Index
A
- active learning / Cold start problem
- adaptable user experience
- building / Building an adaptable user experience
- affective computing / Understanding users emotions
- AI APIs / Machine learning and AI APIs
- AIToolbox / AIToolbox
- algorithmic design / Building an adaptable user experience
- Amazon Machine Learning
- reference link / Machine learning and AI APIs
- Analytics Edge
- reference link / Machine learning
- anticipatory design / Building an adaptable user experience
- Apple Developer site
- reference link / Delivering perfect user experience
- Apple lzfse compression library
- reference link / Lossless compression
- Apriori algorithm
- about / The Apriori algorithm
- implementing, in Swift / Implementing Apriori in Swift
- executing / Running Apriori
- executing, on real-world data / Running Apriori on real-world data
- cons / The pros and cons of Apriori
- pros / The pros and cons of Apriori
- artificial general intelligence (AGI) / What is AI?
- artificial NNs / What are artificial NNs anyway?
- association measures
- used, to assess rules / Using association measures to assess rules
- supporting association measures / Supporting association measures
- confidence association measures / Confidence association measures
- lift association measures / Lift association measures
- conviction association measures / Conviction association measures
- association rule learning / Unsupervised learning, Seeing association rules
- association rule mining / Seeing association rules
- association rules
- displaying / Seeing association rules
- issues, decomposing / Decomposing the problem
- generating / Generating all possible rules
- autoencoder / Autoencoder neural networks
- autoencoder neural networks / Autoencoder neural networks
B
- backpropagation
- reference link / Training the network
- about / Other approaches
- basic neural network subroutines (BNNS)
- about / Basic neural network subroutines (BNNS)
- reference link / Basic neural network subroutines (BNNS)
- batch normalization / Batch normalization
- black magic / Building an adaptable user experience
- BrainCore / BrainCore
C
- Caffe / Caffe
- Caffe2 / Caffe2
- CAP 5415 Computer Vision
- reference link / Computer vision
- categorical variables
- converting / Converting categorical variables
- centroid / K-means clustering
- classification task
- revisiting / Revisiting the classification task
- linear classifier / Linear classifier
- logistic regression / Logistic regression
- clusters
- count, selecting / Choosing the number of clusters
- CNN architectures
- about / Compact CNN architectures
- SqueezeNet / SqueezeNet
- MobileNets / MobileNets
- ShuffleNet / ShuffleNet
- CondenseNet / CondenseNet
- cold start / Cold start problem
- combinatorial entropy / Combinatorial entropy
- computer vision
- issues / Introducing computer vision problems
- object recognition / Introducing computer vision problems
- object localization / Introducing computer vision problems
- object detection / Introducing computer vision problems
- semantic segmentation / Introducing computer vision problems
- instance segmentation / Introducing computer vision problems
- pose estimation / Introducing computer vision problems
- object tracking / Introducing computer vision problems
- image segmentation / Introducing computer vision problems
- Computer Vision, Algorithms and Application
- reference link / Computer vision
- computer vision domain / Introducing computer vision problems
- CondenseNet / CondenseNet
- confidence association measures / Confidence association measures
- confusion matrix
- calculating / Calculating the confusion matrix
- console / Chatbot anatomy
- conviction association measures / Conviction association measures
- convolution
- in CNNs / Convolutions in CNNs
- convolutional layer / Convolutional layer
- convolutional neural network (CNN)
- about / Introducing convolutional neural networks, General-purpose machine learning libraries
- training, for facial expression recognition / Training the CNN for facial expression recognition
- deploying, to iOS / Deploying CNN to iOS
- convolution filter
- about / Convolution operation
- visualizing / Visualizing convolution filters
- convolution operation / Convolution operation
- Core ML
- URL / Tools
- about / Introducing Core ML
- features / Core ML features
- model, exporting for iOS / Exporting the model for iOS
- random forest, ensembling / Ensemble learning random forest
- random forest, training / Training the random forest
- random forest accuracy, evaluation / Random forest accuracy evaluation
- model, importing into iOS project / Importing the Core ML model into an iOS project
