Starting a synthetic biological intelligence lab from scratch

Summary: Recent advances in artificial intelligence (AI) have led to the development and deployment of gigantic models trained on billions of samples. While training these models consumes enormous energy, the human brain produces similar outputs with dramatically lower data and energy requirements....

Full description

Saved in:
Bibliographic Details
Main Authors: Md Sayed Tanveer, Dhruvik Patel, Hunter E. Schweiger, Kwaku Dad Abu-Bonsrah, Brad Watmuff, Azin Azadi, Sergey Pryshchep, Karthikeyan Narayanan, Christopher Puleo, Kannathal Natarajan, Mohammed A. Mostajo-Radji, Brett J. Kagan, Ge Wang
Format: Article
Language:English
Published: Elsevier 2025-05-01
Series:Patterns
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666389925000807
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Summary: Recent advances in artificial intelligence (AI) have led to the development and deployment of gigantic models trained on billions of samples. While training these models consumes enormous energy, the human brain produces similar outputs with dramatically lower data and energy requirements. This has increased interest in synthetic biological intelligence (SBI), which involves training in vitro neurons for goal-directed tasks. This multidisciplinary field requires knowledge of tissue engineering, biomaterials, signal processing, computer programming, neuroscience, and AI. As a result, starting SBI research is highly nontrivial and time-consuming, as most labs specialize in either the biological aspects or the computational ones. Here, we propose how a computational lab can become familiar with the biological aspects of SBI and also discuss computational aspects for biological labs that are interested in SBI. We describe general strategies as well as step-by-step processes, risks, and precautions to mitigate delays and minimize costs. The bigger picture: Synthetic biological intelligence (SBI) is an emerging field where neural cells are grown in a dish or other artificial environment to form biological neural networks. They can be trained in a feedback loop to perform tasks similar to deep artificial neural networks that are trained on big data. In recent years, large language models (LLMs) and multimodal models have achieved human-level performance in various tasks with high computational costs and energy consumption. Biological neural networks can perform tasks with minimal time, very few data samples, and only a tiny fraction of the energy required by silicon-based artificial neural networks. Interest has been growing recently in the development of biological neural networks for “NeuroAI,” a combination of neuroscience and AI, paving the way for artificial general intelligence (AGI) or even artificial super intelligence (ASI). For beginners, however, this new multidisciplinary field can be challenging to enter.
ISSN:2666-3899