Efficient and scalable bio-integrated AI system using brain organoids

This technology is a computing system that integrates brain organoids with CMOS interfaces to enhance artificial intelligence (AI) scalability and efficiency.

Unmet Need: Scalable, energy-efficient AI that mimics human brain processes

The current standard in AI systems, based on artificial neural networks, faces significant shortcomings due to its heavy memory usage and energy consumption, making scalability a challenge. These systems also do not closely mimic the complex computational processes of the human brain. Addressing these limitations is critical as it allows for the development of more sustainable and efficient AI models that can handle the increasing volume of global data, while potentially offering a computational approach that more accurately reflects human cognitive processes.

The Technology: Bio-integrated AI with enhanced efficiency and scalable network capabilities

This technology integrates biological components into computing systems by employing brain organoids derived from induced pluripotent stem cells, coupled with high-resolution electrophysiological CMOS interfaces. These organoids serve as dynamic, high-dimensional reservoirs for information processing, utilizing principles of reservoir computing to bypass extensive training required by traditional artificial neural networks. The approach reduces operational voltage and enhances energy efficiency while enabling more complex and adaptable interconnection networks. This hybrid model offers significant improvements in scalability and efficiency for AI systems and data centers, compared to existing solid-state systems.

Applications:

Autonomous system processing
Adaptive AI learning systems
Medical diagnostics and disease modeling
Decision-making process in robotics
Cognitive science simulation tools
Neural network architectures
Research tool for neuroscience and brain mapping

Advantages:

Enhanced scalability
Cost and energy effective
Accurate mimicry of human brain processes
Faster processing speeds with minimal training
Computing performed both biologically and in silico

Lead Inventor:

Kenneth Shepard, Ph.D.

Patent Information:

Patent Pending

Related Publications:

Zeng N, Jung T, Sharma M, Eichler G, Fabbri J, Cotton RJ, Spinazzi E, Youngerman B, Carloni L, Shepard KL. “A Wireless, Mechanically Flexible, 25μm-Thick, 65,536-Channel Subdural Surface Recording and Stimulating Microelectrode Array with Integrated Antennas” In2023 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits) 2023 Jun 11 (pp. 1-2). IEEE.

Tech Ventures Reference:

IR CU24020
Licensing Contact: Greg Maskel