- model performance, evaluating on iOS / Evaluating performance of the model on iOS
- confusion matrix, calculating / Calculating the confusion matrix
- decision tree learning, pros and cons / Decision tree learning pros and cons
- core ML format
- converting / Converting to Core ML format
- Core Motion API
- CS231n Convolutional Neural Networks for Visual Recognition
- reference link / Computer vision
- curse of dimensionality / Reasoning in high-dimensional spaces
D
- data
- loading / Loading the data
- splitting / Splitting the data
- preparing, for creating predictions / Making predictions
- data augmentation / Getting the dataset, Data augmentation
- data kobolds
- about / Data kobolds
- tough data / Tough data
- biased data / Biased data
- batch effects / Batch effects
- Data preprocessing
- categorical variables, converting / Converting categorical variables
- data preprocessing
- about / Data preprocessing, Data preprocessing
- categorical variables, converting / Converting categorical variables
- features, separating from labels / Separating features from labels
- one-hot encoding / One-hot encoding
- data, splitting / Splitting the data
- dataset
- obtaining / Getting the dataset
- references / Getting the dataset
- loading / Loading the dataset
- URL, for downloading / Training the CNN for facial expression recognition
- training set / Model training, evaluation, and selection
- test set / Model training, evaluation, and selection
- validation set / Model training, evaluation, and selection
- data structures
- defining / Defining data structures
- decision boundary / Understanding the KNN algorithm, Linear classifier
- decision tree classifier
- about / Decision trees everywhere
- training / Training the decision tree classifier
- URL / Training the decision tree classifier
- tree visualization / Tree visualization
- predictions, creating / Making predictions
- accuracy, evaluating / Evaluating accuracy
- hyperparameters, tuning / Tuning hyperparameters
- model capacity trade-offs / Understanding model capacity trade-offs
- decision tree learning
- working / How decision tree learning works
- tree, building from data / Building a tree automatically from data
- combinatorial entropy / Combinatorial entropy
- URL / Combinatorial entropy
- performance, evaluating of model with data / Evaluating performance of the model with data
- precision / Precision, recall, and F1-score
- F1-score / Precision, recall, and F1-score
- recall / Precision, recall, and F1-score
- sensitivity / Precision, recall, and F1-score
- K-fold cross-validation / K-fold cross-validation
- confusion matrix / Confusion matrix
- pros and cons / Decision tree learning pros and cons
- decoder network / Autoencoder neural networks
- deep learning
- about / Choosing a good set of features
- reference link / Machine learning
- Deep learning (DL) / General-purpose machine learning libraries
- deep learning frameworks
- about / Deep learning frameworks
- Keras / Keras
- differential privacy
- URL, for documentation / Privacy and differential privacy
- dimensionality / Choosing a good set of features
- dimensionality reduction / Unsupervised learning
- dimensionality reduction algorithms / Choosing a good set of features
- distance
- calculating / Calculating the distance
- DTW / DTW
- DTW, implementing in Swift / Implementing DTW in Swift
- distributional semantics / Distributional semantics hypothesis
- distributional semantics hypothesis / Distributional semantics hypothesis
- dlib / dlib
- drawbacks, rectified linear unit (ReLU), drawbacks
- Leaky ReLU / Rectifier-like activation functions
- Randomized ReLU / Rectifier-like activation functions
- Parametric ReLU (PReLU) / Rectifier-like activation functions
- Maxout unit / Rectifier-like activation functions
- Softplus / Rectifier-like activation functions
- dropout / Dropout
- DTW distance
- about / DTW
- implementing, in Swift / Implementing DTW in Swift
- dummy variables / One-hot encoding
E
- ElasticNet regression / ElasticNet regression
- encoder / Autoencoder neural networks
- environment
- setting up / Environment setup
- Euclidean distance / Calculating the distance
- exploratory data analysis / Exploratory data analysis
F
- Facebook's approach in Caffe 2
- reference link / Facebook's approach in Caffe2
- about / Facebook's approach in Caffe2
- Facebook zstd compression library
- reference link / Lossless compression
- false negative / Evaluating performance of the model with data
- false positive / Evaluating performance of the model with data
- Fast Artificial Neural Network (FANN) / FANN
