Have you ever bitten into a piece of chocolate, expecting a smooth, rich taste, but experienced an unexpected and unpleasant sour flavour?
This taste is “rancidity”, resulting from food spoilage, and it is a problem that “artificial intelligence” can help solve, by directing researchers to some new mixtures of antioxidants, which help extend the life of food products, according to a new study by researchers from Department of Chemistry at Clemson University, USA, published in the latest issue of the Journal of Agriculture and Food Chemistry.
How does food spoil?
Food spoils when it is exposed to air for a period of time, a process called oxidation. Many common ingredients, especially fats and oils, react with oxygen, and the presence of heat or ultraviolet radiation can speed up this process.
Oxidation leads to the formation of smaller molecules such as ketones, aldehydes and fatty acids, which give spoiled foods a distinctive flavor and strong metallic odor, and eating them repeatedly poses a threat to health.
The protective shield against this process, whether in natural or manufactured foods, is antioxidants, which include a wide range of natural molecules, such as vitamin C, and synthetic molecules capable of protecting food from oxidation.
But the process of using antioxidants is much more complex than simply sprinkling a little vitamin C on food and expecting a protective effect, as researchers must choose a group of antioxidant mixtures and accurately calculate the amount of each of them. However, combining them does not always lead to enhanced antioxidants. Their effect, as there are cases where using the wrong antioxidants, or mixing them in the wrong proportions, can reduce their protective effect, and this is called “antagonism.”
Search for new groups
Discovering the appropriate antioxidant combinations for any type of food requires many experiments, which are time-consuming, require specialized personnel, and increase the overall cost of the food, so researchers are stuck with a small number of mixtures that provide only a certain level of protection against rancidity, and this is where the role comes into play. Artificial intelligence tools.
“As chemists, we wanted to teach artificial intelligence how to search for new combinations of antioxidants, so we chose a type that was able to work with... Text representations, which are written symbols that describe the chemical structure of each antioxidant.
They added, "We fed the artificial intelligence program with a list of about a million chemical reactions, and taught the program some simple chemistry concepts, such as how to identify important features of molecules. Once the program was able to recognize general chemical patterns, such as how some molecules interact with each other, we activated it from "While teaching him some advanced chemistry, in this step we used a database of about 1,100 mixtures previously described in the research literature."
The researchers explain, “At this stage, artificial intelligence can predict the effect of combining any two or three groups of antioxidants in less than a second, but these predictions were not completely consistent with the experiments our team conducted in the laboratory, where we found that artificial intelligence was able to... "Only some of the oxidation experiments we conducted using real fats correctly predicted the results, demonstrating the complexities of transferring results from a computer to the laboratory."
Improve the program
Because artificial intelligence programs are not static tools, but rather dynamic, the research team then continued to feed the program with new data in order to improve its predictive capabilities, so that it accurately predicts the effect of each group of antioxidants. The more data the model obtains, the more accurate it becomes. This is very similar to how human performance develops through learning.
“We found that adding about 200 examples from the lab enabled the AI to learn enough chemistry to predict the results of the experiments our team is performing,” the researchers say, “which means that AI can become a very intelligent assistant, and increases the ability of scientists to develop better ways to preserve... food by coming up with the best antioxidant combinations.”
The most important advantage is in the approach that the researchers used, as Lucas Iris explains in special statements via email to Al Jazeera Net, that “the researchers combined the use of a training data set and new data from laboratory experiments, and this strategy enabled the program to learn more about antioxidants and make predictions.” more accurate".
He adds, "Compared to traditional methods for finding new groups of antioxidants, the program allows speed, as we can screen millions of antioxidant groups and then conduct (real) experiments with the most promising ones. This saves a lot of time and a lot of money, and one of the other factors." "What needs to be taken into consideration is that instead of just knowing whether a mixture of antibodies is synergistic (or not), the software allows us to work with the group that shows the highest probability of being synergistic, rather than working with each group alone."
The next step that the research team will work on is to improve the program by implementing alternative methods for training it, so that we can make predictions related to other variables, including time, temperature, and other components in the fat sample.
Advantages of harnessing artificial intelligence
Peter Modiaga Itaware, a professor at the Faculty of Science at the University of Ibadan in Oyo State, Nigeria, praises the mechanism that the researchers used in this study to harness artificial intelligence in the field of arriving at a better mixture of antioxidants, saying in statements via email to Al Jazeera Net, "More lives will be saved if we implement the use of intelligence." "Rationally artificial in studies of human life and others."
Itawyer, who two years ago published a study in the journal Food Chemistry on the effects of cocoa phytochemistry on the nutritional chemistry of chocolate, adds: “You might think that it is easy to combine two powerful antioxidants for the purpose of food preservation, and I don’t need to remind you that if you combine... Strong acid and base, the result will be salt + water, which is an antidote reaction.”
He continues, "Therefore, if there are not sufficient tests to show whether combining two or more antioxidants will lead to a synergistic or antagonistic result, the global food industry will fail to extend the shelf life of its products, and the coordinated use of artificial intelligence in these trials will do better." "The increasing numbers of people are in dire need of processed foods."
“This will reduce the likelihood of developing diseases related to antioxidants (free radical activities) such as cancer, cardiovascular disease, etc.,” he believes.
He stresses 5 conditions that must be taken into account when employing artificial intelligence in this field or any field related to human health, which are:
- The use of AI must be by a global body of renowned or experienced scientists.
- There must be safe limits to the use or application of AI in human-related studies.
- Terms and conditions of use must be approved by a global body.
- The global body should evaluate any AI program before issuing the green light for its use.
- Quality assessment of all AI robots or programs should be carried out regularly and systematically.
Post a Comment