Data Science Insider: January 28th, 2022

In This Week’s SuperDataScience Newsletter: Meta is Building the World’s Fastest AI Supercomputer. AI Develops Map to Locate Meteorites. AI Can Analyse Eye Scans to Identify Patients at High Risk of Heart Attack. TinyML is Bringing DL Models to Microcontrollers. Robot Performs Keyhole Surgery on Pigs. Cheers, - The SuperDataScience Team P.S. Have friends and colleagues who could benefit from these weekly updates? Send them to [this link]( to subscribe to the Data Science Insider. --------------------------------------------------------------- [Meta is Building the World’s Fastest AI Supercomputer]( brief: Mark Zuckerberg has announced his social media empire is building what he claims is the world’s fastest AI supercomputer as part of plans to build a virtual metaverse. In this blog post, the Facebook founder says that the metaverse, a concept that blends the physical and digital world via virtual and augmented reality, will require “enormous” computing power. The AI supercomputer, dubbed AI Research SuperCluster (RSC) by Zuckerberg’s Meta business, is already the fifth fastest in the world, the company said. “The experiences we’re building for the metaverse require enormous compute [sic] power (quintillions of operations/second!) and RSC will enable new AI models that can learn from trillions of examples, understand hundreds of languages, and more,” wrote Zuckerberg in the blogpost. Meta researchers added that they expected the RSC to become the fastest computer of its kind when it is completed in the summer. Why this is important: As data scientists, we know that AI mimics the underlying architecture of the brain in computer form and is capable of processing, and spotting patterns in, vast amounts of data. Meta generates significant amounts of data from its 2.8 billion daily users. It was fined $5bn for privacy violations in the wake of the Cambridge Analytica scandal and some argue that Meta is focusing on expansion into new areas when it should be putting more resources into basic safety systems. [Click here to sign up!]( [AI Develops Map to Locate Meteorites]( brief: An estimated 300,000 meteorites could be sitting undiscovered within the ice fields of Antarctica, according to the findings of a new study. Using AI to predict potential landing sites of pieces of space rock over the past few millennia helped experts from the Free University of Brussels to create a 'treasure map' of places to find them. Meteorites that fall in Antarctica typically become embedded in the ice sheet, making them harder to spot, but it seems many are hidden in plain sight. Two-thirds of all meteorites found on Earth have been discovered on the frozen continent, a process made easier due to the contrast between dark rocks and snow, with many discovered by chance during costly reconnaissance missions. In this new study, the team discovered that areas of 'blue ice', where frozen water is visible at the surface as ice rather than snow, could be rich in meteorites. Why this is important: Thanks to the AI-produced treasure map, the team now has a list of locations to look, with many very close to existing research stations on Antarctica. Meteorites are important for planetary scientists, as they provide a unique view into the origin and evolution of our solar system, including the Earth. [Click here to read on!]( [AI Can Identify Patients at High Risk of Heart Attack]( In brief: Scientists have developed an AI system that can analyse eye scans and identify patients at a high risk of a heart attack. Doctors have recognised that changes to the tiny blood vessels in the retina are indicators of broader vascular disease, including problems with the heart. In the research, led by the University of Leeds, DL techniques were used to train the AI system to automatically read retinal scans and identify those people who, over the following year, were likely to have a heart attack. The researchers report that the AI system had an accuracy of between 70% and 80% and could be used as a second referral mechanism for in-depth cardiovascular investigation. During the DL process, the AI system analysed the retinal scans and cardiac scans of more than 5,000 people. The AI system identified associations between pathology in the retina and changes in the patient’s heart. Why this is important: Currently, details about the size and pumping efficiency of a patient’s heart can only be determined if they have diagnostic tests such as echocardiography or magnetic resonance imaging. Those diagnostic tests can be expensive and are often only available in a hospital setting, making them inaccessible for people in countries with less well-resourced healthcare systems — or unnecessarily increasing healthcare costs and waiting times in developed countries. [Click here to discover more!]( [TinyML is Bringing DL Models to Microcontrollers]( In brief: ML provides powerful tools to researchers to identify and predict patterns and behaviours, as well as learn, optimise, and perform tasks. While these algorithms and their architectures are becoming more powerful and efficient, they typically require tremendous amounts of memory, computation, and data to train and make inferences. At the same time, researchers are working to reduce the size and complexity of the devices that these algorithms can run on, all the way down to a microcontroller unit (MCU). An MCU is a memory-limited minicomputer housed in a compact integrated circuit that lacks an operating system and runs simple commands. These relatively cheap edge devices require low power, computing, and bandwidth, and offer many opportunities to inject AI technology— a field called TinyML. Now, an MIT team working in TinyML has designed a technique to shrink the amount of memory needed even further while improving its performance. Why this is important: To increase TinyML's efficiency, Han researchers from EECS and the MIT-IBM Watson AI Lab analysed how memory is used on microcontrollers running various convolutional neural networks (CNNs). CNNs are biologically-inspired models after neurons in the brain and are often applied to evaluate and identify visual features within imagery. In their study, they discovered an imbalance in memory utilisation, causing front-loading on the computer chip and creating a bottleneck. By developing a new inference technique and neural architecture, the team alleviated the problem and reduced peak memory usage by four-to-eight times. [Click here to see the full picture!]( [Robot Performs Keyhole Surgery on Pigs]( In brief: A robot has successfully carried out keyhole surgery on the bowels of pigs mostly autonomously for the first time, which researchers say is a significant step towards human trials. Small parts of human operations are often automated, but they tend to focus on rigid parts of the body that don’t change shape, such as bone. Robots controlled entirely by human surgeons are also becoming more common. Justin Opferman at Johns Hopkins University in Maryland and his colleagues programmed a similar robot to conduct intestinal anastomosis, the joining of two ends of the intestine after a section is removed, with limited human intervention. The robot performed the surgery on four pigs, carrying out 86 stitches in total. Two-thirds of the time, the robot placed the stitch autonomously, while the rest of the time it had to be guided into place manually before attempting the stitch again. Why this is important: Custom software controlled the robot during suturing, using images from a 3D camera on the robot’s arm to sense depth and map the changing layout inside the abdomen. Opferman says the trial is the first step towards fully autonomous surgery in humans. [Click here to find out more!]( [Super Data Science podcast]( this week's [Super Data Science Podcast](, founder and CEO of Merantix Labs, Nicole Büttner, joins us for an in-depth look at how to spark and nurture A.I. innovation within a commercial organization. --------------------------------------------------------------- What is the Data Science Insider? This email is a briefing of the week's most disruptive, interesting, and useful resources curated by the SuperDataScience team for Data Scientists who want to take their careers to the next level. Want to take your data science skills to the next level? Check out the [SuperDataScience platform]( and sign up for membership today! Know someone who would benefit from getting The Data Science Insider? Send them [this link to sign up.]( # # If you wish to stop receiving our emails or change your subscription options, please [Manage Your Subscription]( SuperDataScience Pty Ltd (ABN 91 617 928 131), 15 Macleay Crescent, Pacific Paradise, QLD 4564, Australia