Advanced Machine Learning & Deep Learning Masterclass

Explore python basics by learning variables, assignments, and simple operations in IPython notebook, including x and y examples and using underscore to reference previous outputs.

Explore strings in Python for absolute beginners by working with variables, concatenation, and indexing, including zero-based and negative indexing, plus handling index errors.

Explore python basics for absolute beginners: declare variables, manipulate strings, practice indexing and slicing with exclusive upper bounds, and measure string length.

Explore Python lists in an absolute beginners session, demonstrating saving multiple values in IPython notebook, showcasing lists with mixed data types including strings and integers, and contrasting with Java.

Explore Python lists by appending values, inserting at specific indices, and removing items, with examples of numbers, strings, and mixed lists, and zero-based indexing.

Learn Python list operations, including popping the first element by index and removing it from the list. Compute min, max, and sum, and sort lists to practice data manipulation.

Master python conditional statements and boolean expressions, exploring the boolean type and true/false states, the if-else-elif structure, and indentation with flowcharts and pseudocode.

Learn boolean expressions and conditional expressions in Python, including equality, inequality, and comparison operators, indentation, and how flowcharts guide if-else logic for beginners.

Explore boolean operations (and, or, not) with Python examples, truth tables, and input validation, including boundary testing and practical problem solving concepts.

Explore Python iteration with while and for loops, learn flowchart planning, and master indentation rules to prevent infinite loops and write clear, well-structured code.

Explore Python basics by building a guess the number game with while loops and if conditions, handling user input, integer conversion, and simple decision making in PyCharm.

Explore simple Python demos for absolute beginners, import random, generate a random integer from 1 to 20, and print the result across multiple runs.

Learn how flowcharts visually map a process using standard symbols like terminal, input/output, and process to depict sequence and decision making.

Explore repetition in program design. Understand how constructs control execution: while evaluates before statements, repeat until runs at least once, and for loops use a fixed count.

Master the computational approach to problem solving by clarifying assumptions, discarding irrelevant information, and formalizing solutions as algorithms using flowcharts, pseudocode, and structure charts.

Explore flowcharts as a problem-solving tool, using start, input, processing, output, and stop symbols to visualize algorithms with examples like sum of two numbers and temperature conversion.

Explore how to read and inspect raw data with statistics in python, using pandas to load the pima Indians diabetes dataset, verify headers, and handle file path issues.

learn to check data dimensions and shape, data types, and describe statistics with pandas read_csv to assess class distribution and detect imbalanced data for classification.

Explore the correlation between attributes using Pearson's correlation coefficient and build a correlation matrix in Python with pandas to assess pairwise relations and their impact on linear and logistic regression.

Explore statistics by examining correlation between dataset attributes with Pearson's correlation coefficient and pairwise correlations, correlation matrix, and learn how high correlations impact linear and logistic regression in machine learning.

Explore data pre-processing in machine learning by scaling data to a zero-to-one range using MinMaxScaler in scikit-learn, demonstrated on a diabetes dataset with Python, pandas, and NumPy.

Explore data pre-processing techniques including normalization, binarization with a threshold, and standardization using scikit-learn, pandas, and the diabetes dataset to prepare features for machine learning.

Explore univariate feature selection to identify the best features with strongest relation to the output using statistical testing, chi2, and select k best in scikit-learn.

Explore data visualization for machine learning by loading the Kaggle food prediction dataset, merging tables, and plotting a bar chart of the most popular foods by quantity sold.

Explore creating bar charts, pie charts, and histograms in Python to visualize meals per food item, cuisine orders, and price distributions using Matplotlib.

Explore how artificial neural networks mimic the brain to solve problems, covering neuron models, inputs, weights, biases, activation functions, and applications in medical, banking, and defense, plus transfer function examples.

Build a first single-layer artificial neural network from scratch in Python, summing weighted inputs plus bias to produce the output, with no activation function in this step.

Explore multiple input neurons with weights and biases, and apply transfer functions to compute neuron outputs. Compare single versus multiple input neurons, using matrix notation to model layered networks.

Build a simple neural network from scratch in python, including weights and biases across input, hidden, and a three-neuron output layer.

Explore an illustrative neural network example of a fruit sorter using shape, texture, and weight inputs to classify apples versus oranges with a perceptron.

Create a step-by-step Keras neural network in Python using a diabetes dataset, with eight input features, two hidden layers, a sigmoid output, and training with binary crossentropy and Adam.

Train a handwritten digits recognizer in step-by-step deep learning project using Python and Keras, building a 28x28 image classifier with a sequential model, normalization, dense layers, ReLU, softmax, and Adam.

Explore Naive Bayes classifier, based on Bayes theorem, and implement Bernoulli, Multinomial, and Gaussian Naive Bayes in Python with scikit-learn on a spam dataset.

Set up your machine learning workspace by installing anaconda on macos, launching jupyter notebooks, and creating a ready python environment to run machine learning examples.

Explore tokenization in natural language processing by defining tokens and types, comparing word and sentence tokenization, and addressing corpora challenges with the NLTK toolkit.

Learn to download and set up NLTK for natural language processing using Python, install via pip, configure PyCharm, and download NLTK modules for future tokenization lessons.

