This week’s list of top data news highlights covers June 1, 2024 to June 7, 2024 and includes articles on a robot-run Starbucks in South Korea and using an algorithm to generate customized treatment plans for Parkinson’s patients.
London-based startup Stability AI has released an open-source AI model that can generate sounds and songs up to 47 seconds in length from text prompts. Stability AI trained the model on 486,000 royalty-free samples from several digital music libraries.
The local government in Borders, Scotland has launched a digital portal to promote waste recycling in the region. Residents register their address with the tool and the platform guides users on how and when they should recycle different items, sends collection reminders, and provides updates regarding any changes or disruptions to collection services.
3. Accelerating Antibiotic Discovery
Scientists at the Queensland University of Technology in Australia have used a machine learning algorithm to identify approximately 860,000 antimicrobial peptides that could help defend against pathogens that make antibiotics ineffective, which ultimately kill roughly 1.3 million people globally each year. Researchers have published an open dataset containing the newly discovered peptides to speed up discovery of new antibiotics.
4. Delivering Starbucks Orders
Autonomous robots now deliver coffee orders to more than 5,000 employees working in the headquarters of Naver, a South Korean tech company. The robots, called Rookie, move freely between 28 floors using a robot-only elevator and coordinate movements with each other to avoid collisions.
Amazon has rolled out a program called Private Investigator that uses computer vision to inspect customer orders for damages and verify that the company is shipping the correct product. Amazon uses this program to reduce returns and increase customer satisfaction.
Researchers from the University of Nebraska–Lincoln have developed a technique to better identify gene functions in corn using RNA instead of DNA. They measured RNA levels of over 39,000 genes in approximately 700 corn varieties, combining this data with physical measurements of the plants. Researchers designed special equipment using 3D printers to collect samples quickly and preserve RNA. The researchers also created a software pipeline to analyze the vast dataset, identifying about 10 times as many genes affecting flowering time compared to traditional methods
7. Training Medical Professionals
Researchers at an Israel-based startup called OtheReality have developed virtual reality (VR) simulations that allow medical professionals to experience complex patient scenarios like pain management or end-of-life care. The new training program aims to instill empathy in future doctors to reduce medical errors and improve healthcare quality.
8. Curating Data Automatically
An international team of AI researchers has created a way to automatically sort and improve data for training AI. Their method uses three steps to make sure the data is diverse and balanced, which involves grouping similar data points together. Tests showed their automatic method works as well as, or even better than, manually sorting data, and was particularly effective for curating data to train vision models.
9. Enhancing Parkinson’s Treatment
Researchers at Mass General Brigham, a Boston-based medical research nonprofit, have created a personalized approach to deep brain stimulation (DBS), a treatment for Parkinson’s disease, by mapping specific symptoms to different brain areas. They developed an algorithm called Cleartune, which suggests customized DBS settings for each patient, and tested it on five patients, showing better symptom improvement in most cases compared to standard methods. This new approach aims to make DBS more effective for Parkinson’s disease and potentially other conditions.
10. Improving Access to Public Restrooms
New York City has launched a new Google Maps layer showing all city-owned public restrooms across the five boroughs. The initiative, called “Ur In Luck,” will be available on the city’s open data portal for developers, who can integrate data about public restrooms into various digital applications, thereby improving public access to restrooms and addressing the city’s restroom scarcity.
