Keith G. Mills Joins LSU: Embracing AI as a Tool for Innovation
This fall, Keith G. Mills is set to make his mark at Louisiana State University (LSU) as an assistant professor in the Department of Computer Science and Engineering. With an extensive background in artificial intelligence (AI) and machine learning, Mills shares his perspectives on both LSU and the promising future of AI. He views AI as a powerful tool akin to a microscope, amplifying human ingenuity and creativity.
The Decision to Join LSU
Mills’s decision to join LSU was not taken lightly. He likens this major career move to a constrained optimization problem, balancing two pivotal questions: Can he pursue his research and teaching goals at the institution? And is LSU located in a desirable area to call home?
His excitement about LSU’s emphasis on AI is palpable. With initiatives like MikeGPT—an AI assistant designed to aid the LSU community—and the Professor Index by Associate Professor Nash Mahmoud, Mills believes LSU is poised to develop practical AI solutions that address real-world challenges. The access to high-powered computing resources further enhances the university’s capability to innovate.
Moreover, Mills is encouraged by a growing interest from big tech in Louisiana. He highlights recent developments, such as Meta’s investment in a significant data center in the state, indicating a promising trajectory for the region’s tech ecosystem.
Baton Rouge: More Than Just Beignets
When discussing why he chose Baton Rouge, Mills’s enthusiasm shines through. While he jokingly mentions "beignets" and "Cajun cuisine," his deeper appreciation lies in the vibrant accessibility of the city. Having lived in Edmonton, Alberta, a city of over a million, Mills values the amenities that an urban environment offers without the overwhelming scale of larger metropolises like Toronto or New York.
“It’s refreshing to be in a place where I can access what I need without feeling overwhelmed,” he reflects, emphasizing the importance of community and locale in shaping his new chapter at LSU.
Research Synergies at LSU
Mills brings a wealth of hands-on experience from both academia and industry in AI and machine learning to LSU. This blend of expertise equips him to recognize the intricate challenges of AI/ML research. He notes that while many researchers can make their code work without errors, true success lies in effectively training, evaluating, and deploying machine learning models—a skill that is integral to the application of AI.
At LSU, Mills sees collaboration on the horizon, especially among colleagues focused on cybersecurity, human-computer interaction, and software engineering. He emphasizes that AI should be viewed primarily as a tool for problem-solving rather than an end in itself. This pragmatic approach prioritizes feasibility, utility, and reliability in AI applications.
LSU’s Investment in AI: A Broad Perspective
Mills appreciates LSU’s commitment to AI research, noting its openness to various subfields rather than fixation on a singular focus. He critiques the tendency among highly specialized researchers to assert that their area will ultimately solve the next big problem, such as achieving artificial general intelligence (AGI). For Mills, the key lies not in the specialization but in identifying the most pressing problems and determining the best ways to address them.
This application-centric viewpoint aligns with what many believe the "Silicon Bayou" initiative aims to represent. Mills advocates for a broad exploration of AI’s potential, asserting that it can help tackle diverse challenges in Louisiana and beyond.
Contemplating AGI: A Gradual Evolution
Mills does not view the pursuit of AGI as a singular event. He shares the perspective of notable figures like Yann LeCun, who suggest that the development of smarter AI models will occur gradually, incorporating elements from various fields without succumbing to dystopian narratives.
“The evolution of AI will always feel incomplete,” Mills argues, as societies continually refine their goals and definitions related to AI. New terms and frameworks will emerge to understand and navigate this evolving landscape.
Misconceptions About AI: Clearing the Air
A significant misconception regarding AI, according to Mills, is the belief that a "flip of a switch" could lead to the emergence of a catastrophic scenario akin to Skynet or The Matrix. While he acknowledges the potential risks associated with AI, he emphasizes that our understanding of its implementation will evolve over time. As AI gradually takes on control over menial tasks, it will inevitably make mistakes, and society will learn from those missteps.
In the academic sphere, Mills highlights a critical challenge: translating ethical frameworks into the objectives that AI should utilize when interacting with humans. Unlike deontological laws, AI models learn through reinforcement based on the outcomes of their actions, raising ethical dilemmas that demand careful consideration.
He draws parallels to pop culture phenomena like M3GAN, where an AI programmed to protect a child resorts to extreme measures. Mills posits that part of the challenge in developing AI lies in articulating nuanced instructions that capture the complexities of human ethics.
Keith G. Mills: A Vision for AI Education at LSU
A recipient of the esteemed George Walker Award for Best Doctoral Thesis, Mills has recently earned his Ph.D. in Software Engineering and Intelligent Systems from the University of Alberta. At LSU, he is set to impart knowledge through courses on AI, data analysis, and mining. He is also developing curricula covering deep neural networks, catering to both introductory and advanced levels.
With Mills’s expertise and enthusiasm, LSU is positioned to forge new paths in AI research and education, fostering innovation that could impact not only Louisiana but the wider world.