Learn to tokenize text with nltk in python, splitting sample text into sentences and words for natural language processing.

Learn how normalization, a pre-processing step in natural language processing, uses tokenization, stemming, and lemmatization to produce a uniform word form, enhancing information retrieval.

Explore stemming and lemmatization with NLTK in Python, using the Porter stemmer and WordNetLemmatizer to normalize word variants.

Explore part of speech tagging in natural language processing, label each word with its grammatical role using NLTK, and see how this preprocessing reduces ambiguity in text.

Part of speech tagging serves as a pre-processing step in natural language processing, and the lecture demonstrates tagging with NLTK, tokenization, and practical examples while highlighting ambiguity and uses.

Explore stopwords in natural language processing and learn how to filter noise from text by removing English stopwords with NLTK, and apply this to single sentences and corpora.

Discover named entity recognition as the task of finding and classifying names in text, such as people, organizations, GPE, dates, times, and money, to enable indexing and question-answering systems.

Explore named entity recognition with NLTK, using tokenization and part-of-speech tagging to chunk entities and observe misidentifications, including GPE and person labels, and the binary output option.

Explore text classification with Naive Bayes, including spam filtering, sentiment analysis, and topic, language, and authorship detection. Learn to extract features and build classifiers for movie reviews using Naive Bayes.

Learn to classify text and analyze sentiment in movie reviews using NLTK and Naive Bayes, then preprocess by organizing words and creating labeled documents.

Create a text classifier with the Natural Language Toolkit by selecting the top 3000 words from movie reviews, building feature dictionaries, and forming labeled positive and negative feature sets.

Split the data into a 75/25 training and test set, train a Naive Bayes classifier with NLTK on 1500 training features and 500 test features, and achieve about 82% accuracy.

Experiment with classification in natural language processing by tuning training/test splits and word features, applying normalization and stopword removal to surpass the prior 82% Naive Bayes accuracy with NLTK.

Explore natural language processing with a Twitter sentiment analysis project that gauges New Zealand and coronavirus sentiments, using a Python workflow and a CSV corpus.

Create a Twitter application to obtain keys and tokens, submit for approval, and authenticate with the Python Twitter API to access tweets and verify credentials.

Fetch 100 tweets for a keyword via the Twitter API to build the test set for the sentiment analyzer, handling exceptions and returning tweet texts with placeholder labels.

Prepare a training set by retrieving tweet text from ids, pairing with sentiment labels, and writing a tweet data file with id, text, label, and topic.

Preprocess tweets with a Python class using NLTK to lowercase, replace URLs and mentions, remove punctuation and hashtags, tokenize, and filter stopwords, then build labeled training and test CSV sets.

Create a vocabulary from preprocessed training data and derive a feature set of all words. Apply features using NLTK to train a Naive Bayes classifier for tweet sentiment.

Run a Twitter sentiment analyzer to classify tweets as positive or negative, preprocess tweets, tokenize, and build training and test sets for a classifier.

Compare linear regression and logistic regression, highlighting why linear regression cannot model a binary insurance outcome. Learn how the logistic model handles nonlinear separability and contrasts the linear probability model.

Explore univariate linear regression through a hands-on demo that predicts happiness scores from GDP per capita, using a scratch implementation, data loading, and train-test split.

Explore a univariate linear regression demo that trains a model to predict happiness score from GDP per capita, using gradient descent, learning rate, and iterations, with training and testing analysis.

explore how clustering groups data into high intra-cluster similarity and low inter-cluster similarity. see how unsupervised learning uses k-means, centroids, euclidean distance, and iris data.

Explore deep learning and artificial neural networks, from single-layer perceptrons to multi-layer nets and CNNs, with activation functions like ReLU, backpropagation, and transformer models.

Explore end-to-end learning, multitask learning with task descriptors, transfer learning with pre-trained models, and meta learning techniques to optimize rapid adaptation across tasks.

Explore the design of convolutional neural networks, from three-channel image processing and three-dimensional filters to architectures like LeNet, AlexNet, ResNet, and Inception, with pooling, padding, and hyperparameter effects.

Explore how large language models evolve from deep learning, pre-train on vast data, enable transfer learning, and support translation, summarization, and question answering.

Explore transformers, self-attention and multi-head attention, encoder-decoder architectures, and how they overcome rnn and lstm limits for sequence modeling.

Explore deep generative models that learn probability distributions to generate data samples. Learn about autoencoders, variational autoencoders, generative adversarial networks, and diffusion models, and their latent variables.

Explore deep neural networks: from biology-inspired neurons and layer architectures to forward and back propagation, vectorization, activation functions, and regularization techniques.

Explore deep sequence models and their applications, from RNNs, GRUs, and LSTMs to bidirectional architectures, attention mechanisms, and the rise of transformers for language modeling.

Explore reinforcement learning fundamentals, including agents, states, actions, and rewards, and balance exploration and exploitation using value-based and policy-based methods with discounting.

Assess the current status of deep learning, from neural networks to transformers, and discuss training challenges, architectures, and transfer learning. Examine potential societal, economic, and ethical implications.

Explore deep neural networks, from neuron models to multi-layer architectures, and master forward and backpropagation with vectorized computations, activation, and regularization such as L2 and dropout.