Course:EOSC311/2023/Artificial Intelligence in Natural Geology
An overview on how artificial intelligence has been implemented into the field of geology over the years and how it has helped or can help enhance the jobs of geologists.
Introduction
Cognitive Systems is the major that focuses around the design and implementation of artificial intelligence by looking into and understanding how the human brain works in relation to our bodies. As humans continue to explore ways to replicate human intelligence in machines, artificial intelligence has become increasingly popular for many fields of work, including geology. Geology has always been done manually with the help of machines, but it had never seen fully automated processes until the rise of artificial intelligence. Processes such as mining, transportation, and dating have all been done by humans and tools despite there being a risk to human health and safety. With the development of artificial intelligence, we can reduce (or in some areas completely eliminate) high risk jobs for humans and provide a safe and consistent way to perform geological tasks. It is important that geology incorporates more and more artificial intelligence to minimize the risk certain tasks have on individuals, to better prevent or predict major disasters to allow for safety measures to be implemented ahead of time, and to further aid our knowledge of the planet we occupy and hope to preserve for years to come. This page aims to explain how artificial intelligence correlates to geology while exploring the current and potential uses of artificial intelligence in the field of geology.
Artificial Intelligence
Artificial intelligence had an initial birth period between 1940-1960 with the introduction of cybernetics. The field saw a massive increase in the 2010s when computing power and data power spiked significantly. [1] The field of artificial intelligence and Cognitive Systems concerns itself with understanding the human cognitive mechanisms that go into processing and producing actions. It combines the fields of psychology, philosophy, linguistics, and computer Science for the design and implementation of artificial intelligence. The basis of artificial intelligence can be briefly divided into two main components: Predictive Models and Automated Machines. Both distinguishing parts of artificial intelligence can be used to automate tasks used in many different profession such as advertising, finance, and of course, geology. Lots of websites and companies make use of predictive modelling to create suggestions for users, also known as a recommendation system. A popular company that uses this is the movie streaming service Netflix.[2] Artificial intelligence makes use of many training techniques such as supervised and unsupervised training. Supervised training involves training a machine with data that has a "correct" solution for the targeted value and unsupervised training involves training without a "correct" target value.
Predictive Models
One area of artificial intelligence that is used quite commonly is predictive models. With the vast increase in data collecting and storing powers in machines, this component of artificial intelligence has become a popular and consistent way to use past data in predicting outcomes. Essentially, the machine uses a set of data to train its predictive behavior and it is tested on another portion of the data before being released. The past data given to the machine would include multiple columns that we call features and a column that we identify as the "target column" or what it is we want to predict. This dataset is split into what we call a train split and a test split. The train split is the data that the machine will use to learn the features that correlate to a certain target outcome, and test the accuracy, precision, and recall on the test split. Accuracy, precision, and recall are three of the many ways we can evaluate the reliability and correctness of a predictive model.
Accuracy- The percentage of correct predictions over all total predictions
Precision- The percentage of True positives over all the predicted positives
Recall- The percentage of True positives over all actual positives
Automated Machines
Another area of artificial intelligence that most people think of when presented with the words artificial intelligence is automated machines. These are essentially machines that perform tasks that humans can, but the processes are completely automated. Automated machines can range from robotic human replicas to self-driving cars and can be used in a variety of fields worldwide. We have seen recent examples of automated machines in the self-driving capabilities of Tesla cars and in artwork. A great example of this is the Can't Help Myself robot.
Challenges of Manual Geology
Geologists are concerned with many different aspects within the field of geology which can sometimes be stressful and very time consuming. There are also some major aspects of geology that we consider to be more urgent and of extreme importance. One of these aspects is the continuing efforts to predict and prevent natural disasters and climate change. The Earth is made up of tectonic plates that can do one of three things: diverge, converge, or transform. Each of these plate boundaries are the primary causes for many of Earth's catastrophic disasters such as earthquakes, tsunamis, volcanic eruptions, and landslides.[3] In order to predict natural disasters, geologists must look into the history of tectonic plates which can take an extensive amount of human power and time. Many of these disasters occur when tectonic plates collide, where the Oceanic plate subducts beneath the Continental plate causing tremors, shakes, and potential expulsion of magma, as depicted in the diagram. Predicting natural disasters is also notoriously difficult and no scientist has every predicted a major earthquake before.[4] Predicting earthquakes has been difficult for scientists because turning early warning cues into a where, when, and how big is not a trivial thing to do.[4] Yet, this is one of if not the highest stake tasks that geologists and scientists face today.
Another important aspect of Geology that may be difficult to do manually is dating. There are two forms of dating that geologists use, absolute dating and relative dating. The latter of the two encompasses the idea of fossil dating. Fossil dating is the act of determining the relative age of a fossil in relation to its time period. Using fossil dating, geologists are able to put a relative age to a rock that the fossil was found in or around by comparing it to other rocks ages. A lot of steps go into dating a rock and manually doing this is very time consuming and requires a lot of detail and research. From excavation, transportation, preservation, examination, to research, geologists have a lot of tasks at hand that may be benefitted by the application of Artificial Intelligence into Geology.
