Month: July 2021

eSIMs

Shabbir Chunawalla

IDEMIA, a leading French SIM card maker makes in India standard SIM cards — its production capacity is over 600 million SIMs, which accounts for 67 per cent of its global production. It has its largest plant in Noida. It has provided over a billion standard SIMs to Indian customers, and controls 40 per cent of this market. It has manpower strength of over 5000 employees. It has its global R&D centre in India. It has worked with Aadhar programme for its biometrics.

It now intends to enter embedded SIM (eSIM) market. This is a next generation technology. An IC card is embedded in a mobile device. Consumers here can change their networks remotely or through a QR code. They need not wait for a physical SIM to come to their home or office.

It intends to invest heavily in India to set up the eSIM manufacturing facility.

eSIM devices would make it easier for the telecom customers to switch to a local network while travelling abroad. They need not buy a local SIM.

These cards can be used in machine-to-machine IoT domain, especially in connected vehicles.

There is going to be some resistance to this new product. It is likely to wane in time to come. Telcos may not push eSIMs aggressively as customers switch networks easily. In the next few years, some 30 per cent devices will be powered by eSIMs. The market is expected to witness a growth rate of 30 per cent.

Clubhouse

Shabbir Chunawalla

Clubhouse was initiated as a fun app. It had chats on anything, say best food to music to movies. It acquired popularity in India. It has 2-6 million users. From Apple’s iOS it has extended to Android. However, the fun and banter are being replaced by bigotry and intolerance. It reveals locker room conversations which are unpalatable. The onus of moderation is on users. The quality of debate has been deteriorating. There are discussions on contemporary politics and political rivalry. The plus point is that youngsters have found a voice. There can be hostility in the discussions towards those upholding a counterpoint. Political leanings decide the tenor of debate. There is scope for hate speech. Some clubs have civil conversations. The discussions are rooted in facts and figures. On an audio app, there is no scope of trolling, but the negativity spills over outside. Clubhouse should set guidelines to operate and moderate. The issue is that the app should not become polarised just like other social media. Though other platforms can take corrective steps, here it being an audio medium, it is pretty difficult.

Open Source

Shabbir Chunawalla

Open Source refers to open source software (OSS) which is publicly accessible. It is developed in a decentralised and collaborative way. Anyone can see, modify or distribute the code. It relies on peer review and community production. Such software is flexible and economical. It lasts longer than its proprietary versions. It saves costs.

Open sourcing allows people to suggest improvements or point out issues. They also use the code under a license to develop similar products. Indian government’s policy allows the use of source code of any government application and such use will be free from royalty. Co-Win, India’s digital vaccination platform, is to be open source. Red Hat is the world’s leading open source software solutions company. Another name for open source is FOSS — free and open-source software. Aadhar and UPI use open source technology. GST too uses it, but the code has not been opened for the larger community to assess. The government fears that opening up the entire code of an app or platform makes it vulnerable to attack.

In proprietary software there is black-box testing. In open source, there can be much more scope than black-box testing as the code is open. A fresh examination can identify securities issues better.

REvil (Ransomware Evil) Goes Off-line

Shabbir Chunawalla

As we know, there are ransomware groups who attack the targets and extract money. The US at the highest level had demanded that Russia should shut down such groups. REvil, short for Ransomeware Evil, the most aggressive of the groups, went offline in July 2021. The group’s sites on the dark web suddenly disappeared. Happy blog — their publicity material listing the victims and earnings — too was gone. They have virtual conference rooms to negotiate with the victims. They too disappeared. May be, the US asking their Cyber Command and FBI to take down the sites worked. Maybe, Russia at the highest level ordered its shutting down. May be REvil itself took the decision fearing the consequences when two powers were after it.

Digital IDs for Digital Currencies

Shabbir Chunawalla

China is soon going to release e-CNY, their official digital currency issued by the central bank. The currency is not mere token. The monetary authority or some private players maintain credit-debit accounts. These accounts must be identified to prevent fraudulent transactions. How to share these identities across the borders? Can the central banks of another country do it? Will it be acceptable?

One way to solve this is to have compatibility of technical and regulatory standards between the two countries. Secondly, there could be interfaces of their systems. Here there are no middlemen. Thirdly, many countries can come together on one platform to manage their digital currencies. These approaches involve identify establishment but that happens at the national level.

Another way is to have a jointly operated payment system. This system supports diverse digital currencies.

To facilitate exchange, a competent national authority must verify you as an individual and your balance amount. The country accepting foreign digital currency approves the transaction since it already is satisfied by the anti-money laundering standards of the issuing country. Thus two wallets in two different currencies can settle their accounts without telling each other more about the clients identity.

In the absence of suitable systems, the digital currencies will operate in silos, and these will be non-starters. It may so happen that one country’s digital cash is more preferred at some places than others.

Edge Computing

Shabbir Chunawalla

In autonomous cars, there are sensors and cameras which transmit the data to the data centre. Here the data is analysed and sent back to the car, and the appropriate decision is taken. Due to faster internet speeds this has become possible. Still, there are situations where instant decisions are needed. At times, there is no internet too. In such situations, we rely upon edge computing .

