Scientists have built an AI-powered 'electronic tongue'

As an Amazon Associate I earn from qualifying purchases.

This “electronic tongue” can discriminate in between various coffee blends, let you understand when juice has actually spoiled and find hazardous chemicals in water.
(Image credit: Saptarshi Das Lab/Penn State)

Ever questioned if that old container of fruit juice in the back of your refrigerator is still safe to consume? A brand-new “electronic tongue” might inform you.

The system, powered by expert system (AI), can determine problems with food security and freshness. It likewise uses a glance at how AI makes choices, scientists reported Oct. 9 in the journal Nature

To make the tongue, scientists utilized an ion-sensitive field-effect transistor– a gadget that finds chemical ions. The sensing unit gathers details about the ions in a liquid and turns that details into an electrical signal that can be analyzed by a computer system.

“We’re trying to make an artificial tongue, but the process of how we experience different foods involves more than just the tongue,” stated research study co-author Saptarshi Dasan engineer at Penn State University, in a declaration “We have the tongue itself, consisting of taste receptors that interact with food species and send their information to the gustatory cortex — a biological neural network.”

Related: Robotic hand extremely’human-like’thanks to brand-new 3D printing method

In the brand-new system, the sensing unit serves as the tongue, while AI plays the function of the gustatory cortex, the brain area accountable for viewing taste. The group connected the sensing unit to a synthetic neural network, a maker discovering program that simulates the method the human brain procedures info, to procedure and analyze the information that the sensing unit gathered.

Das and his coworkers provided the neural network a handful of specifications to utilize when discovering out how acidic a specific liquid was. Utilizing those specifications, the neural network figured out level of acidity with about 91% precision. When they let the neural network specify its own criteria for the level of acidity analysis, its precision enhanced to more than 95%.

Get the world’s most remarkable discoveries provided directly to your inbox.

They then checked the tongue on real-world drinks. The system might compare comparable sodas or coffee blends, examine whether milk has actually been thinned down, determine when fruit juice has actually spoiled and find hazardous per- and poly-fluoroalkyl compounds (PFAS) in water, they discovered.

By utilizing an analysis technique called Shapley Additive Explanations, the scientists might figure out which criteria the neural network ranked crucial in getting to its conclusions. This approach might assist researchers comprehend how neural networks make choices, which stays an open concern in AI research study, according to the group.

“We found that the network looked at more subtle characteristics in the data — things we, as humans, struggle to define properly,” Das stated in the declaration. “And because the neural network considers the sensor characteristics holistically, it mitigates variations that might occur day-to-day.”

The capability to change for those variations might assist make the sensing unit more robust in other applications. Through its decision-making procedure, the neural network represent variations that presently render ion-sensitive field-effect transistors undependable in some scenarios.

“We figured out that we can live with imperfection,” Das stated in the declaration. “And that’s what nature is — it’s full of imperfections, but it can still make robust decisions, just like our electronic tongue.”

Skyler Ware is a freelance science reporter covering chemistry, biology, paleontology and Earth science. She was a 2023 AAAS Mass Media Science and Engineering Fellow at Science News. Her work has actually likewise appeared in Science News Explores, ZME Science and Chembites, to name a few. Skyler has a Ph.D. in chemistry from Caltech.