Use of Artificial Intelligence in Geology and its Connection
Artificial intelligence is connected to the field of geology through many ways. Geologists currently use machine learning applications in mining, predicting, and for data optimization. Artificial intelligence is essentially a tool that can help geologist optimize and accurately produce results on a consistent basis, taking away from the strain and time it takes to perform tasks manually. The use of artificial intelligence ranges from autonomous transportation to modelling and predicting tectonic plate movement. All of which has been remarkably beneficial to the field of geology.
Mining
With many aspects within Geology requiring a vast amount of time, data, and labor, the use of artificial intelligence has started to impact the way geologists perform tasks. Due to the mining industry being under more and more pressure to increase efficiency[5], artificial intelligence has become a theme that will have significant impact on mining companies[5]. Artificial intelligence has gotten to the point where issues such as safety and sustainability at mine sites can be monitored and predicted through the use of artificial intelligence and tools such as computer vision, smart robots, data science, and machine learning[5]. The emergence of artificial intelligence in the mining industry allows for better resource site predicting capabilities which minimizes the expenses needed to explore potential mine sites.
Artificial intelligence is also being used through machine learning techniques to help companies find minerals for extraction[6]. This is critical to any form of mining as it reduces costs and creates efficiency to opening mines. Being precise when finding areas to mine makes the mining industry more profitable which is why mining companies like Goldspot Discoveries Inc and Goldcorp and IBM Watson have been using artificial intelligence as of late[6].
Although not all mining companies have made the switch over to artificial intelligence, some companies such as Fortescue Metals Group, Rio Tinto, and BHP have led the charge in the introduction of artificial intelligence to mining. These companies use artificial intelligence in a variety of ways, ranging from autonomous haulage projects (Fortescue Metals Group has almost 200 autonomous trucks in Solomon and Chichester hubs[5]) to automating operations to improve safety, efficiency, and to reduce the cost operations. This can be found in the use of trucks, drills, and trains that remove the risk of driver error to improve overall safety. Mining companies around the world have been slowly shifting towards the new age of mining with the help of artificial intelligence, but there is still a long way to go in the collaboration of artificial intelligence and geology.
Predicting Geology
Artificial intelligence is a beneficial tool that can analyze large datasets which is perfect for a geologist. Geologists have to work with massive amounts of data at times as their work includes the usage of satellite imagery, seismic data, and geological samples[7]. Finding trends and patterns within large datasets is often difficult to do, but with the use of artificial intelligence algorithms, processes like these can be done quickly and efficiently. Currently, machine learning algorithms have been used to analyze seismic data to identify previously unidentified micro-earthquakes[7] which provides valuable information on tectonic plates. This is a major advance in geology because the identification allows geologists and researchers to better understand the processes that drive natural disasters caused by plate tectonics. Artificial intelligence can not only help us understand Earth's past and present, but it can also help us to predict its future.
The capabilities of artificial intelligence has also been exploited to generate new data through computer simulations to model complex geological processes[7]. These simulations can be used by researchers to better predict trends of tectonic action and may potentially predict natural disasters as well. Another major implication of artificial intelligence and its predicting power is in the study of climate change. With the multitude of simulations artificial intelligence can run, geologists and researchers can gain insight on the drivers of climate change and make more accurate predictions. Climate change is a fast approaching issue that society must tackle, and with the use of artificial intelligence, our researchers can inform governments on how to better handle the ever changing temperatures. From the discovery of climate change drivers to the action plans developed by nations, artificial intelligence has been a major aid in combating the global problem.
Big Data
Geology has recently been introduced to the concept of big data which is a result of an accumulation of historical data, the emergence of new technologies for data storage, and the overall quantity, value, diversity and timeliness of geological data. Big data means that the old ways of studying data in geology is no longer able to meet the requirements needed in terms of data processing methods and speed[8]. Due to the lack of sustainable methods to process data, geology has turned to artificial intelligence. The use of artificial intelligence on big data allows for analysis on vast amounts of geological data to be performed, returning a summary and an impartial, quick, and scientific analysis on the geological work. Geologists have also been using artificial intelligence and machine learning to solve geochemical anomaly identification[8] because traditional mathematical and statistical methods are ineffective for this. The use of machine learning also benefits geologists in their work on geochemical anomaly identification because the methods do not make assumptions about the distribution of data whereas humans are innately biased to assume[8]. Overall, the introduction of big data to geology can really only be solved through the use of artificial intelligence and the continuous growth in collaboration of the two fields is undeniably important.
Further Applications of Artificial Intelligence in Geology
Although artificial intelligence is becoming more and more popular within the field of geology, there are still many different types of applications that are not being utilized. We have seen the use of machine learning in geology through the use of autonomous transportation, computational modelling and predicting, and in the optimization of big data, but there is still a lot artificial intelligence can provide for geology. As humans learn more to decipher the code behind artificial intelligence, the field of geology can potentially see things like automatic fossil identification and dating, as well as improved predicting capabilities catered to more large scale tectonic movements.
