamentals, Motivation, and Application Domains of Artificial Neural Networks
  • Feed‑Forward Neural Network Architecture and Multi‑Layer Perceptron Design
  • Forward Propagation Mechanics in Feed‑Forward Neural Networks
  • Backpropagation Learning Algorithm and Gradient Computation for MLPs
  • Practical Implementation of Feed‑Forward Neural Networks with Keras and Parameter Accounting

• Kuis 1

  • Latihan Soal Ensemble Learning I
  • Solusi Kuis 1 2025

• Convolutional Neural Network

  • Motivation and Fundamental Concepts of CNNs
  • Local Connectivity and Parameter Sharing Mechanisms
  • Convolutional Layer Operations and Hyperparameters
  • CNN Architectural Design Patterns
  • Backpropagation in Convolutional Networks

UAS

• Recurrent Neural Network (RNN Pt. 1)

  • RNN Fundamentals and Sequential Data Motivation
  • RNN Architecture and Parameter Sharing
  • Dataset Preparation and FFNN vs RNN for Time-Series
  • Forward Propagation Worked Example in RNNs
  • RNN Implementation and Bidirectional RNNs

• Long Short-Term Memory (RNN Pt. 2)

  • LSTM Architecture - Cell State and Gates
  • Forward Propagation in LSTM Networks
  • Backpropagation Through Time and Cross Entropy Loss
  • BPTT for LSTM - Worked Example
  • Truncated BPTT and the Vanishing Gradient Problem
  • Encoder-Decoder Models in Sequence Learning

• Kuis 2

  • CNN - Kuis 2
  • RNN - Kuis 2
  • LSTM - Kuis 2

• Attention and Transformers

  • Attention Mechanism in RNNs
  • Encoder-Decoder with Attention - Bahdanau and Luong
  • From Attention to Transformers - Self-Attention
  • Transformer Architecture - Multi-Head Attention and Positional Encoding

• Reinforcement Learning

  • Reinforcement Learning Fundamentals and Motivation
  • Core Components of RL Agents
  • Categorization of RL Agent Architectures
  • Optimal Policy Formulation
  • Temporal Difference Learning Methods
  • Deep Q-Networks and Advanced Techniques

• Latihan & Persiapan UAS

  • Cheatsheet + Panduan Tracing UAS (RNN · LSTM · Attention · Transformer · RL)
  • Cheatsheet Konsep + Alur Hitung — Encoder-Decoder · Attention · Transformer · RL
  • Paket Soal Latihan UAS 01 · Pembahasan 01
  • Paket Soal Latihan UAS 02 · Pembahasan 02
  • Paket Soal Latihan UAS 03 · Pembahasan 03
  • Simulasi RNN · Attention · RL — alur penghitungan dinamis (buka di tab baru bila tampilan tidak muncul)