Emerging Trends

Written and reviewed by Rohit Singh, Computer Science & Commerce faculty at UniDrill (MCA, Osmania University · 6 years teaching). Based on NCERT Informatics Practices Class XI, Chapter 2.

How does your phone know about the traffic jam?

Open a map app and it suggests a route that avoids a jam that formed ten minutes ago. Upload a group photo and the faces get tagged without you typing a name. Nobody wrote a rule for each road or each face. The software learned from data. This chapter covers six ideas behind that kind of magic: AI, Big Data, the Internet of Things, cloud computing, grid computing and blockchains.

Emerging trends are state-of-the-art technologies that gain popularity and set a new trend among users. Many new technologies fade away. Some stay and shape the digital economy and the way societies interact online.

Artificial Intelligence: machines that imitate thinking

Artificial Intelligence (AI) tries to simulate the natural intelligence of human beings in machines, so that they behave intelligently. An intelligent machine imitates cognitive functions such as learning, decision making and problem solving. The aim is to let machines perform tasks with minimum human intervention.

An AI system is programmed to build a knowledge base, a store of facts, assumptions and rules, and to take decisions from it. It can also learn from past experiences or outcomes to take new decisions. Intelligent digital assistants such as Siri, Google Now, Cortana and Alexa are powered by AI, and so are route suggestions based on live traffic and automatic tagging in photos.

Four areas sit under the AI umbrella in this chapter.

Machine Learning: learning from data

Machine Learning (ML) is a subsystem of AI in which computers learn from data. Instead of a human writing every rule, the program contains algorithms that use data to learn on their own and make predictions. These algorithms are called models.

The learning has a pattern worth memorising because it is easy to scramble in an exam:

The ML cycle in one line: training data trains the model → testing data checks it → repeat until accuracy is acceptable → use it to predict on new, unknown data.

Why train and test on different data? Imagine a student who memorises last year's paper and scores full marks on it. That proves memory, not learning. Keeping separate testing data shows whether the model can handle data it has never seen.

Pause and predict: a spam filter has been trained on thousands of labelled emails. Which data is used to check whether it works, and which kind of data does it finally predict on?

Check your answer: the testing data checks it, and new unknown emails are what it finally predicts on. Models are trained with training data, tested with testing data, and used on unknown data once their accuracy is acceptable.

Natural Language Processing: talking to computers in Hindi or English

Natural Language Processing (NLP) deals with the interaction between humans and computers using human spoken languages such as Hindi or English. It makes it possible to search the web or control a device by voice.

An NLP system can do text to speech and speech to text conversion. Machine translation, where machines translate text from one language to another with fair correctness, is a fast growing field. Another is automated customer service, where software answers customer queries and complaints.

Beyond the textbook. For the exam: predictive typing and spell checking are the standard NLP examples. In reality: the same family of techniques handles search suggestions, subtitles on videos and voice assistants, and it improves as it sees more data. Why it matters: the examiner's wording is usually the chapter's wording, so recall predictive typing and spell check first, then add the others.

Immersive experiences: VR and AR

An immersive experience stimulates your senses so that you can visualise, feel and react. It makes interaction more realistic and engaging, and it is used in training such as driving and flight simulators.

Virtual Reality (VR) is a three-dimensional, computer-generated situation that simulates the real world. You explore and interact with it by being immersed in it, at present with the help of VR headsets. To feel more real it can add sound, smell, motion and temperature. It is used in gaming, military training, medical procedures, entertainment, psychology, engineering and other fields where simulation helps learning.

Augmented Reality (AR) is the superimposition of computer generated perceptual information over the existing physical surroundings. It does not create a new world. It adds digital components to the real one. A location based AR app lets a traveller point a camera at a historical place and see real time information about it.

Why it matters. A quick test: if you can still see your real room in the experience, it is AR. If your room disappears and a new world replaces it, it is VR.

Robotics: programmable machines

A robot is a machine that carries out one or more tasks automatically with accuracy and precision. What makes it different from other machines is that it is programmable: it follows instructions given through computer programs. Robots were first imagined for repetitive, boring, stressful or labour intensive industrial work. Sensors are one of their prime components.

Robotics is an interdisciplinary branch needing mechanical engineering, electronics and computer science, and is concerned with the design, fabrication, operation and application of robots. It is used in industry, medical science, bionics, scientific research and the military.

Drones are used in aerial photography, short distance delivery, disaster management, search and rescue, agriculture, wildlife monitoring, mapping, structural safety inspection, law enforcement and border patrol.

Big Data: too much, too fast, too messy

Data is being produced at a colossal rate. At the current pace, around 2.5 quintillion bytes of data are created each day, and the rate keeps increasing with the Internet of Things. The result is data sets of enormous volume and complexity called Big Data.

Such data cannot be handled with traditional data processing tools. It is not only huge but also unstructured: posts, chats, photographs, tweets, blog articles, news items, polls and comments, audio and video chats. It brings challenges of integration, storage, analysis, searching, processing, transfer, querying and visualisation. It matters because it can hold rich information of high business value.

The five Vs. These five characteristics distinguish big data from traditional data. Every year examiners test them.

Worked example: which V is it? A weather service collects readings from thousands of sensors every second, in numbers, images and text reports. Some sensors give faulty readings. Name the Vs.

The common wrong step is to say "messy data" is Variety. Messy and untrustworthy is Veracity. Variety is about different forms of data, such as text with images.

Data analytics is the process of examining data sets to draw conclusions about the information they contain, with the help of specialised systems and software. Businesses use it for informed decisions, and researchers use it to confirm or reject models, theories and hypotheses. Pandas, a Python library, makes data analysis much simpler.

