معجم مصطلحات التعلم الآلي والتعلم العميق وعلم البيانات

المحتويات

A 24
A/B testing 25
Accuracy 25
Action 26
Activation function 26
Active learning 28
AdaDelta 28
AdaGrad 28
Adaptive Learning Rate Algorithms 28
Adaptive Moment Estimation Algorithm (Adam) 29
Adversarial Learning Strategy 29
Agent 29
Agglomerative clustering 29
Anchor Boxes 29
Anomaly detection 29
Apache Spark 30
AR 30
Area under the PR curve 30
Area under the ROC curve 30
Artificial intelligence 30
Attention 30
Attribute 31
Attribute sampling 31
AUC (Area under the ROC curve) 31
Augmented reality 32
AutoEncoder 32
Automatic Differentiation 32
Auxiliary Loss 33
Average Pooling 33
Average precision 33
Axis-aligned condition 33
B 34
Backpropagation 35
Bagging 35
Bag of words (BOW) 36
Bar Chart 36
Baseline 37
Batch 37
Batch normalization 37
Batch size 37
Bayes Theorem 38
Bayesian neural network 38
Bayesian optimization 38
Bellman equation 39
BERT (Bidirectional Encoder Representations from Transformers) 39
Bias 39
Bias metric 40
Bias-Variance Trade-off 40
Big Data 41
Bigram 41
Bidirectional 41
Bidirectional language model 41
Binary classification 42
Binary Cross-Entropy 42
Binary condition 42
Binning 42
BLEU (Bilingual Evaluation Understudy) 43
Boosting 43
Bounding box 43
Broadcasting 43
Box Plot 44
Bucketing 45
C 46
Caffe 47
Calibration layer 47
Candidate generation 47
Candidate sampling 47
Categorical Cross-Entropy 48
Categorical data 48
Causal language model 48
CBOW 48
Centroid 49
Centroid-based clustering 49
Checkpoint 49
Class 49
Classifier 49
Classification 50
Classification model 50
Classification threshold 50
Class-imbalanced dataset 51
Clean Data 52
Clipping 52
Cloud TPU 52
Clustering 52
Co-adaptation 53
Colab 54
Collaborative filtering 54
Computer Vision 54
Condition 54
Confusion matrix 55
Continuous feature 56
Convenience sampling 56
Convergence 56
Convex function 57
Convex optimization 58
convex set 58
Convolution 58
Convolutional filter 59
Convolutional layer 59
Convolutional Neural Network (CNN) 60
Convolutional operation 60
Correlation 61
Correlation matrix 61
Corpus 61
Cost 62
Cost Function 62
Co-training 62
Covariance 62
Coverage bias 63
Crash blossom 63
Critic 63
Cross-entropy 63
Cross-validation 63
CUDA 63
D 64
Data analysis 65
Data augmentation 65
DataFrame 65
Data Mining 65
Data Pipeline 66
Dataset 66
Data Transformation 66
DBSCAN 66
Decision boundary 67
Decision forest 67
Decision Rules 67
Decision threshold 67
Decision tree 67
Deep model 68
Deep Learning 68
Dependent Variable 68
Decoder 69
Deep Neural Network 69
Deep Q-Network (DQN) 69
Denoising 69
Dense feature 70
Dense layer 70
Dense Network 70
Depth 70
Depthwise separable convolutional neural network (sepCNN) 71
Derived label 71
Device 71
Dimensions reduction 71
Dimensions 71
Discrete feature 72
Discriminative model 72
Discriminator 72
Divisive clustering 73
Dot Product 73
Downsampling 73
DQN 73
Dropout 73
Dropout regularization 74
Dummy Variable 74
Dynamic 74
Dynamic model 74
E 76
Edge detection 77
EigenValues 77
EigenVector 77
Early stopping 77
Embedding layer 77
Embedding vector 78
Emotion recognition 79
Empirical risk minimization (ERM) 79
Encoder 79
Ensemble 80
Entropy 80
Environment 81
Episode 81
Epoch 81
Epsilon greedy policy 82
Example 82
Experience replay 83
Experimenter's bias 83
Exploding gradient problem 83
exploratory data analysis (EDA) 84
Evaluation Metrics 84
F 85
False negative (FN) 86
False negative rate 86
False positive (FP) 86
False positive rate (FPR) 86
Feature 86
Feature cross 87
Feature engineering 88
Feature extraction 89
Feature importance 89
Feature Map 89
Feature Reduction 89
Feature Selection 90
Feature set 90
Feature vector 90
Federated Learning 92
