Artificial intelligence is making headlines at a rapid pace, and at the same time the science of AI is advancing. In just a few years, its influence has surpassed trend status and become firmly entrenched in technology and our daily lives.
The University of Alabama is taking steps to ensure UA and its students are at the forefront of the AI wave. Artificial intelligence is no longer limited to the world of science fiction, but even if we use it every day, the term can mean more or less than many people realize. may be.
Build AI capabilities
Allen Parrish, executive director of the Alabama Cyber Institute at the University of Alabama, says what’s not AI is cause for concern.
Dr. Allen Parrish (contributor)
“Many universities have taken a defensive stance on this,” Parrish said. “But instead, we need to think about how AI will change our society and workforce and prepare our students to educate themselves in ways that will help them seize that opportunity.”
The university has begun moving forward with a project that Parrish sees as the gateway to a new era of computing at UA: a high-performance computing and data center project to be announced in 2023.
The $100 million facility will significantly expand the university’s computing capabilities. For scale, HPC has 125 racks to house servers. The university’s current capacity would fit into one of these racks.
“In some ways, we will be expanding our ability to conduct high-level research by a factor of 100,” Parrish said. “We will be able to ingest larger amounts of data and process it in-house.” This opens up new possibilities, for example, for projects that require a higher level of security. “At the same time, we can focus on new ways to teach students how to use AI, which will be an important part of the future job market.”
Researchers across campus are harnessing the power of data for a variety of tasks, including flood prediction, designing chemical catalysts, and better understanding when and how structures fail. HPC enables more of these projects and increases the speed and power available to drive them. HPC can also support art and visual learning, such as the work of art historian Dr. Jennifer Feltman, who helped create a virtual reality experience of Notre Dame Cathedral in Paris for middle school students.
“This project is campus-wide,” Parrish said. “What we’re seeing in AI today is a microcosm of what we’ll see in the next five years, and we want to be ready for it.”
The University of Alabama’s High Performance Computing and Data Center will house petascale computing systems. This means it can perform trillions of calculations per second. (contribution)
Defining next generation AI
The term artificial intelligence itself contains many ideas. Some of them, such as self-aware robots bent on world domination, exist firmly in the realm of fiction. Dr. Sergei Glaser is an associate professor in the Department of Physics and Astronomy who uses AI to investigate the nature of the universe. He is one of the people around the world who is pushing the boundaries of the field itself.
He said the more accurate term he means when talking about AI is “machine learning.” It’s just the science of algorithms that can learn from data and use analysis of that data to make predictions about new data. A simple example is separating cats and dogs. The more times the model completes this task and corrects errors when they occur, the more successful it will be at identifying cats and dogs in the future.
Dr. Sergey Glaser (contributor)
Around 2010, the “deep learning revolution” occurred. A class of algorithms based on the workings of our brains, called neural networks, can be layered to achieve more sophisticated decision-making processes. Technological innovations have made it possible to build deeper neural networks with more layers.
This is a development that has advanced machine learning in areas such as image processing and natural language processing, and laid the foundation for generative AI models such as ChatGPT and Midjourney.
Gleyzer is active in developing deep learning algorithms for use in physics and astronomy, including research at the Large Hadron Collider in Cologne, Switzerland. The next frontier, he says, is quantum machine learning.
Quantum computing is still in its infancy, but it’s learning how to crawl. In classical computing, information is encoded in 1s and 0s, or binary code. Quantum computing makes it possible to encode data in superpositions, or weighted combinations of 0s and 1s in which both 1s and 0s are potentially present, Gleyzer said. That is the quantum state.
“This is not a theoretical thing. Quantum computers already exist,” Glaser said. And he’s writing algorithms for them.
Quantum computing is the equivalent of computing in the 1950s. This technology is bulky and expensive, and few people have access to machines that are already built. “We asked, if we had a quantum computer, could we design better machine learning algorithms for that hardware? And that’s what we’re doing,” Glaser said. Quantum machine learning has been around for less than a decade, but Gleyzer’s group has already developed quantum versions of machine learning algorithms that perform as well as or better than their classical computing equivalents. is designing.
Glaser believes that both classical and quantum machine learning will be used in the future, just as radio and television are still used alongside the internet. But as the nature of computing evolves, he and his students at the University of Alabama are at the forefront.
A quantum computer that visually represents the qbit, which is the basic unit of data in quantum computing. (contribution)
Alabama Generated AI Task Force
UA is sitting with Alabama leaders as they consider both the possibilities and caveats of widespread use of AI tools.
Dr. Matthew Hudnall, assistant professor of management information systems, has 20 years of experience with data systems in Alabama. When a car is registered with the Land Transportation Office, the paperwork is processed by software he helped create. Earlier this year, Hudnall joined Governor Kay Ivey’s Task Force on Generative AI.
“Our mission is to develop a set of recommendations and plans that leverage generative AI in the best and safest way possible for the betterment of Alabama,” Hudnall said. Four working groups within the task force meet regularly to develop recommendations that will be included in the final report.
Generative AI is just one part of the broader field of machine learning, but it is widely available and at the forefront of popular culture. A unique feature of generative AI is that once the model is trained, it does not make predictions about unseen content. They create new content.
Dr. Matthew Hudnall (contributor)
Regulations such as the Health Insurance Portability and Accountability Act (HIPAA) and the Family Educational Rights and Privacy Act (FERPA) protect our sensitive data, but there are various levels of sensitivity related to AI. guidance is urgently needed. An employee may be able to help. “For example, we don’t want to send tax data for Alabama residents to ChatGPT because there are data security concerns,” Hudnall said. The final report includes a data matrix devised by one of his work groups, based on a system used by the UA’s Office of Information Technology, which helps advise state agencies on the level of data sensitivity. Included.
Each of the four working groups provides similar recommendations for different aspects of generative AI.
As part of his work on the task force, Hudnall guides a group of UA graduate students in the creation of a large-scale language model designed to answer questions about the Alabama Code. This project provides students with a front-row seat to data policy creation and hands-on learning opportunities.
Students will use large language models created by other developers and available for free online. Students downloaded the model and trained it on the Alabama Code. Once trained, anyone can use this software to answer even the most obscure Alabama law questions in seconds. It’s also hosted and executed on local hardware, so no training data is fed back to any external party.
This student project will be included in the report as an example of how to safely harness the power of generative AI.
“I can only see further integration into our society, from process automation to faster access to data,” Hudnall said. “It is our duty as institutions to ensure that our students have these tools.”
This story originally appeared on the University of Alabama website.
