Connect with us

Technology

Faster, More Accurate Diagnoses: Healthcare Applications Of AI Research

Published

on

When Google DeepMind’s AlphaGo shockingly defeated legendary Go player Lee Sedol in 2016, the terms artificial intelligence (AI), machine learning and deep learning were propelled into the technological mainstream.

AI is generally defined as the capacity for a computer or machine to exhibit or simulate intelligent behaviour such as Tesla’s self-driving carand Apple’s digital assistant Siri. It is a thriving field and the focus of much research and investment. Machine learning is the ability of an AI system to extract information from raw data and learn to make predictions from new data.

Deep learning combines artificial intelligence with machine learning. It is concerned with algorithms inspired by the structure and function of the brain called artificial neural networks. Deep learning has received much attention lately both in the consumer world and throughout the medical community.

READ MORE | The Robots Are Coming: How To Stay Relevant And Re-Skill Yourself

Interest in deep learning surged with the success of AlexNet, a neural network designed by Alex Krizhevsky that won the 2012 ImageNet Large Scale Visual Recognition Challenge, an annual image classification competition.

Another relatively recent advancement is the use of graphical processing units (GPUs) to power deep learning algorithms. GPUs excel at computations (multiplications and additions) needed for deep learning applications, thereby lowering application processing time.

In our lab at the University of Saskatchewan we are doing interesting deep learning research related to healthcare applications — and as a professor of electrical and computer engineering, I lead the research team. When it comes to health care, using AI or machine learning to make diagnoses is new, and there has been exciting and promising progress.

Extracting blood vessels in the eye

Detecting abnormal retinal blood vessels is useful for diagnosing diabetes and heart disease. In order to provide reliable and meaningful medical interpretations, the retinal vessel must be extracted from a retinal image for reliable and meaningful interpretations.

Although manual segmentation is possible, it is a complex, time-consuming and tedious task which requires advanced professional skills.

My research team has developed a system that can segment retinal blood vessels simply by reading a raw retinal image. It is a computer-aided diagnosis system that reduces the work required by eye-care specialists and ophthalmologists, and processes images 10 times faster, while retaining high accuracy.

READ MORE | Businesses Of The Future: 20 New Wealth Creators On The African Continent

Detecting lung cancer

Computer tomography (CT) is widely used for lung cancer diagnosis. However, because visual representations of benign (non-cancerous) and malignant (cancerous) lesions in CT scans are similar, a CT scan cannot always provide a reliable diagnosis. This is true even for a thoracic radiologist with many years of experience.

The rapid growth of CT scan analysis has generated a pressing need for advanced computational tools to assist radiologists with the screening progress.

To improve radiologists’ diagnostic performance, we have proposed a deep learning solution. Based on our research findings, our solution outperforms experienced radiologists. Moreover, using a deep learning-based solution improves diagnostic performance overall and radiologists with less experience benefit from the system the most.

READ MORE | Meet The Top Women Investors Of The Midas List In 2019

Limitations and challenges

Although great promise has been shown with deep learning algorithms in a variety of tasks across radiology and medicine, these systems are far from perfect. Obtaining high-quality annotated datasets will remain a challenge for deep learning training. Most computer vision research is based on natural images, but for healthcare applications, we need large annotated medical image datasets.

Another challenge from a clinical standpoint will be the time to test how well deep learning techniques perform in contrast to human radiologists.

There needs to be more collaboration between physicians and machine learning scientists. The high degree of complexity of human physiology will also be a challenge for machine learning techniques.

Another challenge is the requirements to validate a deep learning system for clinical implementation, which would likely require multi-institutional collaboration and large datasets. Finally, an efficient hardware platform is required to ensure fast processing of deep learning systems.

In the complex world of healthcare, AI tools can support human practitioners to provide faster service and more accurate diagnoses, and analyze data to identify trends or genetic information that may predispose someone to a particular disease. When saving minutes can mean saving lives, AI and machine learning may be transformative for healthcare workers and patients.

Seokbum Ko; Professor, University of Saskatchewan

The Conversation

Continue Reading
Advertisement
Comments

Technology

Google Is Making Android As Difficult To Hack As iPhone—And Cops Are Suffering

mm

Published

on

By

Apple makes the most secure phones in the world. At least, that’s the common assumption.

But increasingly Android models are as difficult, if not trickier, to break into as the iPhone, according to search warrants and forensic industry sources.

