IBM Builds an AI-Powered Electronic Tongue

A diverse team of electrochemists, material scientists, and electrical and software engineers at IBM Research worked together to create the “tongue”—called Hypertaste—over the last few years. While portable, single-job systems like pH testers already exist (as do laboratories’ massive and expensive molecular analysis tools), the scientists were interested in creating a relatively small and affordable system that could identify liquids’ complex chemical compounds. The end product uses an array of 16 conductive polymeric sensors to reveal a variety of characteristic data in just one minute.

To use Hypertaste, an operator dips the array into a liquid. The sensors react to the chemicals present in the liquid by initiating a series of voltage signals, which are considered the liquid’s “chemical fingerprint.” These signals are relayed to the operator’s mobile device, where a trained machine learning algorithm compares the signals against a database of chemical compositions. Once a matching composition is identified, the operator can dive into the details of the liquid’s makeup.

The team at IBM Research had to train the algorithm themselves. “Like a budding sommelier learning the intricacies of wine tasting, the Hypertaste sensor array needs to be trained to identify the liquids of interest before being put to the test,” IBM said in a blog post. “This is done by measuring the sensor array response in those liquids multiple times and then feeding the resulting data into a machine learning model which extracts the characteristic features associated with each liquid.”

IBM’s electronic tongue isn’t the first robotic body part to perceive sensory information typically reserved for us living organisms. A couple months ago we reported on the “e-nose,” a scent-sensing prototype developed by researchers at the University of Technology Sydney. The e-nose uses gas sensors to pick up odor molecules, which are then deciphered by an algorithm capable of sorting the liquid’s characteristics. In fact, Hypertaste and the e-nose share a common purpose: to detect counterfeit versions of valuable liquid products, like liquor and perfume.

But Hypertaste’s uses aren’t limited to the upscale black market. IBM has already used the electronic tongue to look for signs of ocean acidification aboard the Mayflower, an autonomous marine research vessel. The system might also be used to ensure consistent food and beverage quality in manufacturing, as well as develop new flavors. Hypertaste might even be useful for assessing one’s metabolic fingerprint: one “taste” of a patient’s urine could reveal insights regarding their lifestyle and nutrition, thus aiding the medical industry.

Continue reading

Introducing TRACBench, a New AI-Powered Transcoding Benchmark

Three of the fastest workstation CPUs on Earth slug it out in an AI-powered transcoding benchmark we designed to scale from dual-core to 64-core systems.

AI-Powered Electronic Design Automation Tools Could Redefine Chipmaking

Samsung has announced it's using AI EDA tools to build its Exynos processors. Expect more announcements in the future, but geared more towards gradual adoption than a sudden surge.

Google Is Testing Its Latest AI-Powered Robot In a Kitchen

Using a 540-billion parameter language network, this wheeled robot can assess its own capabilities, its environment, and the task at hand, then break that task into smaller sub-tasks to achieve the desired goal.

Scientists Create AI-Powered Laser to ‘Neutralize’ Cockroaches

Why use traps and pesticides when you could zap a roach away?