The machine learning can reveal optimal growth conditions to optimize taste and other characteristics – ScienceDaily



[ad_1]

What needs to be done to make the plants taste good? For scientists at the MIT Media Lab, it is necessary to combine botany, machine learning algorithms and good traditional chemistry.

Using all of the above, researchers at the Media Lab's Open Agriculture Initiative report that they've created basil plants that are probably more delicious than the ones you've ever tasted. No genetic modification is involved: researchers used computer algorithms to determine optimal growth conditions to maximize the concentration of flavoring molecules known as volatile compounds.

Caleb Harper, senior researcher at MIT's Media Lab and director of the OpenAg group, says Caleb Harper is just the beginning of the new field of "e-agriculture". His group is currently working to improve the properties of disease-fighting plants, and also hopes to help producers adapt to climate change by studying how crops grow under different conditions.

"Our goal is to design an open source technology at the intersection of data acquisition, detection and machine learning, and apply it to agricultural research and development. A way that has never been done before, "said Harper. "We are really interested in building networked tools that can take the experience of a plant, its phenotype, all the constraints it meets and its genetics, and digitize this to allow us to understand the plant-environment interaction. "

In their study of basil plants, published in the April 3 issue of PLOS ONE, the researchers discovered, to their great surprise, that exposing the plants to light 24 hours a day gave the best taste. Traditional farming techniques would never have given this idea, says John de la Parra, head of research for the OpenAg group and author of the study.

"You would not have been able to discover that otherwise – if you're not in Antarctica, there's no 24-hour photoperiod to test in the real world," he says. "You must have had artificial circumstances to discover that."

Harper and Risto Miikkulainen, professor of computer science at the University of Texas at Austin, are the lead authors of the paper. Arielle Johnson, a board member of the Media Lab, and Elliot Meyerson of Cognizant Technology Solutions are the lead authors, and Timothy Savas, special project assistant at the Open Agriculture Initiative, is also an author.

Maximize the flavor

Located in a warehouse in Middleton, Massachusetts, OpenAg plants are grown in suitable shipping containers so that environmental conditions such as light, temperature and humidity can be carefully controlled.

This type of agriculture has many names – controlled environmental agriculture, vertical agriculture, urban agriculture – and remains a niche market, but is growing rapidly, Harper says. In Japan, such a "plant plant" produces hundreds of thousands of lettuce heads each week. However, many efforts have also failed and very little exchange of information between companies trying to develop such facilities.

One of the goals of the MIT initiative is to overcome this type of secrecy by making all OpenAg hardware, software and data available for free.

"There is a big problem right now in the agricultural sector in terms of lack of publicly available data, insufficient standards for data collection and lack of data sharing," Harper said. "While machine learning, artificial intelligence and advanced algorithm design have progressed so rapidly, the collection of meaningful and well-labeled agricultural data has lagged behind. Our tools being open source, we hope that they will spread faster and create the ability to do it … networked science together. "

in the PLOS ONE study, the MIT team has been striving to demonstrate the feasibility of their approach, which involves growing plants in different sets of conditions in hydroponic containers that they have baptized "food computers" ". This configuration allowed them to vary the duration of the light and the duration of exposure to ultraviolet light. Once the plants were fully developed, the researchers assessed the taste of basil by measuring the concentration of volatile compounds in the leaves, using traditional analytical chemistry techniques such as gas chromatography and mass spectrometry. These molecules contain valuable nutrients and antioxidants. Thus, enhancing the flavor can also offer health benefits.

All the information from the factory experiments was then integrated into the machine learning algorithms developed by the MIT and Cognizant (formerly Sentient Technologies) teams. The algorithms evaluated millions of possible combinations of light duration and UV rays and generated sets of conditions to maximize aromas, including the 24-hour daylight regime.

Beyond the flavor, researchers are currently developing basil plants containing more compounds that can help fight diseases such as diabetes. Basil and other plants are known to contain compounds that help control blood sugar, and in previous work Parra has shown that these compounds can be enhanced by various environmental conditions.

Researchers are currently studying the effects of adjusting other environmental variables such as temperature, moisture and light color, as well as the effects of adding plant hormones or other nutrients. In one study, they expose plants to chitosan, a polymer found in insect shells, which allows the plant to produce different chemical compounds to ward off insect attack.

They are also interested in using their approach to increase yields of medicinal plants such as Madagascar periwinkle, which is the only source of anticancer compounds, vincristine and vinblastine.

"You can see this document as the starting point for many applications that can be applied, and it's an exhibition of the power tools we've built so far," Parra said. "It was the archetype of what we can now do on a larger scale."

Climate adaptation

The researchers say that another important application of e-agriculture is adaptation to climate change. While it usually takes years, even decades, to study the impact of different conditions on crops, in a controlled agricultural environment, many experiments can be performed in a short period of time.

"When you grow plants in a field, you have to rely on the weather and other factors to cooperate, and you have to wait for the next growing season," Parra said. "With systems like ours, we can dramatically increase the amount of knowledge that can be acquired much faster."

The OpenAg team is currently conducting a hazelnut study for candy maker Ferrero, which consumes around 25% of the world's hazelnuts.

As part of their educational mission, researchers have also developed small "personal computers for food", boxes that can be used to grow plants under controlled conditions and return data to the community. MIT team. These are now used by many high school and college students in the United States, within a network of diverse users spread across 65 countries, who can share their ideas and results via a forum online.

"For us, every box is a data point that interests us a lot, but it's also an experimental platform for teaching environmental science, coding, chemistry and physics. mathematics in a new way, "Harper says.

The research was funded by Target Corp., the health product group Lee Kum Kee, Welspun, Sentient Technologies and Cognizant Technology Solutions.

[ad_2]

Source link