- feature/data normalization / Feature scaling
- feature engineering / Choosing a good set of features
- feature extraction / Choosing a good set of features
- feature map / Convolutional layer
- features
- about / Features
- types / Types of features
- selecting / Choosing a good set of features
- separating, from labels / Separating features from labels
- feature scaling / Feature scaling
- feature selection / Choosing a good set of features
- feature space / Understanding the KNN algorithm
- feature standardization
- about / Feature standardization
- multiple linear regression / Multiple linear regression
- Feed-forward (FF) / General-purpose machine learning libraries
- feed-forward neural network / Building the network
- fully-connected layer / Fully-connected layers
- fully-connected neural network / Building the network
G
- garbage in, garbage out (GIGO) / Data preprocessing
- Gaussian mixture model (GMM) / General-purpose machine learning libraries
- general-purpose machine learning libraries
- about / General-purpose machine learning libraries
- AITOOLbox / AIToolbox
- BrainCore / BrainCore
- Caffe / Caffe
- Caffe2 / Caffe2
- dlib / dlib
- Fast Artificial Neural Network (FANN) / FANN
- LearnKit / LearnKit
- MLKit / MLKit
- multilinear-math / Multilinear-math
- MXNet / MXNet
- Shark / Shark
- TensorFlow / TensorFlow
- tiny-dnn / tiny-dnn
- Torch / Torch
- YCML / YCML
- Google Cloud Platform
- reference link / Machine learning and AI APIs
- gradient descent
- using / Where to use GD and normal equation
- about / Where to use GD and normal equation
- used, for function minimization / Using gradient descent for function minimization
- for multiple linear regression / Gradient descent for multiple linear regression
- Graphviz
- URL / Tools
H
- Hamming distance / Calculating the distance
- HDF5 format
- model, saving / Saving the model in HDF5 format
- hidden Markov model (HMM) / Utilizing state transitions
- high-dimensional spaces
- reasoning / Reasoning in high-dimensional spaces
- High-Performance Swift Code
- reference link / Porting or deployment for a mobile platform
- Homebrew
- human motion
- recognizing, with KNN / Recognizing human motion using KNN
- Hyperbolic tangent (tanh) / Step-like activation functions
- hyperparameters
- about / Training the decision tree classifier
- tuning / Tuning hyperparameters
I
- IBM Watson
- reference link / Machine learning and AI APIs
- image segmentation
- used, for k-means clustering / Image segmentation using k-means
- imbalanced / Evaluating performance of the model with data
- Immersive Linear Algebra
- reference link / Mathematical background
- inertial sensors
- using, for people motion recognition / People motion recognition using inertial sensors
- inference-only libraries
- about / Inference-only libraries
- Keras / Keras
- LibSVM / LibSVM
- Scikit-learn / Scikit-learn
- XGBoost / XGBoost
- Information Gain (IG) / Combinatorial entropy
- input layer / Input layer
- instance-based models
- used, for clustering / Using instance-based models for classification and clustering
- used, for classification / Using instance-based models for classification and clustering
- internal covariate shift / Batch normalization
- iOS application
- about / iOS application, Putting it all together
- chatbot anatomy / Chatbot anatomy
- voice input / Voice input
- NSLinguisticTagger / NSLinguisticTagger and friends
- Word2Vec / Word2Vec on iOS
- text-to-speech output / Text-to-speech output
- UIReferenceLibraryViewController / UIReferenceLibraryViewController
- iOS motion sensors
- gyroscope / Recognizing human motion using KNN
- accelerometer / Recognizing human motion using KNN
- magnetometer / Recognizing human motion using KNN
- compass / Recognizing human motion using KNN
- iOS project
- Core ML model, importing / Importing the Core ML model into an iOS project
- IPython notebook / IPython notebook crash course
- item sets
- finding / Finding frequent item sets
J
- Jupyter
- URL / Tools
K
- k-means++ / K-means++
- k-means clustering
- about / K-means clustering
- implementing, in Swift / Implementing k-means in Swift
- update step / Update step
- assignment step / Assignment step
- issues / K-means clustering – problems
- used, for image segmentation / Image segmentation using k-means
- reference / Image segmentation using k-means
- k-nearest neighbors (KNN) / General-purpose machine learning libraries
- Keras
- KNN algorithm
- about / Understanding the KNN algorithm
- implementing, in Swift / Implementing KNN in Swift
- used, for recognizing human motion / Recognizing human motion using KNN
