With the advent of technology, machines have been increasing the efficiency of human tasks. Completion of tasks with less time and enhanced accuracy makes them reliable. It sometimes does things that otherwise be impossible for humans to do. When organizations and governments are revamping their mode of work with technology, researchers worldwide are finding means to improve the existing technology by pushing its limitations further.
In this context, the IBM researchers are stepping forward with a new project to decipher if it can give AI a sense of taste. To study this, they chose a liquid free of flavor and is very difficult for man to tell apart- water. For testing, the computer was familiarized with different water samples that had slight variations in composition due to their mineral content. Then, the computer was put against human competitors to see who could identify a familiar water sample. In this test, humans lost, and the computer won.
According to Patrick Ruch, the AI system was better than human tasters in distinguishing four different kinds of mineral water. Patrick Ruch is the lead researcher on IBM’s AI-assisted e-tongue technology called Hypertaste. Hypertaste caters to a wide range of industrial and scientific users with a growing need to identify liquids swiftly without access to high-end laboratories.
Scope and functioning
IBM researchers are currently working with their industry partners to identify the possible use cases of the model. Hypertaste is a few years away from its commercial use. However, the team is now finding its scope with food and beverage companies to predict different kinds of flavors and identify coffee, soft drinks and other offerings. Patrick Ruch said that Hypertaste is not intended to replace human experts. But instead offload some of the most mundane or complex tasks. It can also be used to ensure the authenticity of wines or whiskeys or assist in product innovation by different flavors. Future applications of the model could also include finding contaminants in drinking water, tracing raw materials throughout the supply chain and checking for the presence of foodborne illness.
Since complex liquids contain many different molecules, it would be inefficient to identify them by sensing each component separately, so Hypertaste uses combinatorial sensing. This resembles our natural senses of taste and smell, where we don’t have a receptor for each molecule occurring in every kind of food or drink. It uses electrochemical sensors comprised of pairs of electrodes, each responding to the presence of a combination of molecules utilizing a voltage signal. These electrochemical sensor’s function using polymer coatings.
IBM has built sensor arrays and combined them with off-the-shelf electronics that they configure to measure the voltages across the electrodes in an array and relay them on a mobile device. The mobile app transfers the data to a cloud server where a trained ML algorithm compares the digital figure listed from the database of the liquids—the algorithm figures out which liquid in the database is similar to the one under investigation. The result is reported back to the mobile app. This process, which they call Classification, will take less than a minute. The sensors are trained by measuring the response in liquids multiple times and feeding the data into the ML model. The probability and speed of the sensors are achieved using combinatorial sensing with an array of cross-sensitive sensors combined with intelligent software outsourced from the cloud.
Increased value with Hypertaste
With Hypertaste, it can be proved that a portable device could be capable of rapid fingerprinting complex liquids. Industries and services would benefit from such technology. For instance, once food and drinks are packaged, there is little ability to verify that the package contains what is on the label apart from seeing the product to a lab for testing. It is much harder for suppliers to insert lower-quality products into the supply chain to fool Hypertaste. The model can also be used in diagnostics or preventive medicine, clinical trials etc. The spectrum of possible applications is vast and spurs the imagination. IBM researchers are confident that AI-assisted portable chemical sensors will meet the needs of many industries.
Source: indiaai.gov.in
