In 2024, the Nobel Prize in physics was awarded to John Hopfield and Geoffrey Hinton for their foundational work in artificial intelligence (AI), and the Nobel Prize in chemistry went to David Baker, Demis Hassabis, and John Jumper for using AI to solve the protein-folding problem, a 50-year grand challenge problem in science.
A new article, written by researchers at Carnegie Mellon University and Calculation Consulting, examines the convergence of physics, chemistry, and AI, highlighted by recent Nobel Prizes. It traces the historical development of neural networks, emphasizing the role of interdisciplinary research in advancing AI. The authors advocate for nurturing AI-enabled polymaths to bridge the gap between theoretical advancements and practical applications, driving progress toward artificial general intelligence. The article is published in Patterns.
"With AI being recognized in connections to both physics and chemistry, practitioners of machine learning may wonder how these sciences relate to AI and how these awards might influence their work," explained Ganesh Mani, Professor of Innovation Practice and Director of Collaborative AI at Carnegie Mellon's Tepper School of Business, who coauthored the article. "As we move forward, it is crucial to recognize the convergence of different approaches in shaping modern AI systems based on generative AI."
In their article, the authors explore the historical development of neural networks. By examining the history of AI development, they contend, we can understand more thoroughly the connections among computer science, theoretical chemistry, theoretical physics, and applied mathematics. The historical perspective illuminates how foundational discoveries and inventions across these disciplines have enabled modern machine learning with artificial neural networks.
Then they turn to key breakthroughs and challenges in this field, starting with Hopfield's work, and go on to explain how engineering has at times preceded scientific understanding, as is the case with the work of Jumper and Hassabis.
The authors conclude with a call to action, suggesting that the rapid progress of AI across diverse sectors presents both unprecedented opportunities and significant challenges. To bridge the gap between hype and tangible development, they say, a new generation of interdisciplinary thinkers must be cultivated.
These "modern-day Leonardo da Vincis," as the authors call them, will be crucial in developing practical learning theories that can be applied immediately by engineers, propelling the field toward the ambitious goal of artificial general intelligence.
This calls for a paradigm shift in how scientific inquiry and problem solving are approached, say the authors, one that embraces holistic, cross-disciplinary collaboration and learns from nature to understand nature. By breaking down silos between fields and fostering a culture of intellectual curiosity that spans multiple domains, innovative solutions can be identified to complex global challenges like climate change. Through this synthesis of diverse knowledge and perspectives, catalyzed by AI, meaningful progress can be made and the field can realize the full potential of technological aspirations.
"This interdisciplinary approach is not just beneficial but essential for addressing the many complex challenges that lie ahead," suggests Charles Martin, Principal Consultant at Calculation Consulting, who coauthored the article. "We need to harness the momentum of current advancements while remaining grounded in practical realities."
The authors acknowledge the contributions of Scott E. Fahlman, Professor Emeritus in Carnegie Mellon's School of Computer Science.