One warrant unearthed by Forbes detailed a case in which cops seized an LG Android phone after swooping on suspected drug dealer, Angel Angulo, who’d allegedly sold methamphetamine to an undercover cop. Upon Angulo’s arrest the police obtained the phone from inside the his Ford Mustang, though the search warrant didn’t detail what exact LG model.

READ MORE | The World’s Most Reputable CEOs 2019

The police obtained the LG in January and even had permission to force open the device by holding it up to Angulo’s face or depressing his fingerprint to unlock with facial or fingerprint recognition.

But agents at the Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) have not been able to find a way to bypass the lock screen with their current forensic tools and techniques.

They’ve now asked for an additional 120 days to access the device, according to a filing from the ATF. (No legal representation was listed for the defendant. The Department of Justice hadn’t responded to a request for comment at the time of publication).

READ MORE | Huawei Lawyer Says CFO Meng A ‘hostage’ After US Presses Charges

iPhones less secure than Android now?

Sources in the forensics community tell Forbes that the security of Google’s operating system has become, in some cases, too strong to allow police (or anyone else with direct access to an Android phone) inside. “Some would say iPhones are less secure,” said one source from a large forensics provider.

Google, like Apple, has been continually adding security features to Android, said Vladimir Katalov, CEO of Russian forensics provider Elcomsoft. One of the key updates to Android over the years remains Secure Startup, which encrypts all internal storage so the data within should be accessible only to someone with a passcode or other form of authentication, like a face or a finger.

The anonymous source noted that a considerable amount of resources go into penetrating the defenses of iPhone operating systems, and once a hack works on one iOS device, it should work on all the others. The same cannot be said for Android, which is fragmented across manufacturers’ home-brew updates and countless phone models. If a police officer or forensics professional finds a way to break into a Google Pixel, for instance, the same hack might not work on any other devices.

READ MORE | Google, Facebook, Twitter Fail To Live Up To Fake News Pledge

“While some Android phones can be accessed using generic methods, each new model may contain unique features which require a more customized approach,” said Peter Sommer, professor of digital forensics at Birmingham City University in the U.K. “In addition there are instances of standard Android phones being heavily modified to have secure features.”

The issue of Android security is particularly pressing for Huawei, the world’s largest manufacturer of Android phones, which is set to lose access to Google’s security updates on newer devices. That’s because of the Trump administration’s ban on American companies doing business with the Chinese telecoms giant.

iPhone still trumps Android in some cases, though. When Forbes tested a range of Android phones’ facial recognition systems, there were clear disparities in the security of the technology across the different devices. A 3D-printed head was able to open all the Android phones tested, but it worked much quicker and in all light conditions on a OnePlus 6, whereas Samsung and LG phones were tougher to crack. Forbes also tested Apple’s facial recognition with the fake head. In that case, it proved impenetrable.

Thomas Brewster;Forbes Staff

Continue Reading

Technology

The 4IR Strategy To Move Forward

mm

Published

on

South Africa has created the Presidential Commission on the Fourth Industrial Revolution. The President of South Africa, Cyril Ramaphosa, chairs this commission and I am the Deputy Chairman. The commission consists of 30 members.

It is tasked with developing South Africa’s strategy around the Fourth Industrial Revolution (4IR).

Why this commission and why is the 4IR so important? To answer these questions, we ought to understand first what the 4IR is and what it entails.

As a point of departure, we ought to understand the first, second and third industrial revolutions. The British scientific revolution that gave us Isaac Newton’s laws of motion, gravitation and the study of heat, catalysed the first industrial revolution.

READ MORE | South Africa’s Informal Sector: Why People Get Stuck In Precarious Jobs

The first industrial revolution gave us coal-powered steam engines that drove trains and ushered the era of using machines to produce goods and services. The luddites, who were a group of activists hell-bent on stopping the first industrial revolution, disappeared into the ash heap of history and the technological revolution marched on.

The scientific revolution that facilitated us to leverage magnetic and electrical forces called electromagnetism ushered the second industrial revolution. Electromagnetism gave us electricity and an electric motor, and these, in turn, spurred the development of the assembly line, which vastly improved production and introduced mass production in factories.

The discovery of semiconductors and the invention of a transistor in the United States (US) ushered the electronic age, which gave us computers, mobile phones, and ultimately led to the invention of the internet.