- cold start issues / Cold start problem
- balanced dataset / Balanced dataset
- k, selecting / Choosing a good k
- pros / KNN pros
- cons / KNN cons
- enhancing / Improving our solution
- probabilistic interpretation / Probabilistic interpretation
- data sources / More data sources
- time series chunking / Smarter time series chunking
- hardware acceleration / Hardware acceleration
- trees, used for speeding up inference / Trees to speed up the inference
- state transitions, utilizing / Utilizing state transitions
L
- label / Revisiting the classification task
- LASSO regression / LASSO regression
- learning resources / Recommended learning resources
- LearnKit / LearnKit
- Least Absolute Shrinkage and Selection Operator (LASSO) / LASSO regression
- least squares method
- used, for fitting regression line / Fitting a regression line using the least squares method
- lexical tokens / Tokenization
- libraries / Libraries
- LibSVM
- lift association measures / Lift association measures
- linear classifier / Linear classifier
- linear regression
- limitations, overcoming / Understanding and overcoming the limitations of linear regression
- issues, fixing with regularization techniques / Fixing linear regression problems with regularization
- tikhonov regularization / Ridge regression and Tikhonov regularization
- ridge regression / Ridge regression and Tikhonov regularization
- ElasticNet regression / ElasticNet regression
- Linear regression (LinReg) / General-purpose machine learning libraries
- logical functions
- building, with neurons / Using neurons to build logical functions
- logistic (sigmoid) function / Step-like activation functions
- logistic regression
- about / Logistic regression
- implementing, in Swift / Implementing logistic regression in Swift
- implementing / The prediction part of logistic regression
- training / Training the logistic regression
- cost function / Cost function
- Logistic regression (LogReg) / General-purpose machine learning libraries
- Long short-term memory (LSTM) / General-purpose machine learning libraries
- loss function / Mathematical optimization – how learning works, Loss functions
- lossless compression / Lossless compression
- loss values
- plotting / Plotting loss
- lossy compression
- about / Lossy compression
- tools / Tools
M
- machine learning
- about / The motivation behind ML, What is ML ?, Machine learning and AI APIs
- applications / Applications of ML, Other applications of ML
- Digital signal processing (DSP) / Digital signal processing (DSP)
- computer vision / Computer vision
- Natural language processing (NLP) / Natural language processing (NLP)
- used, for building iOS applications / Using ML to build smarter iOS applications
- implementing / Time to practice
- using, for extra-terrestrial life explorers / Machine learning for extra-terrestrial life explorers
- best practices / Best practices
- benchmarking / Benchmarking
- privacy / Privacy and differential privacy
- differential privacy / Privacy and differential privacy
- debugging / Debugging and visualization
- visualization / Debugging and visualization
- documentation / Documentation
- training / Goblins of training
- product design ogres / Product design ogres
- machine learning app
- prototyping / Prototyping the first machine learning app
- tools / Tools
- environment, setting up / Setting up a machine learning environment
- implementing, in Swift / Implementing first machine learning app in Swift
- machine learning gremlins
- about / Machine learning gremlins
- data kobolds / Data kobolds
- Manhattan distance / Calculating the distance
- map
- objects, clustering / Clustering objects on a map
- Markov decision process (MDP) / General-purpose machine learning libraries
- mathematical background, learning resources
- about / Mathematical background
- machine learning / Machine learning
- computer vision / Computer vision
- NLP / NLP
- mathematical optimization / Mathematical optimization – how learning works
- metrics / Evaluating performance of the model with data
- Microsoft Azure Machine Learning
- reference link / Machine learning and AI APIs
- Microsoft Cognitive Services
- reference link / Machine learning and AI APIs
- mirror neurons / The motivation behind ML
- MLKit / MLKit
- mobile machine learning
- versus server-side machine learning / Mobile versus server-side ML
- life cycle / Mobile machine learning project life cycle
- preparatory stage / Preparatory stage
- prototype creation / Prototype creation
- mobile platform, porting / Porting or deployment for a mobile platform
- mobile platform, deployment / Porting or deployment for a mobile platform
