The Alexander Kofkin Faculty of Engineering
Even Better Than the Real Thing
A novel, AI-powered alternative to standard spectroscopy can put the power to protect our health in our hands—literally.
Sensing the Light
Wondering if that olive oil is really extra virgin? Or that honey real or fake? Soon enough, you may not need to: Thanks to a microscale device developed by Prof. Doron Naveh of the Alexander Kofkin Faculty of Engineering, you’ll be able to confirm the physical properties of food items via your smartphone, right there in the supermarket aisle. Like so many of today’s breakthrough innovations, Naveh’s technology was made possible by AI, this time when applied to the problem of bulky, costly instruments for measuring the properties of light.
Called computational spectroscopy, Naveh and his international team of researchers use algorithms to perform a function that currently requires the use of a prism to refract light. By measuring a substance’s absorption and emittance of the refracted light’s wavelengths, we can determine that substance’s chemical concentrations, whether they’re the proteins and sugars in our food or the molecules in our samples of blood. But now, as Naveh demonstrates with his prism-free, intelligent sensing device, we can also collect data points on a spectrum of light in real time, then use AI to “predict” the chemical concentrations of the substance under analysis. To confirm the results, Naveh’s device then uses a deep-learning neural network for spectral recognition that’s been trained to interpret against calibrated references. The more data the network is fed, the better it gets at correcting predictions—and delivering spectroscopy-level accuracy. The development was the subject of an article in the prestigious journal Science Advances last year.
Above and banner: Prof. Doron Naveh in his 2D Electronics Laboratory.
And the best part? As compared to a standard spectrometer, which can easily run in the hundreds of thousands of dollars, Naveh’s computational version costs as little as a few cents. “Both the device’s small size and extreme cost efficiency make it a candidate for a vast range of interpretative possibilities,” he says. He notes that last year, when IDF soldiers fell seriously ill from an infection caused by pathogens in the Gaza Strip, he realized that a medic using his device in the field could have determined the presence of the pathogens, thus enabling immediate intervention. “My goal with this and all the devices on which I work is simple: Do good in the world,” says Naveh, who is also involved in a NATO project that’s developing super-efficient, safe imaging sensors for use in homeland-security and medical applications. “If it can help people make more informed decisions and encourages industries’ accountability, that will indeed make the world much healthier. And that would be a very good thing indeed.”