Seventy years after the discovery of a transistor, there is no significant semiconductor company on the African continent. To solve this serious gap in the political economy of the African continent, industries, society and universities on the continent must come together to develop a strategy and the plan of action on how the continent should enter the primary economy of the third industrial revolution.

READ MORE | The Robots Are Coming: How To Stay Relevant And Re-Skill Yourself

The 4IR is a confluence of digital technologies of the third industrial revolution, developments in biotechnology through innovations such as molecular motors, as well as the developments in the physical space through breakthroughs in new materials and robotics technologies.

One of the technologies driving the 4IR is artificial intelligence (AI). Because of AI, machines are gaining intelligence and thus planes are intelligently flying with minimal interference by pilots, the factory floor is increasingly automating, and self-driving cars are reaching maturity. AI is revolutionizing the medical field and making careers such as radiology redundant.

The implications of the 4IR will be extensive. It will reduce the world of work and usher a post-work era, where factories will employ fewer people than ever before. The presidential commission should study this post-work era and the impact thereof on the labor market, equality and tax collection.

This commission will explore the viability of strategies such as the introduction of universal basic income, virtual economic zones and imposing tax on robots in order to prevent the escalation of poverty and inequality.

In the 4IR era, some jobs will disappear, some will change and new types of jobs will be created. It is important that the presidential commission study these changes and design a strategy on how we should move forward. This should include the identification of the required set of skills in order to thrive in the 4IR, and the mechanisms in which the education sector should plan for the development of such skills.      

READ MORE | Educated, Re-engineer and Lead Through Uncertainty

Kai-Fu Lee, the Chairman of the World Economic Forum Council on Artificial Intelligence, wrote in his influential book, AI Superpowers, that in the 4IR, data is becoming the new oil. Countries and companies, such as the US, China, Google, Facebook, and Alibaba, which collect more data will become so powerful that they will influence all aspects of our lives.

The company Cambridge Analytica allegedly influenced the US election through data analytics.

This commission should study and recommend how South Africa should position itself with respect to defence, peace, security, industry, trade, society and politics in the light of the 4IR. Failure to capture this 4IR moment will relegate us to the dustbin of history.

– Tshilidzi Marwala is a professor and Vice-Chancellor and Principal of the University of Johannesburg in South Africa.

Continue Reading

Technology

‘A Tweet Can Tank An Economy’

mm

Published

on

Many perspectives have been shared on influencer marketing, but very few from the brands themselves. What are its business imperatives and ROI?


In the heart of Sandton, Africa’s richest square mile, is Brittany Preece, a social media manager at Investec Bank.

As part of the Private Banking marketing team, Preece and her team have managed to successfully implement disruptive digital strategies to meet the growth objectives of the bank.

Traditionally, the role of marketing has been to support business objectives. Yet when companies experience financial difficulty, more often than not, marketing is seen to be the first in line on the chopping block. As a way of adapting to this reality, most marketing managers have had to look at cost-effective, yet creative, ways of achieving business impact.

READ MORE | How Cryptocurrency Scams Work

Social and digital media have given room for new ways businesses can engage and sell to their customers.

Increasingly gaining traction in South Africa is influencer marketing, an entirely new branch of the sales and marketing funnel which has seen brands leverage popular personalities on social media to promote their products and services online. Globally, the influencer marketing industry is forecast to be worth $5 billion to $10 billion by 2020, according to a study by Mediakix.

For Investec, attracting and diversifying into new markets has informed their decision to leverage influencers as part of their marketing strategy.

“Since 2015, we as Investec, have predominantly used influencer marketing to reach the young professional audience,” Preece says.

READ MORE | Google, Facebook, Twitter Fail To Live Up To Fake News Pledge

“We (Investec Private Bank) have recently evolved our qualifying criteria. We’re no longer just the preferred banking partner for accountants, engineers, lawyers, doctors and actuaries. We’re positioned to be the bank of choice for those who are under the age of 30, consistently earning more than R600,000 ($42,582)a year, with a university degree and working in their area of expertise.”

Seated in front of a glass wall which looks out to a view of a bustling and upbeat office environment, she elaborates that to effectively reach the young professionals’ target market, it is important to identify individuals that can speak to that audience.

“What influencer marketing gives you, that a lot of the other marketing channels can’t, is great reach and engagement with your target audience. Using [influencers] brings a lot of authenticity to the brand, where Investec Private Banking was seen as unattainable to a lot of young professionals,” Preece says.

