Geotechnic engineers have a long history of using artificial neural networks to make decisions about how to manipulate soil in a field.
The ability to use a neural network to identify specific soil types and the associated conditions that need to be remediated is one of the most valuable tools in the geotechatical engineering industry.
In a paper published in Science Advances, a team from Georgia Tech, the University of Alabama at Birmingham and the University de Villemont examined the use of a neural machine learning system to identify soil conditions in a lab-scale system.
They discovered that this machine learning method outperformed traditional approaches and also could be used to identify plant species in a similar way.
The team first applied their approach to the field of agriculture.
They trained a neural net on the soil conditions to identify the plant species and their root zones.
This network then compared the results with the plant’s actual soil conditions.
The system’s predictions for root zones and soil conditions were also used to classify different plant species.
In a separate experiment, the researchers tested the neural machine-learning system against existing plant species to predict the soil characteristics of the same plant species that the system was trained on.
They found that the neural net outperformed the traditional methods by more than 50 percent.
This shows that the approach could be very useful in geotechanical engineering, the authors say.
It’s important to note that this neural network system does not have any natural or human-made components.
It is designed to work in a laboratory setting.
The team says that the technology could be a useful tool in agricultural applications, and they plan to integrate the technology into an agricultural automation system.
However, the system will not be used for anything other than this one specific field.
“This is not a replacement for other techniques for identifying plant species,” the researchers write in their paper.
They also say that it is important to be cautious about using neural machine intelligence to make complex decisions.
The researchers say that while they are confident that the technique can be used in the lab, they are not sure that the model they used is robust enough to make use of real-world conditions.
“We have used these methods to identify and classify plants from the ground, but we are not certain that this technique will work in field settings,” they write.
“It is important for us to be aware of the potential of these methods and be aware that the process is very crude.”