Google Cloud for Machine Learning 2020 Master Course

Learn how network communication works, explore http, build a small api on Google Cloud, and study python basics, json data storage formats, and the flask framework through browser simulations.

Explore basic network communications, how data is transferred between machines, and how a protocol uses start-of-message markers (four zeros) and IP addresses to ensure sender identity, data integrity, and sequencing.

Explore the hypertext transfer protocol, its request–response model, and common methods (get, post, delete) along with status codes (200, 404, 4xx) to design and test web APIs from a browser.

Learn to build an API on App Engine in Google Cloud, as a middleman between users and data, plus hello world in Python with Cloud Console and Cloud Shell.

Explore Python basics, including declaring variables, printing values, defining and calling functions, and looping through lists with for loops.

Learn to build a basic web API on Google Cloud by using Flask in Python, including creating routes, handling JSON responses, and running a simple get/post endpoint.

Build a Google Cloud rest api that accepts a number via the url path /calculate/{num} and returns a json with the number, its square, cube, square root, and cube root.

Explore building a simple API with arrays by creating a fruits list, defining GET and POST methods, and exposing the /fruits endpoint for retrieval and updates.

Create a python api route '/add fruit' that appends a fruit name to the fruit list. Demonstrate with apple, mango, banana, grapes, and guava.

Develop a simple API on App Engine with Flask to handle get requests and return calculated values, review JSON objects, and preview Google Cloud compute engine tasks.

Explore Google Cloud Compute Engine and Cloud Functions, deploy programs on remote compute instances, understand remote computing workflow, and integrate Cloud Functions into a project.

Contrast local computer workflows with remote compute via an API to cloud-powered processors. Identify benefits and drawbacks like scalability, redundancy, and network latency.

Create a new Compute Engine virtual machine by selecting region, machine type, disk, and Ubuntu image, then test its compute capability with a factorization problem on 60.

Launch a compute engine vm in a chosen zone, connect via the browser, and develop a Python program for factoring a number using nano and command-line tools.

Explore resource quantisation in Google Cloud for machine learning by building a compute engine that computes prime factors with Python on a virtual machine.

Explore cloud functions as standalone tasks, define and call Python functions with inputs and outputs, and apply functional abstraction in Google Cloud.

Create and deploy a hello world cloud function in Google Cloud Functions, then extend it to greet users by name via an http trigger with simple input handling.

Create a cloud function that validates an email by checking for an at symbol and a dot, then deploy and test it on cloud functions using Python.

Compare local and remote compute and how APIs enable remote workflows. Examine virtual machines and four virtualization types—application, desktop, network (VPN), and storage—and their impact on scalability and cost.

Develop and deploy a usable cloud function, learn JavaScript basics, and build a simple full-stack demo site with client-side requests and external data sources on Google Cloud.

Learn how cross-origin requests are controlled by the browser using pre-flight checks, with the frontend asking the backend for a resource only after the server grants authorization.

Build a simple HTML boilerplate email checker app on Google Cloud, using Cloud Shell and Cloud Functions, with npm deployment and a basic web page hosted publicly.

Create an html page that uses a cloud function backend to validate email in a mail checker app, with input and textarea fields and a submit button for Google Cloud.

Explore JavaScript basics with a hello world example and core building blocks like variables, functions, conditionals, and loops, demonstrated through browser console testing.

Integrate a cloud function that validates emails with your website, and learn how a cross-origin request is blocked by CORS due to a missing access-control-allow-origin header.

Add CORS to cloud functions by building a Flask-based backend, updating requirements, and returning a structured response with access-control-allow-origin and access-control-allow-methods headers for get and post requests.

Deploy a cloud function to validate emails with access control and origin checks, then trigger it from a button, log outcomes, and view responses like email is invalid.

This lecture shows how to extract user input from a text area with document.getElementById and .value, build a query, and call a cloud function from the frontend.

Explore the need for data storage and basic database architectures, then build a Python CLI e-commerce app using Firebase and cloud storage to perform CRUD operations.

Compare storing data in simple sentences versus tabular databases and object storage, and see how SQL and JSON objects organize product data for efficient retrieval.

Initialize the firebase admin app with service account credentials, create a cloud database, add products such as keyboard and mouse, and safeguard the key file while accessing the e-commerce store.

Learn to perform a read operation on Firestore with Python by listing all products, retrieving a document by user-specified name, converting it to a dictionary, and printing.

Learn to perform update operations on Firestore with Python by building a menu-driven app that searches, updates, and deletes product documents based on user input.

Create a new product document in Firestore using Python by collecting the name, price, and delivery from the user and saving them under a document named after the product.

learn how to delete a Firestore document using Python, guided by a user-driven flow that selects a product by name and removes its document from the database.

Learn to use google cloud to work with big data, compare firebase and equity, and run simple queries on csv datasets like covid-19 and population data.

Compare cloud storage, memory, and non-relational stores like cloud data store and Bigtable with relational tools and warehousing, then cover acquiring data—from self-generated to externally sourced datasets—licenses and csv formats.

Generate a csv dataset with python on Google Cloud, computing number, square, cube, square root, and cube root for 100 to 3000 rows, and write to csv for BigQuery.

Explore how to work with Google Cloud BigQuery: create datasets and tables, upload files, auto-detect schema, and run SQL queries using select from and limit to analyze data.

Explore compound SQL queries and advanced data selection using where clauses, conditions, and limits, with outputs in JSON format and practical examples from audit tables and public data sets.

Analyze public datasets in Google Cloud BigQuery, focusing on a world population grid that tracks country, population, and geographic coordinates across one-kilometer squares while highlighting size and cost considerations.

Learn to query data with Python using Google Cloud BigQuery, create a client, build a select query with where and limit, run it, and print the results.

Contrast BigQuery and Firestore for real-time vs stored data, cover csv basics, and show how to run select queries from Python using Google Cloud BigQuery.

Explore the fundamentals of machine learning, artificial intelligence, and deep learning, and see how neural networks, activation functions, and cloud-based technologies power machine learning development.

Explore the neuron, the basic building block of neural networks, a small input gateway that processes multiple inputs into outputs, inspired by nerve cells and connecting to other neurons.

Learn how neuron inputs, with assigned weights, produce an output: if the sum of weighted inputs is greater than 1, the neuron fires (on), otherwise it stays off, with examples.

Learn how biases alter neuron outputs in neural networks by adding to weighted inputs, shifting activations from 0/1 to a continuous range, across many neurons and connections.

Explore activation functions and forward propagation in neural networks, comparing linear versus nonlinear options like sigmoid and relu, including leaky relu and bias considerations.

Application specific hardware, or ASICs, are built for a single task and deliver exceptional efficiency, with CPUs for sequential processing, GPUs for rendering, and TPUs for cloud machine learning.

Learn to use Google Cloud to write and train machine learning programs, build a neural network to classify handwritten digits, and monitor performance with Google Cloud monitors.

Explore what machine learning is, contrasting classical programming with learning from input-output examples. See how MNIST digit classification trains models through iterative loss optimization to recognize handwritten digits.

Shows how inputs and outputs train a machine learning model to generate rules via weights and biases, then test these rules in a classical programming model.

Set up a cloud linux environment, confirm python, and build a Keras sequential model to classify amnesty images using dense layers, dropout, and softmax, then train and evaluate.

Explore how machine learning teaches computers to perform tasks by learning from data, covering neural networks, loss and reward, and problem types like classification, regression, clustering, and anomaly detection.