Fossil & Rock Dating
Despite the fact that fossil and rock dating has been almost entirely digitalized in the field of geology, there have been little to no applications of artificial intelligence to optimize the process. Although the work to extract the fossils and rocks may be beyond the capabilities of current artificial intelligence, machine learning should still be capable of quickly identifying fossils and rocks if given the necessary information.
With a vast amount of big data in geology, a machine learning algorithm would simply have to pull up records of matching the input data given by a user. If geologists were able to use artificial intelligence to automatically identify and date fossils and rocks, it would save them a substantial amount of time that can be put towards other tasks. The time it takes to pull up large amounts of records to match an extracted fossil or rock to pre-existing discoveries would be obsolete in the face of artificial intelligence.
Predicting Capabilities
Current artificial intelligence algorithms are able to model and predict tectonic plate movements but on a small scale. Majority of the earthquakes that artificial intelligence has been able to identify so far are small-scale ones, but the disasters that drastically affect peoples lives are typically much larger in scale. Therefore, with the vast amounts of historic data surrounding plate tectonics, a future direction that geologists should take artificial intelligence to is towards predicting large-scale disasters caused by tectonic movement. The impact that something like this could have on peoples live is substantial enough to reasonably put more effort and resources into. Although it may not be possible at the end of the day, this particular use of artificial intelligence is one that society should work towards.
Conclusion
Artificial intelligence has grown in leaps and bounds since its introduction in the early 1940's. With the increased amounts of research efforts being put towards the replication of a human mind within robots, it is not surprising to see the emergence of artificial intelligence in other fields of work. Geology is one of those fields that has benefited from the ever growing improvements to machine learning algorithms, and because it is still relatively early in the two fields collaboration with one another, there is still a lot left to build towards. With a surplus of geological tasks requiring a large amount of time and work power, the introduction of artificial intelligence to geology has provided a much needed relief. We have seen the use of artificial intelligence in the mining industry, in computational modeling of tectonic plates, and in the optimization of geological big data, but we still need artificial intelligence to takes a next step towards solving other geological conundrums. As we continue to improve artificial intelligence, further applications can be developed for geological use. Despite a lot of the current usage of machine learning being directed towards optimizing profits, geology can take a major step forward by researching ways to benefit the public. One way of doing so is to put the effort and resources into an algorithm that can predict large-scale disasters caused by tectonic movement. Although we may be a long way from perfecting such an advanced system, geology and artificial intelligence can one day make a significant impact on the lives of many through its continued collaboration with one another.
References
- ↑ History of artificial intelligence - artificial intelligence - www.coe.int. Artificial Intelligence. (n.d.). https://www.coe.int/en/web/artificial-intelligence/history-of-ai#:~:text=1940%2D1960%3A%20Birth%20of%20AI,of%20machines%20and%20organic%20beings.
- ↑ Mathur, V. (2023, March 14). Netflix’s use of artificial intelligence algorithms. Analytics Steps. https://www.analyticssteps.com/blogs/netflixs-use-of-artificial-intelligence-algorithms
- ↑ Gomberg, J. S., & Ludwig, K. A. (2017). Reducing risk where tectonic plates collide. Fact Sheet. https://doi.org/10.3133/fs20173024
- ↑ 4.0 4.1 Nature Publishing Group. (2023, February 17). Why can’t we predict earthquakes? Nature News. https://www.nature.com/articles/s43588-023-00418-1
- ↑ 5.0 5.1 5.2 5.3 Leading mining companies in the Artificial Intelligence Theme. Mining Technology. (2023, May 22). https://www.mining-technology.com/data-insights/leading-mining-companies-in-the-artificial-intelligence-theme/#:~:text=According%20to%20GlobalData%E2%80%99s%20thematic%20research,Gold%2C%20and%20Dundee%20Precious%20Metals.
- ↑ 6.0 6.1 Marr, B. (2021, July 13). The 4th Industrial Revolution: How Mining Companies are using AI, Machine Learning and Robots. Bernard Marr. https://bernardmarr.com/the-4th-industrial-revolution-how-mining-companies-are-using-ai-machine-learning-and-robots/
- ↑ 7.0 7.1 7.2 Bonis, A. D. (2023, May 23). Ai in geology: Analyzing Earth’s past and predicting its future. CityLife. https://citylife.capetown/uncategorized/ai-in-geology-analyzing-earths-past-and-predicting-its-future/26135/
- ↑ 8.0 8.1 8.2 Chen, L., Wang, L., Miao, J., Gao, H., Zhang, Y., Yao, Y., Bai, M., Mei, L., & He, J. (2020). Review of the application of big data and artificial intelligence in Geology. Journal of Physics: Conference Series, 1684(1), 012007. https://doi.org/10.1088/1742-6596/1684/1/012007
 |
This Earth Science resource was created by Course:EOSC311. |