Topic: Technology

Researchers at a School of Engineering have created a new AI system that uses much less energy and is more accurate. This breakthrough could help reduce energy consumption in data centers and AI systems.

Artificial intelligence (AI) has been consuming a lot of electricity in the United States. In fact, AI systems and data centers used about 415 terawatt hours of power in 2024, which is more than 10% of the country's total electricity production. This rapid growth has raised concerns about sustainability. To address this issue, researchers at a School of Engineering have created a new AI system that uses much less energy while being more accurate.

The new AI system is called neuro-symbolic AI and combines traditional neural networks with symbolic reasoning. This approach mirrors how people approach problems by breaking them into steps and categories. The team, led by Matthias Scheutz, Karol Family Applied Technology Professor, has developed this method to reduce energy use by up to 100 times while also improving performance on tasks.

The researchers tested their system using the Tower of Hanoi puzzle, a classic problem that requires careful planning. The neuro-symbolic AI achieved a 95% success rate, compared with just 34% for standard systems. When given a more complex version of the puzzle that it had not encountered before, the hybrid system still succeeded 78% of the time. Traditional models failed every attempt.

The new system learned the task in only 34 minutes, while conventional models required more than a day and a half. Energy consumption was also reduced dramatically. Training the neuro-symbolic model required only 1% of the energy used by a standard AI system. During operation, it used just 5% of the energy needed by conventional approaches.

The team's work will be presented at the International Conference of Robotics and Automation in Vienna in May and will appear in the conference proceedings.

Why It Matters

This breakthrough could help reduce energy consumption in data centers and AI systems, which is important for India's growing tech industry. It also shows that innovative solutions can be found by combining different approaches to a problem.

Key Facts

• The new AI system uses much less energy than traditional AI systems, reducing energy consumption by up to 100 times.
• The neuro-symbolic AI combines traditional neural networks with symbolic reasoning to improve performance on tasks.
• The researchers tested their system using the Tower of Hanoi puzzle and achieved a 95% success rate.

Key Terms

Neural Networks

Computer systems inspired by the human brain that can learn from data.

Implications

This breakthrough could help reduce energy consumption in data centers and AI systems, which is important for India's growing tech industry. It also shows that innovative solutions can be found by combining different approaches to a problem.