Feedback loop 92
Feedforward neural network (FFN) 92
Filter 92
Few-shot learning 93
Fine tuning 93
Flatten layer 93
Forget gate 93
Forward propagation 93
Full softmax 94
Fully connected layer 94
F-Score 94
G 95
Gated Recurrent Unit (GRU) 96
Generalization 96
Generalization curve 96
Generalized linear model 97
Generative adversarial network (GAN) 97
Generative model 98
Generative Pre-trained Transformer (GPT) 98
Generator 98
Gini impurity 99
Goodness of Fit 99
GoogLeNet 99
Gradient 99
Gradient boosting 100
Gradient boosted (decision) trees (GBT) 100
Gradient clipping 100
Gradient descent 100
Graph 100
Greedy policy 101
Ground truth 101
H 102
Hadoop 103
Hard Margin 103
Hallucination 103
Hashing 103
Heuristic 104
Hidden layer 104
Hierarchical clustering 104
Hinge loss 105
Histogram 105
Holdout data 106
Hyperparameter 106
Hyperplane 106
I 108
Image recognition 109
Imbalanced dataset 109
Implicit bias 109
Imputation 109
Inception Network 110
Independently and identically distributed (i.i.d) 110
Inference 110
Inference path 111
Information gain 111
Input layer 112
Instance 112
Interpretability 112
Intersection over union (IoU) 113
IRIS Dataset 114
Item matrix 114
Items 114
iteration 114
J 116
Jupyter Notebook 117
K 118
Keras 119
Kernel 119
Kernel Support Vector Machines (KSVMs) 119
k-means 119
k-median 120
k-nearest neighbors(KNN) 121
L 122
L0 regularization 123
L1 loss 123
L1 regularization 123
L2 loss 124
L2 regularization 124
Label 125
LaMDA (Language Model for Dialogue Applications) 125
lambda 125
landmarks 125
language model 125
Large language model 125
Lasso Regression 126
Layer 126
Leaf 127
Learning rate 127
Least squares regression 128
Lemmatization 128
Line Chart 128
Linear model 128
Linear 129
Linear regression 129
Local Optima 129
Logistic regression 130
Logits 131
Log Loss 131
Long Short-Term Memory (LSTM) 131
Loss 131
Loss curve 131
Loss function 132
Loss surface 133
M 134
Machine learning 135
Machine Translation 135
Majority class 135
Markov decision process (MDP) 135
Markov property 135
Mask R-CNN 136
Masked language model 136
matplotlib 136
Matrix factorization 136
Matrix Multiplication 136
Max Pooling 136
Mean 137
Median 137
Mean Absolute Error (MAE) 137
Mean Squared Error (MSE) 138
Metric 139
Mini-batch 139
Mini-batch stochastic gradient descent 139
Minimax loss 140
Minority Class 140
Missing value 140
ML 140
MNIST 140
MobileNet 141
Mode 141
Modality 141
Model 141
Model deployment 142
Model capacity 142
Model parallelism 142
Model Selection 142
Model training 143
Model Validation 143
Momentum 143
Multi-class classification 143
Multi-class logistic regression 144
Multi-head self-attention 144
Multimodal model 144
Multinomial classification 145
Multinomial regression 145
Multivariate Analysis 145
Multivariate Regression 145
MxNet 145
N 146
Naive Bayes 147
Named entity recognition (NER) 147
NaN 147
NaN trap 147
Natural Language Processing (NLP) 147
Natural Language Toolkit (NLTK) 148
Natural language understanding 148
Negative class 148
Neural network 148
Neuron 149
NLU 150
N-gram 150
Node (neural network) 151
Node (decision tree) 151
Noise 151
Nominal Variable 151
Non-binary condition 152
Nonlinear 152
Nonstationarity 152
Normal Distribution 153
Normalization 153
Novelty detection 154
Numerical data 154
NumPy 154
O 155
Object Detection 156
Objective 156
Objective function 156
Oblique condition 156
Observation 157
Offline 157
Offline inference 157
One-hot encoding 157
One-shot learning 158
One-vs.-all 159
Online 159
Online inference 159
OpenCv 159
Optimizer 159
Out-of-bag evaluation (OOB evaluation) 160
Outlier detection 160
Outliers 161
Output layer 161
Overfitting 162
Oversampling 162
P 164