- production / Production
- MobileNet / MobileNets
- model
- selecting / Choosing a model
- machine learning algorithms, types / Types of ML algorithms
- supervised learning / Supervised learning, Unsupervised learning
- reinforcement learning / Reinforcement learning
- mathematical optimization / Mathematical optimization – how learning works
- mobile, versus server-side machine learning / Mobile versus server-side ML
- mobile platform limitations / Understanding mobile platform limitations
- saving, in HDF5 format / Saving the model in HDF5 format
- model capacity
- trade-offs / Understanding model capacity trade-offs
- multilayer perceptron (MLP) / Building the network
- multiple linear regression
- about / Multiple linear regression
- implementing, in Swift / Implementing multiple linear regression in Swift
- gradient descent / Gradient descent for multiple linear regression
- training / Training multiple regression
- linear algebraic operations / Linear algebra operations
- feature-wise standardization / Feature-wise standardization
- equation / Normal equation for multiple linear regression
- MXNet / MXNet
N
- named entity recognition (NER) / Named entity recognition (NER)
- Natural language processing (NLP)
- about / Natural language processing (NLP)
- in mobile development / NLP in the mobile development world
- approaches / Common NLP approaches and subtasks
- subtasks / Common NLP approaches and subtasks
- tokenization / Tokenization
- stemming / Stemming
- lemmatization / Lemmatization
- Part-of-speech (POS) tagging / Part-of-speech (POS) tagging
- named entity recognition (NER) / Named entity recognition (NER)
- stop words, removing / Removing stop words and punctuation
- punctuation, removing / Removing stop words and punctuation
- network
- building / Building the network, Building the network
- input layer / Input layer
- convolutional layer / Convolutional layer
- fully-connected layer / Fully-connected layers
- nonlinearity layer / Nonlinearity layers
- pooling layer / Pooling layer
- regularization layer / Regularization layers
- training / Training the network, Training the network
- creating / Creating the network
- network compression
- example / An example of the network compression
- network structure
- plotting / Plotting the network structure
- neural layer
- building / Building a neural layer in Swift
- neural network (NN)
- about / General-purpose machine learning libraries
- preventing, from growing / Preventing a neural network from growing big
- neuron
- building / Building the neuron
- non-linearity function / Non-linearity function
- used, for building logical functions / Using neurons to build logical functions
- new trends / Building an adaptable user experience
- NLP libraries
- noise / Data preprocessing
- non-generalizing machine learning / Understanding the KNN algorithm
- non-linearity function
- about / Non-linearity function
- step-like activation functions / Step-like activation functions
- non-parametric / Understanding the KNN algorithm
- nonlinearity layer / Nonlinearity layers
- numerical optimization
O
- objects
- clustering, on map / Clustering objects on a map
- one-hot encoding / One-hot encoding
- online machine learning / Getting the dataset
- operation
- pooling / Pooling operation
- Optical character recognition (OCR) / Computer vision
- optimizing for inference
- network pruning / Network pruning
- precision, reducing / Reducing precision
- approaches / Other approaches
- outlier/anomaly detection / Unsupervised learning
P
- parametric models / Understanding the KNN algorithm
- people motion recognition
- inertial sensors, using / People motion recognition using inertial sensors
- PEP 8
- perfect user experience
- delivering / Delivering perfect user experience
- pooling layer / Pooling layer
- precision / Precision, recall, and F1-score
- preparatory stage
- about / Preparatory stage
- issues, formulating / Formulate the problem
- constraints, defining / Define the constraints
- existing approaches, research / Research the existing approaches
- data, research / Research the data
- design choices, creating / Make design choices
- Principal components analysis (PCA) / General-purpose machine learning libraries
- product design ogres, machine learning gremlins
- data, preparing / Magical thinking
- technologies / Cargo cult
- feedback loops / Feedback loops
- uncanny valley effect / Uncanny valley effect
- production / Production
- Programming Collective Intelligence
- reference link / Machine learning
- prototype creation
- about / Prototype creation
- data preprocessing / Data preprocessing
