The First Sight
Let us travel back in time to my bachelor’s studies, when I started my journey in obtaining a bachelor’s degree with specialization in communications engineering at Al-Azhar University – Gaza. During the last two years at university, specifically in 2014, I had begun looking for a distinguished graduation project with a new concept and idea, and throughout the process of investigation and research, I came across new concepts in my minds such as classification, neural networks, and other fundamental principles of machine learning. Hence, the spark of desire to include machine learning and artificial intelligence concepts in my engineering and cognitive dictionaries arose.
A Winding Starting Road
I took more than one online course to gain information and establish the fundamental concepts of machine learning, including “An introductory course in machine learning (ML) – Learning from Data, Online Course (MOOC)” [1]. It appeared hefty and bold at first glance, owing to the abstract mathematical formulae within the course. Later I attended an applied course called “Machine Learning” in Coursera by Andrew Ng [2], which covers an overview of machine learning, data mining, and statistical pattern recognition. As a result, I found it more enjoyable to learn, because it was more practical than theoretical science (project-based learning). This provided me with a more comprehensive overview and served as a head start in machine learning programming journey.
Mastery and Applications of ML
The next part of my journey was the departure trip outside the walls of the homeland. After receiving a bachelor’s degree, I was awarded a scholarship to pursue a master’s degree at the University of Padua in Italy. During my master’s studies, there was a course called “Applied Machine Learning in Wireless Communications – Prof. Michele Rossi” [3]. Due to the wide range of applications such as signals, picture, and speech recognition, the course was adequate to begin the applied side of machine learning.
The next step was to broaden my applied and programming knowledge in the field of ML, which I accomplished by enrolling in a series of specialty courses called “Deep Learning Specialization” in Coursera by Andrew Ng” [4] (These courses were a complement to the aforementioned series by Andrew Ng).
The series was a set of five logical, applied, and informative courses as follows:
– Neural Networks and Deep Learning.
– Improving Deep Neural Networks: Hyperparameter Tuning, Regularization, and Optimization.
– Structuring Machine Learning Projects.
– Convolutional Neural Networks.
– Sequence Models.
As a result of completing the series, the student will be prepared to create and train neural network architectures, as well as engage in the development of leading-edge AI technology. I have applied the gained knowledge in my master’s degree thesis with the title “ECG Signal Acquisition Over Resource-Constrained IoT Channels with ML Algorithm”, which had the aim of power-reduction using machine learning algorithm and Autoencoder.
My Career as a Data Analyst using Machine Learning
I was oriented towards the job market after graduating and acquiring a master’s degree, and I have joined a local company in Italy, which focuses on artificial intelligence applications and addressing technological issues. The corporation was assigned the responsibility of developing a classifier to classify and identify fractures, cracks, water leaks, and defects within tunnels. This was accomplished by scanning and capturing photos to identify them afterwards using computer vision and machine learning algorithms.
After that, I have moved to IMDEA in order to apply machine learning algorithms for cellular systems, where I began working as a researcher. In IMDEA I had the opportunity to work on actual challenges, and implementation of reinforcement machine learning algorithms.
Rising to the Challenges
My passion and enthusiasm for obtaining a Ph.D. as a researcher qualified me for the Marie Curie Scholarship in Spain – UOC University and at the same time working within the WindMill project in the field of machine learning to develop Wireless communication networks.
Working in the WindMill project changed my perspective of the labor market. I’ve learned that training and building the model of ML is not always sufficient, but the essential thing is to develop and qualify the model of ML to be suitable for the manufacturing domain as an integrated product. The next step, which I’m currently in, is learning machine language as an integrated engineering environment for manufacturing through a series of 4 courses called “Machine Learning Engineering for Production (MLOps) Specialization” in Coursera by Andrew Ng” [5].
The series is a set of four effective and efficient engineering topics, with well-established tools and informative methodologies:
– Introduction to Machine Learning in Production.
– Machine Learning Data Lifecycle in Production.
– Machine Learning Modeling Pipelines in Production.
– Deploying Machine Learning Models in Production.
Tips Toward Joining the Machine Learning Bandwagon
It is worth mentioning that machine learning necessitates tenacity in studying ideas that you may be unfamiliar with, as well as large time investment in order to gain a thorough understanding of the underlying principles. Here are the top 8 Machine Learning tips for beginners:
Elaborate Mathematics and Numbers
You don’t have to be a statistician to process your data for machine learning. However, in order to grasp your results and where you are headed, you must first understand basic statistical principles, such as:
• Mean and distribution
• Statistical decision theory
• Regression
• Mean Square Error, Least Squares
Master the appropriate programming language
It is not difficult, as many imagine, to master a particular programming language, especially if it was popular, easy to learn, and commonly used for data analysis and machine learning such as Python or R.
Cover the Basics of Machine Learning
To begin, you must devote time and effort to understand the fundamental concepts of machine learning and data science, such as the algorithms and programs used in these areas.
Set Your Goals
Because the area of machine learning is always developing, in order to remain on track while learning you must define precise goals before plunging in. You can concentrate on certain applications or industries, as well as the challenges you wish to tackle with machine learning. In this manner, you may use your goals as a compass to guide you on your journey to mastery.
Conduct exploratory data analysis
Exploratory Data Analysis (EDA) is a method of analyzing data that employs visual tools. It is used to detect trends, patterns, the shape of data, feature correlations, and signals within the data that can be used to build predictive models, or to validate assumptions using statistical summaries and graphical representations.
Dive into Machine Learning models and techniques
Remember that machine learning includes supervised and unsupervised models and other classifications such as reinforcement learning. Solve problems under these different classifications to be aware of what each model does and where it should be used. Moreover, be familiar with some concepts such as autoencoders, clustering, feature separation techniques, labels, neural networks, and optimization.
Manage Big Data and explore Deep Learning Models
Access to enormous amounts of data for use in algorithms to provide valuable results must be handled carefully. It is necessary to understand how it is stored, accessed, and manipulated in order to create solutions and interconnect the data.
The work on the deep learning algorithm is done through connected neural network layers to convert the features of the data into results. Furthermore, by feeding deep learning with big data, you may achieve astonishing outcomes in management, innovation, sales, and productivity.
Do and Complete a Data Project
Now follow the previously mentioned tips and start your application project! Find machine learning projects for beginners such as face detection and voice recognition using pre-made online databases. Work smarter not harder, and get familiar with the popular machine learning libraries such as Keras, Matplotlib, NumPy, Pandas, TensorFlow, and so on.
Online Courses Referencing:
[1] Online Course, ” An introductory course in machine learning (ML) – Learning from Data, Online Course (MOOC)”.
https://home.work.caltech.edu/telecourse.html
[2] Andrew Ng, “Machine Learning”, Coursera.
https://www.shorturl.at/rsxPV
[3] Michele Rossi, “Applied Machine Learning in Wireless Communications”, University of Padova – Italy.
http://www.dei.unipd.it/~rossi/
[4] Andrew Ng, “Deep Learning Specialization “, Coursera. https://www.shorturl.at/nxAMR
[5] Andrew Ng, “Machine Learning Engineering for Production (MLOps) Specialization”, Coursera.
https://www.shorturl.at/abtvA
