Foundations — Samur.AI
What’s up Samurais? Ready to upgrade your mind? Let’s go then!
I have noticed that there exists some confusion as to what are the prerequisites for starting out with machine learning. In this really short article, you will gain knowledge on the basic prerequisites for being a master in machine learning and some basic terminology to help you feel comfortable with the subject.
Machine Learning — A very common term nowadays!
The term machine learning gets thrown around a lot these days. In fact, you might have observed that everybody claims that they have an excellent grasp of the subject, but few people actually understand the mechanics behind all the magic! Thus, whenever you try to study any topic, make sure that you understand the mathematics behind the subject matter. This way, you will be able to expand your knowledge base and be proficient in the domain of artificial intelligence.
To start off, make sure you have some introductory knowledge in the following domains:
- Linear Algebra
- Probability
- Differential Calculus
- Statistics
Don’t be afraid, it is not that difficult.
For fun and animated understanding of the subjects, visit the channel 3Blue1Brown on YouTube. In addition, check out this GitHub repo by Siraj Raval which lays out the topics in a comprehensive manner with numerous resources to augment your mind.
In addition to the above topics, experience with a programming language will be a bonus. If possible, learn python. It has awesome frameworks with a variety of packages for data preprocessing and applied machine learning. It also supports deep learning extensively.
The following section will define some key terminology which is prevalent in the domain of machine and deep learning.
- Training data — It refers to the dataset that is used for training the machine learning model(algorithm). This model is later fed with new examples on which the model predicts the output. More correct predictions result in a good model with optimal performance.
- Testing Data — This data refers to the new set of examples that is fed to the trained model so as to measure the performance of the model and improve the model if required. The performance of the model is improved with the help of hyperparameter tuning. More on that later.
- Model — ‘Model’ is a synonym for a machine learning algorithm. It is the core of any product incorporated with any machine learning concept. The model is trained over time with the help of the dataset. One must always remember, more the data, better the model.
- Preprocessing — Data preprocessing or simply preprocessing is that stage of any machine learning problem that focuses on the cleaning of the data and making the data suitable as per the model architecture.
- Prediction — After the model is trained on the training data, a new set of examples are fed to the model to check its performance. If the model satisfies the requirements, then it may be accepted, otherwise, it may be improved using hyperparameter tuning.
- Visualization — Visualization is that part of the pipeline which focuses on generating a visual representation of the relationship between different variables in the dataset. Some examples are charts, histograms, plots etc. One must always make sure to visualize the data before proceeding with the data preprocessing stage as it will help the team to select the variables having the most impact on the problem at hand. Though many people underestimate the power of data visualization, it is a part of the pipeline with high significance.
So those were the primary terms required to start your machine learning journey. There are numerous other topics that are required to be comfortable with the subject. They will be covered in later blog posts. Subsequently, we will touch on topics like types of machine learning problems(supervised, unsupervised and reinforcement learning) and revolutionary deep learning. Later articles will also include pictorial representations for more immersion.
Feel free to ask doubts in the comments section and mention what else would you like to read about.
Join the AI movement. Upgrade your mind, upgrade your life!
Check out the link below for an introduction to linear regression.
If you like this article, consider buying me a coffee :)
