AI inspires the working of IT systems undertaking even the most challenging tasks. Advances in the field of robotics and AI often appear to be slow to the majority, but that is because the tasks performed by these technologies have become so completely ingrained into our lives that most don’t even notice them performing.
Moreover, the deep learning revolution that allows neural networks to perform algorithms has led to innovations that have taken AI one step closer to becoming power tools with amazing capabilities that everyone had hoped for from as far back as the 70s.
News from the U.S. and around the world this past week continue to prove that breakthroughs are going to keep coming in from the fast-paced world of technology and they are becoming even more noticeable than ever before.
In the week that passed the Coronavirus epidemic has escalated. Amid fears of its outcome and the race for a vaccine we have seen how important AI is in helping to tackle any dilemma, whether it is helping choose what people will watch on Netflix, finding medication for illnesses or even improving transportation systems. George McCormick, a leader in agency-side social media communication, said: “it is astonishing to see the psychographic data and user-generated participation on the Facebook and Instagram channels of both governments and health organizations. The willingness of governments to invest in social media as an alternative way to reach the masses and gather information from them is quite notable.”
Fighting COVID-19
With an expansive surveillance system that uses facial recognition working toward curtailing the spread of the coronavirus in China, AI and big data are the surest way to fight it fast.
Artificial Intelligence is helping scientists from SRI Biosciences in California to speed up the research for the anti-viral drug for COVID-19 to only 6 months when the normal time needed would have been 2 years. Robotics are used to determine what molecules are the best for fighting the virus and how the compound needs to be synthesized and tested.
This week SRI announced their collaboration with Iktos, the AI technology firm that has accelerated other drug discovery programs. Avi Ben Ezra, the CTO of SnatchBot, who is a key backer of affordable government AI and RPA systems globally, said: “Now, with RPA and AI teamed up, response centers can be much quicker, more automated – and provide the kind of realtime information that will make a difference.”
Real-time assistance for neurobiological movement disorders
People with Parkinson’s and other neurodegenerative diseases experience severe hand tremors. Scientists from Canada and the US have used a pioneering machine learning model that used small sensors to analyze patient motions from the 60s and 70s. They then applied a data-driven neural network modeling technique in a novel way to extract information about predictive hand motions that can be applied to all patients.
Their model is ready to be used and promises to precisely predict involuntary movements for patients in real-time. The current model requires substantial computational power and the scientists plan to develop new models of exoskeletons that use low-power cloud-computing so that they can be operated in the homes of patients.
Advanced computer vision
Computer vision or image recognition has become integral for many AI applications from driverless cars, security, and industrial applications. However, the computer power needed to process the data received usually slows the sensors that give the required visual data.
From the MIT Technology Review, the news is that scientists have announced the arrival of a new type of artificial eye that can process the data of what it’s seeing in nanoseconds, faster than any image sensors until now. The new sensor uses a chip made of tungsten diselenide that is etched with light-sensing diodes. The network can be trained by adjusting the sensitivity of the diodes until the correct responses are received.
Even though the eye still has limited abilities because it only consists of 27 detectors, it can perform both standard supervised and unsupervised machine learning tasks and the scientists believe that scaling the neural network to larger sizes will be easy.
Autonomous transport news
After its 2018 setback following the 2018 accident in Tempe, Uber published a post that gave readers a glimpse into how complex it is to develop and deploy self-driving cars. However, self-driving cars are not the only autonomous mode of transport making it to the news this week.
IBM and Promare, a U.K. based marine research and exploration charity, announced that they are running a trial on an AI-powered marine navigation system prototype. The Mayflower Autonomous Ship (MAS) is due to cross the Atlantic Ocean in September this year following the same routes as the original Mayflower in 1620.
The crewless MAS will travel at 20 knots an hour as opposed to the 2.5 knots of the original Mayflower and will be propelled by the generation of solar-and wind-powered and will have a diesel generator in case it needs backup. According to the developers, edge computing is the technology that will allow the MAS to navigate the waters of the Atlantic as its onboard “AI Captain” senses the environment and acts on information by making smart decisions quickly.