- model, training / Model training, evaluation, and selection
- evaluation / Model training, evaluation, and selection
- selection / Model training, evaluation, and selection
- field, testing / Field testing
- python
- URL / Tools
- Python
- Python NLP libraries / Python NLP libraries
- Python packages
R
- radius-based neighbor learning / Choosing a good k
- random forest
- ensembling / Ensemble learning random forest
- training / Training the random forest
- accuracy, evaluation / Random forest accuracy evaluation
- recall / Precision, recall, and F1-score
- rectified linear unit (ReLU) / Rectifier-like activation functions
- regression model
- selecting, for issues / Choosing the regression model for your problem
- bias-variance trade-off / Bias-variance trade-off
- regression task / Understanding the regression task
- regularization layer
- about / Regularization layers
- dropout / Dropout
- batch normalization / Batch normalization
- regularization techniques
- about / Fixing linear regression problems with regularization
- linear regression, issues fixing / Fixing linear regression problems with regularization
- reinforcement learning / Reinforcement learning
- representation learning / Autoencoder neural networks
- residual / Fitting a regression line using the least squares method
- residual sum of squares (RSS) / Fitting a regression line using the least squares method
- ridge regression / Ridge regression and Tikhonov regularization
S
- scikit-learn
- URL / Tools
- Scikit-learn / Scikit-learn
- scikit-learn converter
- scipy
- URL / Tools
- sensitivity / Precision, recall, and F1-score
- server-side machine learning
- versus mobile machine learning / Mobile versus server-side ML
- ShuffleNet / ShuffleNet
- signal / Data preprocessing
- simple linear regression
- about / Introducing simple linear regression
- fitting, least squares method used / Fitting a regression line using the least squares method
- forecasting / Forecasting the future with simple linear regression
- softmax loss function
- vanishing gradient problem / Vanishing gradient problem
- biological analogies, viewing / Seeing biological analogies
- SqueezeNet / SqueezeNet
- Stochastic gradient descent (SGD) / Training the network
- supervised learning / Supervised learning
- supporting association measures / Supporting association measures
- Support vector machine (SVM) / General-purpose machine learning libraries
- Swift
- machine learning app, implementing / Implementing first machine learning app in Swift
- k-means clustering, implementing / Implementing k-means in Swift
- implementing, in Swift / Implementing k-means in Swift
- k-means clustering, implementing in Swift / Implementing k-means in Swift
- multiple linear regression, implementing / Implementing multiple linear regression in Swift
- neural layer, building / Building a neural layer in Swift
- layers, implementing / Implementing layers in Swift
T
- taxicab distance / Calculating the distance
- TensorFlow / TensorFlow
- textual corpuses / Textual corpuses
- threshold function / Step-like activation functions
- tikhonov regularization / Ridge regression and Tikhonov regularization
- time series classification / People motion recognition using inertial sensors
- tiny-dnn / tiny-dnn
- toolbox
- linear algebra / Machine learning toolbox
- probability theory / Machine learning toolbox
- data input-output / Machine learning toolbox
- data wrangling / Machine learning toolbox
- data analysis/statistics / Machine learning toolbox
- visualization / Machine learning toolbox
- symbolic computations / Machine learning toolbox
- interactive prototyping environment / Machine learning toolbox
- machine learning packages / Machine learning toolbox
- tools, machine learning app
- Torch / Torch
- Twenty Questions game
- Twitter text
- about / Twitter text
- reference link / Twitter text
U
- unit step function / Step-like activation functions
- universal approximation / Using neurons to build logical functions
- unsupervised learning / Unsupervised learning
- users emotions / Understanding users emotions
V
- vanishing gradient problem / Step-like activation functions
- Vector space properties / Vector space properties
- volume / Convolutional layer
W
- Wit.ai
- reference link / Machine learning and AI APIs
- within-cluster sum of squares (WCSS) / Implementing k-means in Swift
- Word2Vec
- about / Word2Vec, Twitter text
- in Gensim / Word2Vec in Gensim
- reference link / Word2Vec
- Word Association game / Word Association game
- word embedding / Where to go from here?
- word vector representations / Word vector representations
X
Y
- YCML / YCML