Artificial Intelligence (AI) is positively impacting healthcare, business and society. Leading technologists with the IEEE Computational Intelligence Society break down ways AI research and innovation is changing our world for the better.
Why AI is important
Born more than half a century ago, Artificial Intelligence (AI) has only recently reached maturity, yet it’s already one of the most influential technologies of the modern era. Becoming increasingly ubiquitous, AI can be found in virtually every industry and government agency, in infrastructure around the world, and in our personal lives. Its power provides innovative solutions to simplify or improve our everyday life and also provides many industries with business advantages. The most popular of these technologies are Neural Networks and Machine Learning, the basis for the famous Deep Learning thrust into the spotlight of most marketed AI systems. Fuzzy techniques are other such technologies that provide flexibility and friendly interactions with users and key advantages for AI systems. Evolutionary computation and adaptation methods are promoted by the IEEE Computational Intelligence Society (IEEE CIS) to tune AI systems and provide optimized solutions to business leaders.
As the largest engineering community in the world with more than 430,000 members worldwide, the IEEE has been at the forefront of AI since its inception. The IEEE CIS is dedicated to the core technologies used in currently trending AI realizations and is an international leader in innovative interdisciplinary research, technology transfer, real-world applications, and education in artificial intelligence.
AI’s impact on COVID-19 and healthcare
During the pandemic, AI has had an opportunity to shine, with scientists from around the world tapping the technology’s power to develop innovative public health solutions. For example, with COVID-19 putting long-term healthcare facilities at risk, families are having to find alternative solutions to keep aging loved ones safe at home. Researchers at the University of Missouri’s Center for Eldercare & Rehabilitation Technology have found smart technology can help. Marjorie Skubic, a Professor of Electrical Engineering and Computer Science, and Marilyn Rantz, a Curators’ Professor Emerita in the School of Nursing, lead a large interdisciplinary team that has spent more than a decade developing and tracking smart monitoring systems designed to keep older adults at home longer. The team investigates non-invasive bed sensors that monitor pulse, respiration, and restlessness; motion sensors that track activity in the home; and depth sensors that can track gait and fall risks with the aid of AI-based techniques, and even alert designated loved ones in the event of a fall.
The team also uses AI to communicate health findings with older adults. Participants in the research group’s latest study are given devices equipped with an AI voice assistant. The assistant is synced with an individual’s health sensors and has been programmed to report sensor-based health information using visual and voice interfaces. “Aging in your home is no longer a luxury for those most at-risk of transmitted viruses,” Skubic said. “Artificial intelligence is key to addressing this challenge by giving older adults better tools for managing their own health.”
To develop more innovative solutions that help keep people safe and make the world a better place, the IEEE CIS recommends that all organizations actively combating COVID-19 leverage AI, one of the most powerful tools at our disposal. IEEE CIS has also launched a COVID-19 initiative to encourage all researchers working on any aspect of AI contributing to the fight against COVID-19 to submit their papers to special tracks of the most appropriate IEEE CIS journals.
One of the first papers published in IEEE Transactions on Artificial Intelligence (IEEE TAI), a journal dedicated to the theory and practice of AI, is a very comprehensive review of how data science research can assist in COVID-19. While a common analysis is using a classifier to detect COVID-19 from chest x-rays, data science also makes it possible to analyze sentiment regarding the virus from tweets, and to detect the rate and the use of modeling, simulation, and optimization to design drugs for the epidemic.
AI’s impact on business
Optimization problems occur in every industry, such as the allocation of beds to patients in hospitals, scheduling aircraft, and identifying the best mix to create a personalized drug for a patient to reduce the interaction with other drugs the patient is consuming. For businesses, the challenge when it comes to AI is knowing which algorithm to deploy in order to achieve the optimal result. An IEEE TAI paper presents an AI that can assist a novice user to decide on the algorithm needed to solve such business problems.
How AI benefits astronomy
As is the case with many other emerging technologies, AI’s influence on science, technology, engineering, and math (STEM) fields is massive, particularly in the areas of research and development.
An important area of learning, astronomy, is undergoing a paradigm shift due to the large volume of data generated by new sky surveys. For example, the Zwicky Transient Facility (ZTF), a survey telescope located in San Diego, California, generates an alert when a source in the sky is changing in brightness like supernovae, or in position, such as asteroids.
The ALeRCE broker collaboration pipeline is classifying the voluminous alert streams coming from the ZTF — an average of 150,000 alerts nightly — which are processed in real-time. Using an AI-based image classifier the ALeRCE team developed, in just a single alert they can identify five classes of objects: asteroids, supernovae, variable stars, active galactic nuclei, and bogus.
The ALeRCE team is also preparing for the Rubin Observatory (also known as LSST), which is expected to commence operations in 2023 with a robotic telescope. The LSST will scan the entire southern hemisphere sky every three days, collecting information from 40 billion objects, and it is expected to generate approximately 10 million alerts per night. The LSST 3.2 Gigapixel camera puts it in a unique position to discover rare and exotic events and novel unexpected phenomena.
“In preparation for the LSST era, we are developing novelty detection algorithms based on AI capable of finding rare and new phenomena,” said Professor Estevez, who is with the University of Chile and the Millennium Institute of Astrophysics.
The future of AI is education
How reliable and accurate is AI? Are we maximizing the benefits AI can generate? These questions are necessary in order for humans to trust AI, especially as smart systems become more sophisticated. Human education literature can help, researchers at the University of New South Wales Canberra and the University of Canberra believe.
Professor Hussein Abbass is leading a multi-disciplinary team funded by the Australian Research Council Discovery Program to revolutionize the design of trustworthy AI. In the same way humans are taught to perform certain tasks, a recent sub-field of AI, known as Machine Teaching, transfers knowledge from people to machines. The Platform for Interactive Concept Learning for human teachers is one of the first machine teaching platforms to teach a machine learner. Machine education structures task-specific knowledge into a curriculum for machine teaching. The outcome is an education system for graduating trustworthy AI systems certified for a group of skills. “We are designing a new type of university where students are not humans, but AIs. Our AI university is one where the graduate AI is guaranteed to have a particular set of skills acquired using an accredited curriculum,” said Professor Abbass.