Beyond the textbook. For the exam: big data is defined by the five Vs. In reality: the exact count of Vs varies in different books, and some lists add more. Why it matters: answer with the five that the chapter gives (Volume, Velocity, Variety, Veracity, Value) and do not be thrown by a longer list in another source.

Internet of Things: when bulbs and fridges join the network

A computer network connects laptops, desktops, servers and smartphones. The Internet of Things (IoT) is a network of devices that have embedded hardware and software to communicate, that is, connect and exchange data, with other devices on the same network. Today many household devices have advanced hardware (microcontrollers) and software but work in isolation and need a person to give directions and inputs. IoT brings them together so they collaborate and form an intelligent network. For example, a microwave, air conditioner, door lock or CCTV camera connected to the Internet can be controlled remotely from a smartphone.

Web of Things (WoT). With IoT you may need a different app for each device. WoT uses web services to connect anything in the physical world, giving a single interface through the web. It paves the way for smart homes, offices and cities.

Sensors. Rotating your phone changes the screen from vertical to horizontal because of two sensors. The accelerometer detects the orientation of the phone. The gyroscope (gyro) tracks rotation or twist and adds to the accelerometer's information. A smart sensor takes input from the physical environment, uses built in computing resources to perform predefined functions on detecting specific input, and processes the data before passing it on.

Pause and predict: a phone's screen turns sideways when you tilt it. Which sensor detects the orientation, and which one tracks the twist of your hand?

Check your answer: the accelerometer detects the orientation, and the gyroscope tracks the rotation or twist.

Smart cities. A smart city uses computer and communication technology with IoT and WoT to manage and distribute resources efficiently. Sensor based examples: a smart building senses earthquake tremors and warns nearby buildings, a smart bridge uses wireless sensors to detect a loose bolt, cable or crack and alerts the authorities by SMS, and a smart tunnel detects leakage or congestion and sends the data to a central computer for analysis. Transport, power plants, water supply, waste management, law enforcement, schools, libraries and hospitals work together to improve city efficiency.

Cloud computing: services like electricity

Cloud computing delivers computer based services such as software, hardware (servers), databases and storage over the Internet, so that they can be accessed from anywhere on any smart device. The companies that provide them are called cloud service providers, and they usually charge pay per use, much like electricity. You already use the cloud when you back up pictures and files on the Internet or host a website.

Why it works. A user can run a bigger application or process a large amount of data without owning the storage or processing power, as long as the Internet is connected. It is cost effective and gives on demand, need based resources.

The three service models. Understand the cloud as "everything as a service".

The Government of India has taken an initiative called GI Cloud, named MeghRaj.

Worked example: pick the model. A company wants basic services such as virtual server provisioning and on demand storage that it can combine into a platform to deploy and run its own customised applications. Which model fits?

Pause and predict: you edit a spreadsheet in your browser and never install anything. Which cloud model is this?

Check your answer: SaaS. You use ready made application software on demand and are not concerned with installing or configuring it.

Grid computing: a virtual supercomputer

A grid is a network of geographically dispersed and heterogeneous computing resources. Unlike the cloud, whose main focus is providing services, a grid is more application specific and creates the sense of a virtual supercomputer with enormous processing power and storage. Its resources are called nodes, and they temporarily come together to solve one large task and reach a common goal.

Because countless mobile devices, PCs and workstations are connected through LAN or the Internet, it is economical to reuse their memory and processing power. Grids solve computationally intense scientific problems without buying costly hardware.

A middleware is required to set up a grid. The Globus toolkit is one open source toolkit used for building grids, with software for security, resource management, data management, communication and fault detection.

Why it matters. IaaS looks like grid computing because both give you computing power. The difference is in who supplies it. In IaaS a provider rents infrastructure to you. In a grid, many nodes pool their own resources to solve a common problem.

Blockchains: a ledger nobody owns alone

Traditionally, transactions on a ticket booking site or in a bank are stored in a centralised database and updated one by one. Because everything is in one place, the data may be hacked or lost.

Blockchain works on a decentralised, shared database, where each computer holds a copy. A block is a secured chunk of data or a valid transaction. Each block has a header that every node can see, while only the owner can access the private data of the block. Blocks link together into a chain.

Formally, a blockchain is a system that lets a group of connected computers maintain a single, updated and secure ledger. It is an append only open ledger, updated only after all the nodes authenticate the transaction. Every node receives a full copy.

The transaction flow in one line: someone requests a transaction → request is broadcast to all nodes → all nodes authenticate it → the block is added to the existing chain → transaction complete.

Why it is secure. Every member keeps a copy of the whole blockchain, so no single member can change the data alone. To cheat you would have to alter the records on every computer at the same time.

Its most popular application is digital currency. Because it is decentralised, open and secure, it is also being considered for transparency and accountability in business and governance: healthcare data sharing, land registration records to avoid ownership disputes, and voting systems to avoid vote alteration. It can be used in banking, media, telecom, travel and hospitality too.

Beyond the textbook. For the exam: blockchain is a decentralised, append only, shared ledger where all nodes authenticate transactions. In reality: "blockchain" and digital currency are not the same thing. Digital currency is one application, and the underlying ledger idea can be used for records of any kind. Why it matters: questions ask for the technology, not the currency.

Exam traps

60-second revision

Where this chapter connects

• Previous: Computer System explains the hardware and software that every technology here runs on.
• Next: Brief Overview of Python starts the programming half of the book. Pandas, mentioned under data analytics, is a Python library.
• Related: Understanding Data revisits data collection, structured and unstructured data, and the statistics used to analyse them.