READ MORE | Facebook Recommits To A More Personalized Dating App, Your Privacy

In the digital universe, content is king. Authenticity is paramount. A study conducted by the Mobile Marketing Association reported that most consumers have banner blindness and suffer from advertising fatigue. They cannot recall the last digital banner ad that they saw.

Ad-blocking continues to be a growing phenomenon, further demonstrating the shift in consumer behavior. Consumers simply don’t want to be marketed to. To this effect, influencers assist brands to reach their customers through content that is more relevant to those consumers.

According to Business Insider’s 2018 Influencer Marketing Report: Research, Strategy & Platforms for Leveraging Social Media Influencers, the average social media engagement rate from influencer marketing averages 5.7% per post.

This is a feat compared to content generated by brands which fluctuates around 2% – 3% per post.

With this in mind, it’s even more essential for brands to involve influencers upfront to co-create an envisioned campaign.

“You want your influencer to buy into what you stand for as a brand. It shouldn’t be a hard sell – consumers see through that. If not done properly, it can look very fake and [we] never want to make it feel like an ad,” Preece says.

READ MORE | Crazy Aviators: The Eerie Similarities Between Billionaire Howard Hughes And Elon Musk

Brands and agencies measure awareness and engagement by analyzing a few parameters, including, inter alia, number of likes, comments, shares and followers. Proponents of this theory believe an increase in these parameters lead to brand success.

Detractors of the theory believe awareness and engagement alone are not sufficient indicators of brand success. “Likes and follows are unimpressive without an increase in the bottom line,” says Sinesipho Maninjwa, a financial analyst and commentator.

But technology has its limitations. Dianne Joseph, the Commercial Director of Digital at Nielsen, a leading global information, data and measurement company, elaborates that there are real challenges in measuring sales attribution as a result of influencer marketing.

“The influencer models are difficult to measure because most influencers use their own personal pages to post, and these organic posts don’t offer us the opportunity to implement a tag (for tracking purposes) and, therefore, we aren’t able to track the outcomes,” Joseph says.

The ability to measure what is happening on and off digital platforms, or across social media platforms, is a challenge that most digital marketers face. Different platforms such as Facebook and Twitter have their own unique methods of tracking and storing data. The link between these platforms is, therefore, a lot more tenuous.

But in support of influencer marketing, a number of global studies have come out and shown that influencer marketing does yield a positive return on investment (ROI).

Pierre Cassuto, of social media influencer platform Humanz, who recently set up shop in South Africa from Tel Aviv in Israel, believes that the use of influencers can lead to increased sales. “In the US market, there is about a four times ROI on money spent in the influencer marketing space for retailers, cosmetics and FMCG brands,” he adds.

Cassuto agrees that engagement on its own doesn’t yield conversion. For influencer marketing to yield conversion, the message being broadcast by influencers has to be well thought out and carefully constructed to achieve that objective.

There has to be a reason for someone to do something now as opposed to anytime. That can be around creating urgency around something specific, it can be around creating a limited time availability offer.

Essentially, brands can become creative and purposeful in how they design their entire campaign to track conversion from the ground up. “The way we suggest measuring conversion is by making sure that the influencers have a way that [brands] can track the relationship between the influencers post and the actual conversion,” Cassuto says.

The first looks at providing the influencer a coupon code that they can distribute and share. In doing so, the specific brands will know that the people who redeemed the coupon code came from the influencers. The second options measures sell over time related to increase in traffic due to influencer marketing towards a landing page that the brand is driving to. With this, Cassuto states that brands can do A/B testing to see the impact on days influencers are posting and days when they’re not.

With all the advantages of influencer marketing, should brands solely invest in this sales channel?

Preece firmly holds the view that influencer marketing on its own isn’t the be all and end all. “I do believe that it should be a mix. I believe that an integrated marketing plan that uses out-of-home, billboards, TV and digital is where you see the most results.”

“Things can go wrong with influencer marketing. You are putting faith and trust into a human being who is imperfect. We have a saying: A tweet can tank an economy,” Preece says.

With consumers becoming increasingly selective about what they consume, influencer marketing continues to grow as an attractive sales channel. It is imperative that brands place the necessary care and due diligence before partnering with any influencer.

The perfect fit between an influencer and the brand will determine the overall ROI. After all, effective influencer marketing is the online equivalent of the highly valuable word-of-mouth advertising that marketers have always coveted.

Continue Reading

Trending