Padding 165
Pandas 165
Parameter 165
Parameter update 166
Partial derivative 166
Pattern Recognition 166
Perceptron 166
Permutation variable importance 167
Perplexity 167
Pie Chart 167
Pipeline 168
Pipelining 168
Policy 168
Polynomial Regression 168
Pooling 169
Positive class 170
PR AUC (area under the PR curve) 170
Precision 170
Precision-recall curve 171
Prediction 171
Prediction bias 171
Pre-trained model 171
Principal Component Analysis (PCA) 172
Prior belief 172
Probabilistic regression model 172
Proxy labels 172
Pruning 173
P-Value 173
Python 173
Pytorch 173
Q 175
Q-function 176
Q-learning 176
Quantile 176
Quantile bucketing 176
Quartile 177
Quantization 177
R 178
Random forest 179
Random policy 179
Ranking 179
Rank (ordinality) 179
Rater 179
Recall 179
Recommendation Engine 181
Recommendation system 181
Rectified Linear Unit (ReLU) 181
Recurrent neural networks (RNN) 182
Region-Based Convolutional Neural Network (R-CNN) 183
Regression model 183
Regularization 184
Regularization rate 184
Reinforcement learning (RL) 185
ReLU 185
Replay buffer 185
Representation 185
Re-ranking 185
Residual Network (ResNet) 186
Return 186
Reward 186
Ridge regularization 186
RMSPorp 187
RNN 187
ROC (receiver operating characteristic) Curve 187
Root 188
Root Mean Squared Error (RMSE) 189
Rotational invariance 189
S 190
Scalar 191
Scalar Multiplication 191
Scaling 191
scikit-learn 191
SciPy 192
Self-attention (also called self-attention layer) 192
Self-supervised learning 193
Self-training 193
Segmentation 194
Semantic Segmentation 194
Semi-supervised learning 194
Sentiment analysis 194
Sensitivity 195
Seq2Seq Model 195
Sequence model 195
Sequence-to-sequence task 195
Shape (Tensor) 195
Skewness 196
Shrinkage 196
Sigmoid 196
Skip-Gram 197
SKLearn 197
Sliding-Window 197
Similarity measure 197
Size invariance 197
Sketching 198
SMOTE 198
Soft Margin 198
Softmax 199
Sparse feature 199
Sparse representation 199
Sparse vector 200
Sparsity 200
Spatial pooling 200
Specificity 200
Spectogram 201
splitter 201
Squared hinge loss 201
Squared loss 201
Staged training 201
Standard Deviation 202
Standardization 202
State 203
State-action value function 203
Static 203
Static inference 203
Stationarity 203
Stemming 204
Step 204
Step Size 204
Stochastic gradient descent (SGD) 204
Stride 204
Structural risk minimization (SRM) 205
Style Transfer 205
Subsampling 205
Supervised machine learning 206
Support Vector Machine (SVM) 206
Synthetic feature 206
T 208
Tabular Q-learning 209
Tanh 209
Target 209
Target network 209
Temporal data 209
Tensor 209
TensorBoard 210
TensorFlow 210
TensorFlow Playground 210
TensorFlow Serving 210
Tensor Processing Unit (TPU) 210
Tensor shape 210
Tensor size 210
Termination condition 210
Test 211
Test loss 211
Test set 211
Text Mining 212
TF-IDF 212
tf.keras 212
Threshold (for decision trees) 212
Time series analysis 212
Timestep 213
Token 213
Tokenization 214
Torch 214
Tower 214
TPU 214
Training 214
Training loss 214
Training-serving skew 215
Training set 215
Trajectory 216
Transfer learning 216
Transformer 216
Translational invariance 217
Trigram 217
True positive (TP) 217
True positive (TP) 217
True positive rate (TPR) 218
Type 1 Error 218
Type 2 Error 218
U 219
Underfitting 220
Undersampling 220
U-net 220
Unidirectional 221
Unidirectional language model 221
Unlabeled example 221
Unsupervised machine learning 221
Uplift modeling 222
upweighting 222
User matrix 223
V 224
Validation 225
Validation loss 225
Validation set 225
Vanishing gradient problem 226
variable importance 226
Variance 226
VGG Network 227
W 228
Weight 229
Weighted sum 229
Wide model 229
Width 230
wisdom of the crowd 230
Word embedding 230
Word2vec 230