🤖 Artificial Intelligence & Machine Learning Terminology
Comprehensive glossary of AI, Machine Learning, Data Science, and related emerging technology terms.
🧠 Core AI/ML Concepts
| Istilah |
Definisi |
| Algorithm |
Sekumpulan aturan atau instruksi yang diberikan kepada AI untuk membantu memecahkan masalah |
| Artificial Intelligence (AI) |
Simulasi proses kecerdasan manusia oleh mesin atau sistem komputer |
| Machine Learning (ML) |
Subset dari AI yang memungkinkan mesin belajar dan berkembang dari pengalaman tanpa pemrograman eksplisit |
| Deep Learning |
Subset dari ML yang menggunakan neural network dengan banyak lapisan untuk menganalisis data |
| Neural Network |
Sistem komputasi yang terinspirasi dari jaringan neural biologis yang membentuk otak hewan |
| Natural Language Processing (NLP) |
AI capability yang enables computers untuk understand, interpret, dan generate human language |
| Computer Vision |
Field of AI yang trains computers untuk interpret dan understand visual world |
| Reinforcement Learning |
Type of ML dimana agent learns untuk make decisions melalui trial dan error |
| Supervised Learning |
ML approach dimana algorithm learns dari labeled training data |
| Unsupervised Learning |
ML technique dimana algorithm finds patterns dalam data tanpa labeled examples |
| Semi-supervised Learning |
Combination of supervised dan unsupervised learning techniques |
| Transfer Learning |
Technique dimana model trained pada satu task di-reuse sebagai starting point untuk another task |
📊 Data Science & Analytics
| Istilah |
Definisi |
| Big Data |
Large, complex datasets yang require specialized tools untuk process dan analyze |
| Data Mining |
Process of discovering patterns dan knowledge dari large amounts of data |
| Data Pipeline |
Series of data processing steps dimana output dari one step adalah input untuk next |
| Feature Engineering |
Process of selecting dan transforming variables untuk improve model performance |
| Feature Selection |
Process of selecting subset of relevant features untuk model construction |
| Data Preprocessing |
Technique untuk prepare raw data untuk machine learning algorithms |
| Dimensionality Reduction |
Process of reducing number of features dalam dataset while preserving information |
| Cross-validation |
Statistical method untuk evaluate machine learning model performance |
| Overfitting |
Modeling error yang occurs ketika function fits training data too closely |
| Underfitting |
Occurs ketika model cannot capture underlying trend of data |
| Bias-Variance Tradeoff |
Balance between model's ability untuk minimize bias dan variance dalam predictions |
🔮 Model Types & Architectures
| Istilah |
Definisi |
| Convolutional Neural Network (CNN) |
Deep learning architecture particularly effective untuk image recognition tasks |
| Recurrent Neural Network (RNN) |
Neural network architecture designed untuk sequential data processing |
| Long Short-Term Memory (LSTM) |
Type of RNN architecture yang can learn long-term dependencies |
| Transformer |
Neural network architecture yang relies entirely pada attention mechanisms |
| Generative Adversarial Network (GAN) |
Architecture dengan two neural networks competing untuk generate realistic data |
| Autoencoder |
Neural network trained untuk copy input ke output untuk dimensionality reduction |
| Decision Tree |
Tree-like model untuk decision making berdasarkan feature values |
| Random Forest |
Ensemble method yang combines multiple decision trees |
| Support Vector Machine (SVM) |
Supervised learning model untuk classification dan regression |
| K-Means Clustering |
Unsupervised algorithm untuk grouping data into k clusters |
| Linear Regression |
Statistical method untuk modeling relationship between dependent dan independent variables |
🎯 Training & Optimization
| Istilah |
Definisi |
| Gradient Descent |
Optimization algorithm untuk finding minimum of function |
| Backpropagation |
Algorithm untuk training neural networks dengan computing gradients |
| Learning Rate |
Hyperparameter yang controls how much model adjusts weights during training |
| Epoch |
One complete pass through entire training dataset |
| Batch Size |
Number of training examples processed before model weights are updated |
| Hyperparameter Tuning |
Process of finding optimal hyperparameters untuk machine learning model |
| Regularization |
Technique untuk prevent overfitting dengan adding penalty untuk complexity |
| Dropout |
Regularization technique dimana randomly selected neurons are ignored during training |
| Early Stopping |
Technique untuk stop training ketika performance pada validation set stops improving |
| Loss Function |
Function yang measures difference between predicted dan actual values |
| Activation Function |
Function dalam neural network yang determines output of neuron |
| Istilah |
Definisi |
| Accuracy |
Percentage of correct predictions made oleh model |
| Precision |
Ratio of correctly predicted positive observations to total predicted positives |
| Recall (Sensitivity) |
Ratio of correctly predicted positive observations to all actual positives |
| F1-Score |
Weighted average of precision dan recall |
| ROC Curve |
Graph showing performance of classification model at all classification thresholds |
| AUC (Area Under Curve) |
