ML — Machine Learning Intro
What Is It and How It Evolves Over Time? — AISeries — Episode #00
Hi, What is AI in the eyes of the public?
Maybe movies like Bicentennial Man (1999 — Based on a 1976 novel by Isaac Asimov), or Matrix (1999 — A computer hacker learns from mysterious rebels about the true nature of his reality and his role in the war against its controllers), or Ex Machina (2015— To erase the line between man and machine is to obscure the line between men and gods), or Blade Runner 2049 (2017 — an apocalyptic plunge of what’s left of society into chaos).
And the list goes on…but for now, what do the pioneers said?
Let’s hear what they said about Artificial Intelligence / Machine Learning 👇
I propose to consider the question: Can machines think? Allan Turing 1950
The branch of computer science concerned with making computer behave like human. John McCarty 1956
The science of making machine do things that would require intelligence if done by man. Marvin Minsky 1968
Machine Learning is a way of teaching programs that utilize data to generate algorithms rather than explicitly program an algorithm from the scratch (Oliver R. Simpson book).
In other words, Machine learning is the study of computer algorithms that improve automatically through experience and by the use of data. It is seen as a part of artificial intelligence (Wikepedia).
Machine learning is the study of computer algorithms that improve automatically through experience and by the use of data (Wikepedia).
In this episode of python awesomeness, we are going to make our first try in the field: ML.
Welcome!
But first, a Little History/Theories of AI, for contextualization, right?
AI incorporates
1950AI
1980ML
2010DL
Milestone — History of IA
1960 1969 — Golden Age — XOR Problem — M.Minsky
1970 1986 — Dark Age — MLP — Solution Non-Linear Separable Problem — D. Rumelhart & G.Hinton & R.Williams
1995SVM-Support-vector Machines —Sophisticated Solutions — V.Vapnik & C.Cortez
2006Deep Neural Network, DNN, is an artificial neural network (ANN) with multiple layers between the input and output layers.
Goal of ML
In machine learning, the goal of training is to create an accurate model that answers our questions correctly most of the time.
7 Steps for ML Solution
These principles apply:
1Gathering Data — The quality and quantity of data that you gather will directly determine how good your predictive model can be;
2Preparing That Data — Is the process of transforming raw data so that data scientists and analysts can run it through machine learning algorithms to uncover insights or make predictions;
3Choosing a Model — This is a process that can be applied both across different types of models (e.g. logistic regression, SVM, KNN, etc.);
4Training it — means learning (determining) good values for all the weights and the bias from labeled examples;
5Evaluating it— It helps to find the best model that represents our data and how well the chosen model will work in the future
6hyper-parameter training —Is a parameter whose value is used to control the learning process ;
7Prediction — PREDICT function that can be used to predict outcomes using the model
Source: The 7 Steps of Machine Learning by Yufeng G
Overall Procedure of Building a Model
Machine learning methods:
Supervised machine learning trains itself on a labeled data set. The goal of the program is to learn from the inputs to reproduce the output.
Unsupervised machine learning ingests unlabelled data — lots and lots of it — and uses algorithms to extract meaningful features needed to label, sort, and classify the data in real-time, without human intervention. The computer is left on its own to discover patterns inside the data.
Reinforcement machine learning is a behavioral machine learning model that is similar to supervised learning, but the algorithm isn’t trained using sample data. This model learns as it goes by using trial and error. A sequence of successful outcomes will be reinforced to develop the best recommendation or policy for a given problem (Ia. autonomous vehicle).
Reinforcement Learning vs. Machine Learning vs. Deep Learning
Reinforcement learning is a branch of machine learning.
Unlike unsupervised and supervised machine learning, reinforcement learning does not rely on a static dataset, but operates in a dynamic environment and learns from collected experiences.
Data points, or experiences, are collected during training through trial-and-error interactions between the environment and a software agent.
This aspect of reinforcement learning is important, because it alleviates the need for data collection, preprocessing, and labeling before training, otherwise necessary in supervised and unsupervised learning.
Practically, this means that, given the right incentive, a reinforcement learning model can start learning behavior on its own, without (human) supervision. (https://ww2.mathworks.cn/discovery/reinforcement-learning.html)
Applications of Machine Learning
Facial recognition/Image recognition;
Automatic Speech Recognition;
Financial Services;
Marketing and Sales;
Healthcare;
Spam filter: the algorithm learns a predictive model from data labelled as spam and “no spam” (ham). After training it can predict for new emails whether they are spam or not;
Character recognition;
Object recognition in images;
And more...What’s a Good Model?