To make decisions, edge computing uses processors which are powerful enough to store the data and process it immediately by themselves. These can be used for specific applications such as voice, video and image processing. This is called edge computing. It also saves bandwidth. Here data does not travel through net. In image processing and facial recognition, you can save substantial bandwidth. The edge processor can later send the processed data to the data centre for further analytics.

Industry 4.0 can use edge computing. Here the enterprise operations are connected. There is intelligent decision-making in real time. In ITC plants camera inputs are analysed to exercise quality control over the products. It reduces the quantum of defective products. There are AI models in edge computing. Edge computing can be used in predictive maintenance.

It is expected that a major chunk of data in the next five years will be processed outside a traditional data centre or cloud.

Bug Bounty Hunters

Shabbir Chunawalla

There are cyber attacks all over the world, including India. Such attacks in India run into millions. On account of the flood of attacks in recent times, we should focus on people, process and technology. Though there are professional firms of cybersecurity, there is also an army of ethical hackers who are certified security professionals or security researchers. They are called ‘bug bounty hunters.’ They crawl the web and scan the systems of various organisations. They alert the organisations if there is vulnerability in their systems. They are rewarded in cash or kind by the organisations. Their services are availed of, even if the organisation has its own internal security team. There are vulnerabilities on government sites too. The bounty programme was initiated in the US and Europe in the 80s. Social media companies leverage it. Some organisations do not understand the work these ethical hackers do. They are white-hat people and yet they are treated as black-hat hackers. Some companies are in denial mode. India needs 5 lac plus ethical hackers in the next five years, where as at present it has merely 70,000.

Data breaches happen mainly due to weaknesses in the technology. Though organisations spend a lot to build robust systems, they ignore the two other equally important components — human error and third-party service providers.

AI and Endoscopy

Shabbir Chunawalla

Endoscopy is widely used in gastrointestinal investigations. It has been enhanced now by AI and ML. Software that can detect and flag abnormalities in the GI tract is being used. The machine is fed with thousands of images of cancerous growths and it learns over a period of time to detect such growths accurately. Such growths are recognised by the computer and an alert is given. Earlier detection is an advantage. The software does this by using computer vision. At present, this is used to detect colonic cancer but in future this can be extended to include different parts of the body such as stomach, liver and pancreas. Already, AI is being used in ophthalmology, radiology and dermatology. These are early days in gastro-enterology. The first such device has been approved in April 2021. Many more trials are happening. Many more such devices will be approved in future.

NetraAI : Cloud-based AI Solution

Shabbir Chunawalla

Intel is in the forefront to use AI to deliver transformative healthcare solutions. NetraAI is cloud-based AI solution which uses deep learning to identify retinal conditions in a short span of time with the accuracy level of ophthalmologists. In India, we have a shortage of trained retinal specialists in remote and interior areas. It constrains the screening of asymptomatic patients. They come late with advanced diabetic eye disease.

A Singapore IT firm Leben developed NetraAI for Sankara Eye Foundation. It assesses retina risk comprehensively. It is offered as software-as-a-service platform. It analyses images from fundus camera devices for immediate results of referable DR grading via a cloud-based web portal.

It uses advanced algorithms with a four-step deep convolutional neural network. This helps in detecting DR stage and annotating lesions based on pixel density in the fundus images. This can be extended to other retinal conditions and glaucoma. It is powered by Intel. Xeon Scalable processors and built-in Intel Deep Learning (DL) Boost accurately detect DR.

AI’s History

Shabbir Chunawalla

Cade Matz has written a book Genius Makers which traces the history of Artificial Intelligence (AI) and how it reached its presence status.

What is currently known as AI originated from the idea of neural networks in the 1940s. It was studied how the neurons in the brain functioned. It was studied whether its electronic version could be created.

Frant Rosenblatt, psychology professor, Cornell, demonstrated how a computer could learn to distinguish simple patterns in mid-1950s. This creation was called Perceptron, and was promoted in media. It, however, had little practical application.

Marvin Minsky, Rosenblatt’s contemporary, wrote a book. It proved self learning systems simulating neural networks were useless. He also explored the field of neural networks, but was convinced that this is not the way to go.

Minsky, John McCarthy and Rochester proposed the term Artificial Intelligence in a convention. They proposed Symbolic AI which could teach computer to do specific things by giving very specific instructions. This idea eventually prevailed, and variations of the Symbolic AI evolved over several decades.

Geoff Hinton, however, did not give up on neural network, and his research assisted by his many students led to the field of Deep Learning. Yuan Lecun, a French origin computer scientist who shifted to Silicon Valley contributed to many breakthroughs.

Google hired Hinton and his assistants. Facebook hired Lecun. MS too spent a lot on AI research. It fell behind, as it had initially backed Symbolic AI. It had to catch up with Deep Learning.

GPUs gave AI research a huge boost. Deep Mind was working on AI. Google bought it.

Google Brain and Deep Mind had a rivalry. Google Brain concentrated on using Deep Learning for technologies that could be introduced quickly. Deep Mind worked on abstract work. It was UK-based. Deep Mind beat the world’s best Go Player.

Open AI, an artificial intelligence lab set up by Musk gave away, its research free.