Data Analytics Zone – Ride the waves of Machine Learning and AI

Posted on July 3, 2019

This blog will guide you through

  • AI and Machine Learning
  • Data the root and Mathematics the stem of AI
  • Data Analytics and Machine Learning
  • Deep Learning
  • Applications of Machine Learning

Despite a low tide in the process to unravel the mystery of making the computers behave intelligently as human beings; the time has come to witness a very high tide of developments in the history of mankind. The creation is leveraging its ability to win over the creator. When the creation happens to be a machine then this development depicts the miracle of the iconic pair Machine learning and Artificial Intelligence. The computer systems gain the ability to think and perform like a human without being explicitly programmed.

AI and Machine Learning: These are the buzz words in the technological frontiers of today. Here the human intellectual characteristics such as the ability to reason, discover meaning, a prediction from past experiences, generalize and learn from historical data are ported to computer projects so as to enable the machine to perform intelligently like a human. Though we claim a lot of developments in Artificial Intelligence, as of now we have only reached the level of projects with specific system endowed intellectual process. The road map is much to go from specific intelligence to general intelligence and further to superhuman intelligence. Machine learning could be termed as a part of Artificial Intelligence.

About Origin and growth: The epoch development in the history of computation had its birth in the 1950s with the development of Turing Test by Alan Turing and the development of a game of checkers by Arthur Samuel. The growth rate was not in an exponential pace in its initial stages, but as the data storage capacity and the processing speed of computers reaching a presidential height, Machine Learning and Artificial intelligence has developed into a throb of the pulse in the contemporary technology-driven society.

Data the root and Mathematics the stem of AI: The root node of this technological development is with Data. The ability of the mathematical and statistical process in generating knowledge and insight from data is the core strength behind this development. When this possibility of mathematical science is integrated with the processing capacity of the computers, it has become possible to generate critical insights from huge and multiple data sources with a higher level of accuracy and speed. This combination of data, Mathematics and computer programming form the basis of Machine learning and Artificial Intelligence.

Data Analytics and Machine Learning: In this information society, data is considered as gold and analytics is the process of mining it. In any verticals, whether it is trade – business – industrial world or planning and management of various government portfolios or in socio-economic platforms of society or in life situations strategic decision making is a critical factor that decides its destiny. Insights, Knowledge and visualizations of historical data from these domains will serve as a valid basis for making the decision and proactive action plans. In machine learning strategy the historical data will serve as training data to build appropriate models using a wide range of algorithms. The models that are developed from the training dataset will serve as the input for future predictions, classification or establishing rules of association of the new data. Hence in a machine learning process, the system develops its own models from the available datasets and uses it for future computation, without any explicit program for that purpose.

In a broader sense Machine Learning (ML) is seen as a subset of Artificial Intelligence (AI) and an enabler for Big Data Analytics. According to NASSCOM, the Big Data Analytics application is poised for exponential growth in India. It is expected to witness an eight-fold growth by 2025, to cross $16 billion from the current $2 billion.

Deep Learning: is a category of Machine Learning based on artificial neural network. Deep Learning architecture is applied to numerous fields like natural language processing, audio recognition, speech recognition, computer vision, image recognition and analysis, bioinformatics and game programs. In some cases, the results from deep learning network have proved to be superior to human expertise.

Applications of Machine Learning:

Digital Marketing: We will be familiar with the pop-ups while surfing the webpages, online shopping carts and Facebooks these pop-up recommendations are the application of machine learning models. The large data which are available is analyzed by machine tools and it categorizes the customers with respect to the previous activities and suggests them new products. These functionalities are widely used by most of the e-commerce giants like Amazon, Flipkart etc. The Amazon reports that with the product recommendation system they have enhanced their business by 40%.

Search Engine: Information about the users’ activity are collected and on the basis of that data machine learning model could suggest various topic according to the relevance. This service is very effectively used by Google. We would have noticed when we type a word in Google search different suggestions of the intended search will appear according to the previous search. This is performed by Google’s machine learning algorithm – RankBrain.

Online fraud detection: Here on training the system with the patterns of customer’s financial information’s, the nature of their activities and information about their social network, machine learning model could predict the reliability of customers transactions. Positively, the Machine Learning application thus enables to thwart security breaches by comparing the tractions against account history of the customers, if any unusual activity found will raise red flag which delays the transaction until the user confirmation.

Health Care: This industry is the most potential user of Machine Learning possibilities. It is estimated that the availability of physician for 1000 people is about 1.25. Hence there is a severe dearth in the availability of health services to the community as a whole, moreover, the cases of health issues and epidemic are on a steady increase. In this context, a well-trained machine learning model could supplement the industry as a virtual doctor capable of diagnosis to a higher level of accuracy. Also on examining the health record of the patients Machine Learning model could predict the possibility of a patient to develop some disease with a greater level of accuracy so that the doctors could proactively get involved in remedial treatment.

In a communities, health framework Machine learning model could predict the outburst of some epidemic and provide alert for precautionary measures. Also, ML is also effectively used in bioinformatics and researches.

Financial Sectors: One of the most vibrant financial sectors is the stock market. Here machine Learning model could predict the future fluctuation of stock prices by comparing the principal component of the factors that decide the market prices. With the application of ML, we could build automatic trading technology which makes the trading easier to large and small investors. In recent years ‘Hedge funds” have very effectively implemented machine learning model for predictions.
Insurance and Banking sector has also started implementing ML model for studying the investment trend and also to understand the customer behaviour and fraud detection of automated transactions.

Automatic Language translation: Machine learning could effectively be used in the voice recognition system. It employs text and speech recognition technique and mimics the human language. For example, Apple’s Siri and Microsoft’s Cortana are personal virtual assistants using this technique to understand and respond to your speech. In this model the machine learns from a large dataset of human speech, the model gradually learns to understand and mimic the human language with greater accuracy.

Image Recognition: In this context, machine learning models could be with the capability of computer vision, the ability of software to identify objects, places, writings and even action of images. Here the computer can combine the vision technologies in combination with the camera to achieve image recognition. For instance, if we post a picture on Facebook then immediately it gives suggestions on to whom to tag the photo. The model of image recognition could be effectively used by the police force to detect crimes or identify criminals from a surveillance camera.

Email Smart Reply and spam filtering: Google’s smart reply function is another application of machine learning. Two aspects of machine learning are involved in this case: first, incoming emails serve as a dataset for the system to “learn” the pattern and decipher the meanings of the text. Next, machine learning enables the system to predict possible outcomes and thus automatically suggest options of reply to the emails. 

Likewise, machine learning could be very effectively used to build virtual personal assistants, virtual teachers and chatbot. In this model, the computer program is designed to stimulate conversation with users to get valid information’s or to chart a task that has to be executed. In this process, the students input the text message where the agent executes the task and provides the appropriate messages.

Recently, In Japan, it was found that a farmer using machine learning to sort the grade of the cucumber. Here he developed a deep learning model by training the model with different grade of cucumber. Further, he found that the model is capable to sort out the cucumber into various grades with about 80% accuracy.

In our next blog, we will be discussing the Opportunities and Career path of Data Analytics and Machine Learning in detail

Leave a Reply

Your email address will not be published.