Data Science Insider: November 18th, 2022

[AI aims to make polar ship routes more fuel-efficient]( brief: The British Antarctic Survey (BAS) researchers are in the process of developing a route planning tool that optimizes a ship's journey, significantly reducing its carbon footprint. The system is designed to suggest the fastest and most fuel-efficient routes based on environmental datasets and forecast analysis. The first team to test the system in the waters will be the crew of the U.K.'s polar research ship RRS Sir David Attenborough. "We've created something very similar to the kind of in-car navigation system like Google maps...but with the added complication that in the ocean there are no roads, and the conditions are changing constantly, which affects the routes between destinations," said Professor Maria Fox, the leader of the research project. While at first, the tool's algorithms will mostly rely on the weather and environmental forecasts, the BAS AI Lab team hopes to integrate more live data, such as the speed of the ship, fuel requirements and other ship-related logistics to present a more accurate assessment of ship route optimization. Why this is important: Currently, reducing carbon footprint is critical considering the state of the earth's climate and the dire effects of climate change that take place today — and AI-based technology may be the answer. Not only does this tool can equip researchers with the necessary knowledge to find the most fuel-efficient routes, but it can also shine a light on the relationship between data and ship navigation in terms of its impact on the environment. [Click here to learn more!]( [Scientists discover how to scale up "liquid" neural networks]( news.mit.edu/2022/solving-brain-dynamics-gives-rise-flexible-machine-learning-models-1115In brief: The advanced machine learning algorithms based on the brains of small species (also known as "liquid" neural networks) were a groundbreaking discovery last year. However, the MIT research team took it further and developed a new type of AI "liquid" neural algorithm by solving the equation of how two neurons interact with each other through synapses. These new models are known as "closed-form continuous time" (CfC) neural networks. Before these findings, the main barriers to a better neural algorithm model were the complex maths behind them and expensive computing due to the number of neurons and synapses increasing with every next step in the algorithm. Solving the differential equation that underpins the interactions between two neurons is bound to generate faster and more flexible liquid network algorithms. "CfC models are causal, compact, explainable, and efficient to train and predict. They open the way to trustworthy machine learning for safety-critical applications,” said MIT Professor Daniela Rus. Why this is important: In practical terms, the improved models of liquid networks that translate into faster and more accurate machine learning algorithms also enhance the performance of various tasks as a result. For instance, for a medical prediction task which involved sampling patients, the new CfC neural networks worked 220 times faster than the previous model. Other tasks where the new algorithms outperformed the traditional models included detecting human activity from sensors, image recognition and processing, and modelling the physical dynamics of a robot model's movements. This discovery is vital for data scientists and researchers as it opens new possibilities for grand-level simulations which had never been done before. [Click here to read on!]( [Nvidia and Microsoft partner to build an AI Supercomputer]( In brief: The two giants well-known in the world of tech and business have recently announced an exciting mutual collaboration, which involves them working together to create an AI Supercomputer. It will be hosted on Microsoft's Azure cloud, utilizing Nvidia's A100 and H100 chips that contain thousands of graphics processing units, which is what makes them the most potent company's processors. Nvidia also aims to integrate their Quantum-2 InfiniBand networking software into the cloud, which would make it to be the first public cloud to use that technology. The main objective of the project is to harness the full potential of the Microsoft cloud's capabilities to improve research and automation processes for organisations in the future. As Ian Buck, Nvidia's GM for hyperscale and HPC, said in his interview with Reuters: “We're at that inflection point where AI is coming to the enterprise and getting those services out there that customers can use to deploy AI for business use cases is becoming real." Why this is important: This is a crucial development for businesses, smaller organisations and larger corporations alike, as it allows them to offer Microsoft's AI cloud and application services to customers shortly. As the AI-powered computing infrastructure is bound to become the new reality in the commerce world, it is going to transform everyday processes and functions within enterprise structures, making them faster, more efficient and fully automated. Simultaneously, it is a wake-up call for organisations with more traditional protocols, systems, and hierarchies to embrace the change. [Click here to discover more!]( [AI tool to generate personalised hypertension treatment]( In brief: One of the principal disadvantages of hypertension treatment practices is that there is no consensus within the medical community regarding the level at which antihypertensive medications should be administered. The Yale-led team aims to bring more clarity to the issue with their new machine learning tool called PRECISION, which stands short for Pressure Control in Hypertension. Using the data from two clinical trials that evaluated the cardiovascular benefits in different patients who had taken medication to treat the pressure, the team computed the results, creating a graph that compares outcomes in patients that were assigned a higher dose with those that received the standard treatment. “We hypothesized that developing machine learning-based methods can infer response to therapies from diverse populations enrolled in studies," said Rohan Khera, Yale assistant professor of Cardiology and director of the Cardiovascular Data Science Lab. As the tool can analyze the data of individual patients, it can help determine the effects of treatment on a more personalised level rather than relying on population-level results alone. Why this is important: Implementing the PRECISION software benefits both the patients and medical professionals. With the more personalized approach to the data from the previous trials, future patients will be able to receive a treatment that is best suited for them. Not to mention, the more efficient the methods of detection, the less time and costs it takes for clinical studies to run successfully, which speeds up the progress of discovering and improving new treatments and making them accessible to the public. [Click here to see the full picture!]( [AI will play a major role in World Cup]( In brief: As everyone is gearing up for one of the most exciting sporting events of the year, data scientists will also be on their best game this season, ready to assist with their knack for data analysis and processing. Not only will the players have access to a detailed data breakdown of their performance, including how many runs they made and the number of hassles made by the opposing team, but it will also enable them to formulate a clearer strategy for their future games. In particular, these sets of data will be used to analyse the tactics and behaviour of their own teams as well as those of the opponents'. Matthew Penn, a PhD student at Oxford who had developed a statistical 'double Poisson model' that evaluates and interprets the attacking and defending strengths of the teams, predicts that Belgium is the most likely to win this year's World Cup, followed by Brazil. Why this is important: Data science studies that focus on the progress and outcomes of sports competitions and tournaments convey invaluable information to coaches, players and clubs. Using big data, they can spot past mistakes to improve their strategy and leverage their performance. It also enables coaches and soccer clubs to use this data to scout new players based on how well they did in any particular game or season. Arguably, it can also increase the fairness of matches as it will be easier to determine any violations with the software's aid rather than relying on human bias and error that are inevitable. [Click here to find out more!]( [Super Data Science podcast]( this week's [Super Data Science Podcast](, Erin LeDell, H2O.ai’s Chief Machine Learning Scientist, joins Jon to investigate how AutoML supercharges the data science process, the importance of admissible machine learning for an equitable data-driven future, and what Erin’s group Women in Machine Learning & Data Science is doing to increase inclusivity and representation in the field. --------------------------------------------------------------- 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