Machine Learning — Post By Anjuum Khanna

So Anjuum Khanna is back with a new insight on machine learning. As we all know that machine learning is a part of artificial intelligence. So in Anjuum Khanna’s words, machine learning is a method to make computer machine learn with data to perform a specific task in a progressively improving pattern.

In other words, we can also say that “Machine Learning is an idea to learn from examples and experience, without being explicitly programmed. Instead of writing code, you feed data to the generic algorithm, and it builds logic based on the data given”. So this is all about the simple algorithm which is smart enough to perform in different scenarios.

For more clarification let’s take help of an example, one kind of algorithm is a classification algorithm. It can put data into different groups. The classification algorithm used to detect handwritten alphabets could also be used to classify emails into spam and not-spam. So in Anjuum Khanna’s words, these algorithms learn by their own experience and can be used for different outputs.

This name “Machine Learning” was invented in 1959 by Arthur Samuel. This was evolved from the study of pattern recognition and computational learning theory in artificial intelligence, machine learning explores the study and construction of algorithms that can learn from and make predictions on data — such algorithms overcome following strictly static program instructions by making data-driven predictions or decisions, through building a model from sample inputs.

The requirement of Machine Learning:-

Machine Learning is a part of Artificial Intelligence (AI). Applying AI, we wanted to build better and intelligent machines. But except for some small achievements such as finding the shortest path between point two points, we were unable to program more complex and constantly evolving challenges. So as we have learned that if there is a will there is a way, so we found out that the only way to be able to achieve this task was to let machine learn from itself and then should give output as required. This sounds similar to a child learning from its self. So machine learning was developed as a new capability for computers.

All the results based on data were only possible when the human brain creates a pattern in that data. The data is very massive, the time taken to compute is increased, and this is where Machine Learning comes into action, to help people with large data in minimum time.

So machine learning as a technology helps analyze big chunks of data, easing the task of data scientists in an automated process and gaining equal importance and recognition.

The techniques we use for data mining have been around for many years, but we were processing that data with human efforts and that was quite a time taking procedure. If you run deep learning with access to better data, the output we get will lead to dramatic breakthroughs which are machine learning.

Types of Machine Learning:-

Supervised Learning Major part of machine learning is supervised learning. So in supervised learning machine learns from previous examples. So as per Anjuum Khanna’s simple definition in this learning, we provide both input and output data. This data use an algorithm to derive the mapping function from the input to the output.

In supervised learning we can classify problems into two parts, one is the classification, and the other is a regression.

Classification: A classification problem is when the output variable is a category or a group, such as “black” or “white” or “spam” and “no spam”.

Regression: A regression problem is when the output variable is a real value, such as “Rupees” or “height.”

Unsupervised Learning In this type of learning usually there is only input data. There is no output data. So I, Anjum Khanna defines it in a very interesting line that in unsupervised learning algorithms are left to themselves to discover interesting structures in the data.

As there is no output data so the machine doesn’t have any correct answers, the machine has to find itself all correct answers based upon an algorithm.

Unsupervised learning problems can be further divided into association and clustering problems.

Association: An association rule learning problem is where you want to discover rules that describe large portions of your data, such as “people that buy X also tend to buy Y”.

Clustering: A clustering problem is where you want to discover the inherent groupings in the data, such as grouping customers by purchasing behaviour.

So this is a very interesting and innovative type of machine learning and this is getting evolved day by day. This type of learning is getting used by many big companies to predict consumer behavior etc.

Reinforcement Learning In this type of learning there used to be a computer program which interacts with a dynamic environment where it must perform a particular goal (such as playing a game with an opponent or driving a car). The program is provided feedback in terms of rewards and punishments as it navigates its problem space.

Using this algorithm, the machine is trained to make specific decisions. It works this way: the machine is exposed to an environment where it continuously trains itself using feedback from the last tasks.

What Machine Learning can do:-

I am amazed by seeing a change which is coming with artificial intelligence especially with machine learning. There are few areas where we use machine learning very often without even knowing that this is machine learning. Some of my favorite fields of machine learning are:-

Face detection: Machine learning can be used to detect face or match face in a no of photographs.

Email filtering: With the help of machine learning we can classify spam or no spam emails.

Medical diagnosis: This is another important area where machine learning is quite important. We can diagnose whether patient has the specific disease or not.

Weather prediction: With the help of older data we can forecast about the weather. As there will be rain or not.

Prerequisites to learn Machine Learning:-

Machine Learning requires a lot of logic and calculation hence mathematics is an important part of it. So it meets statistical, probabilistic, computer science and algorithmic aspects arising from learning iteratively from data which can be used to build intelligent applications.

There are many reasons why the mathematics of Machine Learning is necessary, and I am highlighting those points in below points:-

• A lot of mathematical knowledge required to select the appropriate algorithm for problem includes considerations of accuracy, training time, and model complexity, the number of parameters and number of characteristics.
• Again for identification of underfitting and overfitting by following the Bias-Variance trade-off is required.
• Estimating the right determination period and uncertainty.

So there are few basic topics on which an aspirant need to have a good grip. Those topics are:-

1. Linear Algebra
2. Probability Theory and Statistics
3. Calculus
4. Algorithms and Complex Optimisations

So finally I, Anjuum Khanna wants to summarise that machine learning is a part of AI where we set up an algorithm to reduce data input. This algorithm learns from examples and performs in a perfect manner. Forgetting the good hold of machine learning developer should have good knowledge of maths and should learn from past experience to evolve faster.

Originally published at anjuumkhanna.in on September 10, 2018.