Machine Learning (ML) is a buzz word in today’s era with its application lying in day-to-day life. We make use of ML in our daily routine without being aware about it. So, let’s us know how and where we use ML and how dependent are we on these technology.
Let us discuss each application with example in detail:
1. Image Recognition
Source: https://azati.ai/
It is one of the most common applications of machine learning. It deals with identification of any digital image such as person, place, objects etc based on the pixels.
The most common example is automatic friend tagging suggestion on facebook. When you upload an image on facebook, it automatically recognizes your friend and then suggestion name for tagging.
Another example is of biometric attendance marking system in offices. Your attendance is marked based on your face.
Google Lens is another example of image recognition. If you will put google lens in front of an object say, a plant, it will show you all details about that particular plant.
Telling the mood of a person by his/her face is another example of image recognition. To this, anything related to scanning of image comes under image recognition.
2. Speech/ Voice Recognition
Speech and voice recognition are often considered as same. However, there exists a minute difference between both. Speech recognition particularly deals with converting verbal format to text such as translation of voice into text in a video. While voice recognition deals with recognizing the voice of a person and doing desired job such as setting an alarm.
Other examples of speech/ voice recognition is virtual personal assistants such as Google Assistant, Alexa, Siri recognizes voice and do the work such as search, open emails, setting alarm, voice dialing, etc as directed. Smart device such as speakers, google home also make use of speech recognition. Converting voice to text in a video in various languages is another example.
3. Stock Market Trading
People dealing with stocks are always in doubt while buying or selling of stocks due to high risk associated with it. They are in need of advice regarding buying/ selling of stock. Machine learning provides this guide by predicting stock market trends and helps them reach to a decision.
Now-a-days, in addition to machine learning, deep learning models such as LSTM, RNN etc. are also used to predict stocks.
4. Self Driving Cars
Source: https://www.vox.com/
Self-driving cars is another most common application of machine learning. In this, machine learning trains the car for object detection, speed limit, directions, road map, and everything else that is required for driving to avoid any accidental incidents. Tesla is an example of company that manufactures self-driving cars using ML.
5. Commuting Predictions
This banner can be divided into two parts: Traffic Predictions and Online Transportation Networks
Traffic Predictions
Whenever we travel, we make use of maps to check traffic conditions and shortest path to reach the destination in minimum travel time. This is done by saving real-time data of commuters like current locations and velocity of vehicles at central server for managing traffic. Then, this data is used to show shortest path, traffic congestion etc to other commuters.
Another way of providing congestion details is by utilizing already saved data based on daily experience using machine learning. For instance, suppose at 6pm, on usual basis there is congestion in gurgaon, so ML technique will predict congestion at that time in case real-time data is not gathered.
For checking it, we need to use GPS navigation services.
Online Transportation Network
While booking a cab, you must have seen price fluctuation with time on same route. For example, if it takes Rs.400 to travel from Gurgaon to Rohini at 12pm, it might show the travel price of Rs. 800 when travelling at 8pm. The price surge based on riders and other factors are predicted and provided based on ML models.
6. Language Translation
Under this banner, the text can be convert to any language you desired. This is useful in case we need to travel to a place whose language is unknown to use. So, translation can help us with the language.
7. Product Recommendations
Whenever you visit some e-commerce website such as Amazon, it shows you items that match your tastes under the banner “Items you may like”. You might even receive emails for shopping suggestions based on your previous buying/ exploring behavior.
OTT platforms such as Netflix, Voot, etc. makes use of recommendation algorithms for suggesting next watches to users and thereby increasing user’s experience and thus profits of company. From the fact that, 60% of Netflix business is from recommendation algorithm, you can imagine the huge impact ML have on online businesses.
8. Online Fraud Detection
Now-a-days, cyber crime such as fraudulent transactions by stealing login credentials, fake accounts etc often occurs leading to degrading customer’s trust on online transactions. ML provides the solution to this by making any and every transaction secure by comparing million of transactions taking place and differentiate between real and fraud transaction.
9. Spam Email and Malware Filtering
Source: https://www.storyly.io/
Our email account have different folders as spam, important, and inbox. The email we receive automatically gets filtered based on the source from where email is received. The differentiation among categories in which email is kept is based on ML algorithm. Spammers are smart enough to update their code daily so as to reach user’s inbox and mount an attack. New malware introduced are often similar to previous versions with 2-3% changes. ML model built to detect spam emails should be such that it can detect ay spam email irrespective of advanced algorithm used to create it.
10. Medical Diagnosis
Now-a-days, machine learning is used in medical science at large scale for predicting the disease and providing right medication. For example, early heart disease prediction helps save lives of many by providing timely aid to heart patients.
In addition to prediction, medication can also be provided to patients using ML technique. Given the symptoms exhibited in a patient and a database of anonymized patient records, predict whether the patient is likely to have an illness. A model of this decision problem could be used by a program to provide decision support to medical professionals.
Detailed working of each application with codes will be discussed in upcoming blogs.
So, Stay Tuned!!
