In this week’s roundup, we look at how AI (artificial intelligence) technologies are helping people improve their cooking, navigate traffic better, and even determine if their relationships are healthy. We also explore several different research projects, including one that can reduce the energy it takes to train machine learning models and another that can improve predictions about hurricane paths and intensity.
Can Artificial Intelligence Create a Decent Dinner?
By Nell Mackenzie, contributing writer for BBC.com
Several new AI-powered applications are helping home cooks take their cuisine to another level. From one app that helps people make the most of whatever they happen to have in the kitchen to another that can predict how good a recipe is likely to taste, AI is assisting both skilled cooks and those who rarely step into the kitchen in making better food and cocktails.
How Google Maps Is Using AI and Machine Learning to Understand Post-COVID Traffic
By ET Online for EconomicTimes.com
When COVID-19 forced a sudden shift to working and learning from home, it meant that everything we knew about traffic patterns and flow was no longer valid. Predicting travel times based on data sent from travelers’ phones is not a new Google Maps feature. However, Google is improving its predictions by implementing deep machine learning to pair data from phones with real-time information from local governments about accidents and construction backups. Users can expect more accurate time estimates and instant re-routing to reduce driving times.
Brain-Inspired Electronic System Could Make Artificial Intelligence 1,000 Times More Energy Efficient
By University College London for SciTechDaily.com
Because of its high energy use, training a current AI model can generate as much carbon dioxide as the total emissions of five cars, but that soon might not be the case. Researchers at University College London have created a much more energy-efficient method based on a technology called “memristers.” This technology was first introduced in 2008 but was found to be error-prone. Now, scientists have found a way to use memristers to create accurate neural networks that work similarly to human brains.
A Machine Learning Algorithm Uncovers The Top Five Predictors Of A Healthy Relationship
By Mark Travers, contributing writer for Forbes.com
Matchmakers, dating apps, and happy couples have all claimed to know the secrets of a successful relationship. They might be surprised by new research that says healthy relationships are based more on the subjective feelings of the people in them than objective notions of compatibility, such as having the same religion or enjoying the same type of movies. The researchers came to this finding by using machine learning to analyze more than one million data points.
NASA Researchers Use Machine Learning to Better Predict if a Hurricane Will Rapidly Intensify
By Andrea Leinfelder, contributing writer for HoustonChronicle.com
If you’ve ever seen a map of the possible paths of a developing hurricane, it’s clear that predicting the direction and magnitude of these storms is extremely difficult. Factors such as ocean temperatures and the amount of rainfall can quickly change the path and intensity of a hurricane. NASA is hoping to improve predictions with a new machine learning model that is 60–200 percent more accurate than the current model used by the National Hurricane Center.
Did you see an interesting article in the last week? Share it with us! Send it to astuttle [at] lityx.com.