By Charles Carter, 24/09/22

Biomedical engineering innovators from Duke University have developed a new data encryption method using simulated bacterial growth patterns.

The approach involves machine-learning and could have security advantages over existing encoding methods, due to the chaotic nature of bacterial growth.

How does it work?

The team developed a simulation of how bacteria grow in a petri dish based on different initial conditions such as nutrient levels and space constraints.

They allocated each letter of the alphabet and other text-based characters to a specific set of initial conditions and let their virtual bacterial colony reproduce.

The pattern produced by a set of initial conditions is slightly different every time, but the engineers created a machine learning algorithm that could decipher the corresponding character correctly.

The receiver is sent a video of the different patterns that make up the letters and words of the secret message, and the algorithm spits out the decoded sentence.

What are the potential benefits?

The chaotic nature of self-organised bacterial growth means the pattern is different every time even for the same initial conditions.

This means a malicious actor trying to intercept and decode the message would find it extremely difficulty to develop their own AI to solve the puzzle.

The method is also highly scalable and generalisable for different applications including secure communication of more complex information.

Questions for you. Comment below

  1. First thought that comes into your head?
  2. Pros and cons according to you?
  3. Other applications of this approach?
  4. What could this be combined with?

Links

https://pratt.duke.edu/about/news/bacterial-pattern-encoder

https://www.patternencoder.com/

Leave a Reply

Your email address will not be published. Required fields are marked *