Measure of ability of classifier untuk distinguish between classes |
| Confusion Matrix |
Table untuk describing performance of classification model |
| Mean Squared Error (MSE) |
Average of squares of errors between actual dan predicted values |
| R-squared |
Statistical measure representing proportion of variance explained oleh model |
| Cross-entropy Loss |
Loss function commonly used untuk classification problems |
🚀 Generative AI & LLMs
| Istilah |
Definisi |
| Large Language Model (LLM) |
AI model trained pada large amounts of text data untuk understand dan generate language |
| Generative Pre-trained Transformer (GPT) |
Type of LLM architecture untuk generating human-like text |
| Prompt Engineering |
Practice of designing inputs untuk get desired outputs dari AI models |
| Fine-tuning |
Process of adapting pre-trained model untuk specific task atau domain |
| BERT |
Bidirectional Encoder Representations dari Transformers - model untuk understanding language |
| Tokenization |
Process of breaking text into individual tokens atau words |
| Embedding |
Dense vector representation of words atau phrases dalam continuous space |
| Attention Mechanism |
Technique yang allows model untuk focus pada specific parts of input |
| Transformer Architecture |
Neural network architecture yang uses self-attention mechanisms |
| Zero-shot Learning |
Model's ability untuk perform task tanpa specific training examples |
| Few-shot Learning |
Learning approach dengan only few examples per class |
🎨 AI Applications & Domains
| Istilah |
Definisi |
| Computer Vision |
AI field yang enables computers untuk derive meaningful information dari images |
| Natural Language Generation |
AI capability untuk produce human-like text dari structured data |
| Speech Recognition |
Technology yang converts spoken language into text |
| Recommendation System |
Algorithm yang suggests relevant items kepada users berdasarkan preferences |
| Autonomous Systems |
Systems yang can perform tasks tanpa human intervention |
| Robotics |
Field combining AI dengan mechanical engineering untuk create intelligent machines |
| Chatbot |
AI program designed untuk simulate conversation dengan human users |
| Virtual Assistant |
AI-powered software agent yang can perform tasks atau services untuk user |
| Predictive Analytics |
Use of data untuk predict future outcomes atau trends |
| Anomaly Detection |
Identification of rare items yang significantly differ dari majority of data |
| Istilah |
Definisi |
| TensorFlow |
Open-source machine learning framework developed oleh Google |
| PyTorch |
Open-source machine learning library developed oleh Meta/Facebook |
| Scikit-learn |
Machine learning library untuk Python programming language |
| Keras |
High-level neural networks API running on top of TensorFlow |
| Pandas |
Data manipulation dan analysis library untuk Python |
| NumPy |
Library untuk scientific computing dengan Python |
| Jupyter Notebook |
Interactive computing environment untuk creating dan sharing documents |
| Docker |
Platform untuk developing, shipping, dan running applications dalam containers |
| Kubernetes |
Container orchestration system untuk automating deployment dan management |
| MLflow |
Open-source platform untuk managing machine learning lifecycle |
| Apache Spark |
Unified analytics engine untuk large-scale data processing |
🏢 AI Ethics & Governance
| Istilah |
Definisi |
| AI Ethics |
Moral principles dan values yang guide development dan deployment of AI systems |
| Algorithmic Bias |
Systematic errors dalam algorithm yang create unfair outcomes |
| Explainable AI (XAI) |
AI systems yang provide understandable explanations untuk their decisions |
| Fairness |
Principle ensuring AI systems don't discriminate against certain groups |
| Privacy-Preserving ML |
Techniques untuk training models while protecting sensitive data |
| Federated Learning |
Decentralized approach untuk training models tanpa centralizing data |
| Model Interpretability |
Degree untuk which human dapat understand cause untuk model's decision |
| Responsible AI |
Approach untuk developing AI systems yang are ethical, transparent, dan accountable |
| AI Governance |
Framework untuk ensuring ethical development dan deployment of AI |
| Data Sovereignty |
Legal concept regarding ownership dan control of digital data |
🌊 Emerging AI Trends
| Istilah |
Definisi |
| Artificial General Intelligence (AGI) |
Hypothetical AI yang matches atau exceeds human cognitive abilities |
| Multimodal AI |
AI systems yang can process dan understand multiple types of data simultaneously |
| Edge AI |
Deployment of AI algorithms locally pada hardware device |
| Quantum Machine Learning |
Integration of quantum computing dengan machine learning algorithms |
| Neuromorphic Computing |
Computer architecture inspired oleh human brain's neural networks |
| AutoML |
Automated process untuk applying machine learning untuk real-world problems |
| AI as a Service (AIaaS) |
Cloud-based service yang provides AI capabilities through APIs |
| Synthetic Data |
Artificially generated data untuk training machine learning models |
| Digital Twin |
Virtual representation of physical object atau system |
| Swarm Intelligence |
Collective behavior of decentralized systems |
Terakhir diperbarui: August 7, 2025