Molecular AI
DNA Computing Devices and Intelligent Nanobio Agents that Learn from Wet Molecular Data
Overview
This project investigates intelligent nanobiomolecular agents that perceive, act, and learn to make decisions in biomolecular environments. The aim is to develop DNA-based molecular information processing technologies and devices that learn from wet molecular data in both in vitro and in vivo biological systems.
Current progress spans specific milestones such as experimentally implementing molecular evolutionary learning of DNA hypernetworks for handwritten digit classification, and massively parallel distributed pattern recognition using simulations in molecular computing.
There are many applications in the BT and NT industries for the DNA computing devices and molecular intelligent agents developed under this project. Areas of particular interest include molecular diagnostics, personalized drug delivery, intelligent gene therapy, and nanorobot controllers.
Objectives
- Molecular machine-learning devices using in vitro DNA computing
- Molecular pattern recognizers that automatically learn from wet DNA data
- Intelligent nanobio agents that learn to perceive and act in molecular environments
- Robust decision making in changing biomolecular environments over a long period of time
System Architecture
The system architecture (illustrated in the project diagram nano_bio.png) depicts the integration of molecular computing substrates with learning algorithms, bridging the gap between biological wet-lab data and intelligent decision-making at the nanoscale.
Research Team
Principal Investigator: Prof. Byoung-Tak Zhang
Researchers:
- Ji-Hoon Lee
- Christina Baek
- Hyo-Sun Chun
- Je-Hwan Ryu
Contact: Christina Baek (dsbaek@bi.snu.ac.kr)