TERMINOLOGIES TO KNOW AS A MACHINE LEARNING ENGINEER — PART 1
Let me stay away from “cybersecurity” for a while and decided to write something about Machine learning and data science for the community. So, I would like to start with some basic terminologies in ML as this is my first blog on Machine learning.
Artificial Intelligence (AI) which is a popular term nowadays, which comprises the Machine learning (ML) and Deep learning (DL). To be more clear, Deep learning (DL) is a subset of Machine learning (ML). And Machine learning (ML) is a subset of Artificial Intelligence (AI).
What is Machine learning?
First, let me start with the definition of ML. ML is a branch of Artificial Intelligence (AI) which is used to predict the outcome with more accuracy. Machine learning is how a computer system develops its intelligence. One of the ways to train this model is by deploying neural networks, which is a series of algorithms.
Types of Machine learning
The model is trained with a inputs and desired output. This type of learning is done by feeding the dataset with labelled data. Classification, regression, decision tree fall under this category. Supervised learning would be more accurate, simple and has known number of classes.
In this type of learning, the dataset contains unlabeled data. The task of unsupervised learning is to group unsorted data according to similarities, patterns and differences without any prior training data. Clustering and association comes under unsupervised learning.
This is similar to supervised learning but the dataset contains both labelled and unlabeled data. The labelled data contains necessary tags to represent the data, whereas, the unlabeled data lack information. By using this combination, the model tries to label the unlabeled data.
Reinforcement learning is about making decisions sequentially. In simple words, the output depends on the current input and the next input depends on the previous output. This learning decision is dependent, hence the labels are given to sequence of dependent decisions.
Reinforcement learning has an agent, an environment. The agent produces an action and as a result the environment replies with a reaction. Then the model takes that feedback and learns.
Machine learning Vs Deep learning
- Deep learning uses neurons to compute the output. While ML doesn’t uses any neurons instead they use models which are trained to produce the desired output.
- Deep learning is used for complex data such as images, audio, video files etc. ML is used for numerical data or data points.
- Deep learning is used for large number of data. ML is used for minimum data.
In a case, where we have less number of images in a dataset and want to process it, then we can go with Machine learning algorithms instead of deep learning algorithms. For minimum data we can use ML because deep learning algorithms are more complex and consumes more memory. So, if we the data is minimum we use ML to efficiently compute the data even for images, audio, videos too.
Let me stop here with these basics and hope you now have a better understanding about what is machine learning and the difference between ML and Deep learning. We’ll get into some more important terms in machine learning in the next part.