Question

AI, Robotics, Machine Learning, etc.

1. How do we distinguish between robotics and machine learning?

2. What are some ways in which Machine Learning (ML) is being used in business?

3. What is the need for ‘training data?’ Exactly who or what is being trained?

4. What are the two kinds of ‘learning’ that we talk about in ML?

a. What is the difference between the two?

b. Why is it significant?

Answer 1

1. Robotics is something which deals with the technology of robots entirely. A robot is a programmable machine that carries out a set of tasks autonomously in some way. Some robots are manual and some are programmable. Robots which are made to do a particular job repetitively with so many constraints need to be explicitly programmed. Since,they are used for only automation without much thinking and learning. If the robot needs to learn daily and adopt to new situations there is a need of machine learning logic inside of them. For Example, A line-follower robot is one which needs to be programmed well to pass all the constraints,it simply checks and goes in the way it doesn't learn,adopt and move in the path where as, a maze solver is one through which we can use ML algorithms to learn and go in the path easily and fastly. This is a simple example to distinguish robotics and machine learning.

2.There is a great demand for ML in business sector, Almost all the top companies use ML algorithms to yield better income.Some of the finest examples are Image Recognition which is used by Amazon to scan products and it automatically identifies and shows us the results.To have better reviews management system ML is vital. Working with huge data of user review and sorting them is important to bring better products into market. Chatbots for user satisfication using ML.

3. Training data is important. We need to choose a data set for our required problem and fetch our dataset to produce ML model. Our Model is being trained with the dataset.

4. The two kinds of learning in ML are Supervised and Unsupervised Machine Learning.

(a) Supervised Learning :- It is a type of learning model where the model is initially fetched with dataset. Based on predications it learns and when the required level of accuracy is attained it is deployed. It is given a dataset,it learns,later it identifies. For example, Many pictures of apples are given as dataset it keeps on recognising and learning from it later it is used for identifying. It is like a teacher teaching something and later the student solves. Popular one is Linear Regression model,Voice assistant is a good example of supervised learning. Supervised Learning algorithms have expected output.

Unsupervised Learning :- The data instances of a training dataset do not have an expected outcome as in supervised algo. It detects pattern based on init characteristics of the input data. In this type of learning an unlabeled data set is used to train based on the characteristics it groups the data. It can distinguish different data but couldn't identify the meaning as in supervised learning. They are left on there own to learn. There is nothing to teach and train them it should figure out on its own. For example, Recommendations in online shopping is based on this type of learning where it learns from the customer previous orders.

(b) Both are significant learning methods and are used extensively in different sectors based on usage. Both are equally important and cannot be compared on the scale of which is better than the other.