Terminology of Machine Learning
(source: https://www.mygreatlearning.com/blog/what-is-machine-learning/)
Model: Also known as “hypothesis”, a machine learning model is the mathematical representation of a real-world process. A machine learning algorithm along with the training data builds a machine learning model.Feature: A feature is a measurable property or parameter of the data-set. Feature Vector: It is a set of multiple numeric features. We use it as an input to the machine learning model for training and prediction purposes.Training: An algorithm takes a set of data known as “training data” as input. The learning algorithm finds patterns in the input data and trains the model for expected results (target). The output of the training process is the machine learning model.Prediction: Once the machine learning model is ready, it can be fed with input data to provide a predicted output.Target (Label): The value that the machine learning model has to predict is called the target or label.Over fitting: When a massive amount of data trains a machine learning model, it tends to learn from the noise and inaccurate data entries. Here the model fails to characterise the data correctly.Under fitting: It is the scenario when the model fails to decipher the underlying trend in the input data. It destroys the accuracy of the machine learning model. In simple terms, the model or the algorithm does not fit the data well enough.Affine transformation: In Euclidean geometry, an affine transformation, or an affinity, is a geometric transformation that preserves lines and parallelism.Fold change: It is defined as the ratio between the two quantities; for quantities A and B, then the fold change of B with respect to A is B/A. In other words, a change from 30 to 60 is defined as a fold-change of 2Lingo: the language, word style, speech, jargon, or slang, used by a person or group
Related Posts
00#Episode — AISeries — ML — Machine Learning Intro — What Is It and How It Evolves Over Time? (this one:)
01#Episode — AISeries — Huawei ML FAQ — How do I get an HCIA certificate?
02#Episode — AISeries — Huawei ML FAQ Again — More annotation from Huawei Mock Exam
03#Episode — AISeries — AI In Graphics — Getting Intuition About Complex Math & More
04#Episode — AISeries — Huawei ML FAQ — Advanced — Even More annotation from Huawei Mock Exam
05#Episode — AISeries — SVM — Credit Card — Start to Finished — A Complete Colab Notebook Using the Default of Credit Card Clients Data Set from UCI
06#Episode — AISeries — SVM — Breast Cancer — Start to Finished — A Complete Colab Notebook Using the Default of Credit Card Clients Data Set from UCI
07#Episode — AISeries — SVM — Cupcakes or Muffins? — Start To Finished — Based on Alice Zhao post
More AI history:1950-Allan Turing: Can Machine Think?
1956-John McCarty: Can Machine Behave like Human?
1970-Failure of Projects: Machine Translation (AI Winter)
1997-Deep Blue: Defeated Chess World Champion Gary Gasparov
2011-IBM/Watson: Beats the brightest trivia minds at Jeopardy.
2016-Deep Mind/AlphaGo: Defeated Go World Champion Lee Sedol
2018-Deep Mind/AlphaGo Zero: Defeated StartCraft Game World Champion Grzegorz -MaNa- Komincz
Credits & References
INTRODUÇÃO A MACHINE LEARNING PARA CERTIFICAÇÃO HCIA-AI by crateus.ufc.br
The 7 steps of machine learning by Google Cloud Tech
A guide to convolution arithmetic for deep learning by Vincent Dumoulin F and Francesco Visin 2 F † FMILA — Université de Montréal †AIRLab, Politecnico di Milano
The Difference between Data Engineer, Data Scientists & Data Analysts:
Data Scientists and Data Analysts need to have TRUSTED, TIMELY & EFFICIENT ACCESS to the Data to do their best work.Data Enginner's role is the most distinct amongst the three. Data Enginner CREATE & MANTAIN THE INFRASTRUCTURE & DATA PIPELINE that takes terabytes of raw data coming from differente sources into one CENTRALIZED LOCATION with clean, relevant data for the organization.Data Analyst do what's described in their title: ANALYSE DATA & REPORT INSIGHTS from their analysis, and as AI field, for DEVELOPING A PREDICTIVE MODEL for BETTER ALLOCATION OF RESOURCES & FOR INCENTIVE OF PLANNING.Data Scientists INVESTIGATE, EXTRACT & REPORT MEANINGFUL INSIGHTS into the organization's data.In summary, Engineer create data, Scientist investigate data, and Analyst report insights :)
Stay away from negative people. They have a problem for every solution.
Albert Einstein
