I aim to systematically investigate the use of machine learning (ML) techniques for the solution of problems that are relevant to geological applications, with emphasis on petrological, geochemical, and volcanological applications. This line of research was sparked in 2014 thanks to the funding grants “Challenge: Coupling macHine leArning, Leading anaLytical mEthods aNd larGe gEochemical datasets in tephra studies” (PI: Maurizio Petrelli, 2015/2017), “ENGAGE: machinE learNinG Applications for Geological problEms” (PI: Maurizio Petrelli, 2019/2020), the Microsoft Machine Learning Award (PI: Maurizio Petrelli, 2015) provided by the University of Perugia and Microsoft Inc., respectively. It is now supported by many active grants that I lead or act as co-PI.  

I am expecially proud of three manuscripts that sparked widespresd interest on ML witihn the petrologic and  volcanologic community. In 2016, the manuscript titled "Solving Petrological Problems by Machine Learning" introduced the potential of ML to the petrological community by exploring the application of ML classification techniques for tectonic setting recognition based on geochemical data. A 2017 study extended these methods to volcanology, specifically in tephrochronology: The manuscript, "Combining Machine Learning Techniques, Microanalyses, and Large Geochemical Datasets for Tephrochronological Studies in Complex Volcanic Areas: New Age Constraints for the Pleistocene Magmatism of Central Italy," demonstrated how ML could unlock insights in complex volcanic systems. In 2020, the manuscript titled "Machine Learning Thermo‐Barometry: Application to Clinopyroxene‐Bearing Magmas" introduced ML regression tasks into petrology. This work showcased how ML could be used to estimate magmatic pressure and temperature conditions, offering new and complementary tools for advancing thermodynamic modeling in petrological studies. To consolidate the field’s progress, the perspective on petrology manuscript "Machine Learning in Petrology: State-of-the-Art and Future Perspectives" provided a comprehensive overview of the current state of ML applications in petrology and outlined future research directions. I am currently contributing a chapter to the third edition of the Encyclopedia of Volcanoes, titled "Methodological Advances in Volcanology: The Role of Artificial Intelligence in Boosting Automation, Modeling, and Discovery." This chapter aims to highlight recent advancements and explore how artificial intelligence is revolutionizing methodological approaches in volcanology.

From June 21 to July 5, 2023, I spent a fantastic time in Beijing and Hangzhou by visiting the China University of Geoscience and the Zhejiang University, respectively. There, I held seminars on using ML techniques in Earth Sciences and a plenary lecture titled "Machine Learning in Petrology and Volcanology."

Selected Publications:

1. Petrelli, M., Anantrasirichai, N., Bean, C. J., Biggs, J., Malfante, M., Wilding, J. D., Zhu, W. (2025). Methodological Advances in Volcanology: The Role of Artificial Intelligence in Volcano Monitoring, Modelling, and Hazard Assessment. The Encyclopedia of Volcanoes 3rd Edition. Bonadonna C. (Ed). Elsevier.  Accepted for Publication

2. Petrelli M., 2024. Machine Learning in Petrology: State-of-the-Art and Future Perspectives. Journal of Petrology. doi: 10.1093/petrology/egae036

3. Petrelli M., 2023. Machine Learning for Earth Sciences. Springer. ISBN:978-3-031-35113-6

4. Jorgenson C., Higgins O., Petrelli M., Bégué F., Caricchi L., 2022. A Machine Learning-Based Approach to Clinopyroxene Thermobarometry: Model Optimization and Distribution for Use in Earth Sciences. JGR: Solid Earth. doi:10.1029/2021JB022904

5. Petrelli M., Caricchi L., Perugini, D. (2020). Machine Learning Thermo‐Barometry: Application to Clinopyroxene‐Bearing Magmas. JGR: Solid Earth. DOI: 10.1029/2020JB020130

6. Caricchi L., Petrelli M., Bali E., Sheldrake T., Pioli L.,  Simpson G. (2020). A Data Driven Approach to Investigate the Chemical Variability of Clinopyroxenes From the 2014–2015 Holuhraun–Bárdarbunga Eruption (Iceland). Frontiers in Earth Science. DOI: 10.3389/feart.2020.00018

7. Petrelli M., Bizzarri R., Morgavi D., Baldanza A., Perugini, D. (2017). Combining machine learning techniques, microanalyses and large geochemical datasets for tephrochronological studies in complex volcanic areas: New age constraints for the Pleistocene magmatism of central Italy. Quaternary Geochronology. DOI: 10.1016/j.quageo.2016.12.003

8. Petrelli M., Perugini D. (2016). Solving petrological problems through machine learning: the study case of tectonic discrimination using geochemical and isotopic data. Contributions to Mineralogy and Petrology. DOI: 10.1007/s00410-016-1292-2 

This line of research focuses on the geochemical, petrological, and volcanological characterization of magmatic systems, with particular emphasis on time-scale estimates of magmatic processes occurring in volcanic plumbing systems before eruptions. To do that, I combine numerical methods, experimental petrology, and the study of natural samples. The main aim is to provide methods and tools to support public authorities and decision-makers in volcanic hazard assessment and risk mitigation. 

Selected Publications:

1. Petrelli M., Ágreda López M., Pisello A., Perugini, D. (2023). Pre-eruptive dynamics at the Campi Flegrei Caldera: from evidence of magma mixing to timescales estimates. Earth, Planets and Space, doi: 10.1186/s40623-023-01765-z

2. Rooyakkers S.M., Stix J., Berlo K., Petrelli M., Sigmundsson F., 2021. Eruption risks from covert silicic magma bodies. Geology, doi:10.1130/G48697.1 

3. Petrelli M.,  Zellmer G.F., 2022. Rates and Timescales of Magma Transfer, Storage, Emplacement, and Eruption. In Dynamic Magma Evolution (Francesco Vetere Editor), Geophysical Monograph Series, 254, American Geophysical Union and Wiley.

4. Petrelli M.,  El Omari K.,  Spina L.,  Le Guer, Y.,  La Spina G.,  Perugini D., 2018. Timescales of water accumulation in magmas and implications for short warning times of explosive eruptions. Nature Communications, doi: 10.1038/s41467-018-02987-6

5. Petrelli M., El Omari K., Le Guer Y., Perugini D., 2016. Effects of chaotic advection on the timescales of cooling and crystallization of magma bodies at mid crustal levels. Geochemistry, Geophysics, Geosystems, doi: 10.1002/2015GC006109 

6. Petrelli, M., Perugini, D., Poli, G., 2011. Transition to chaos and implications for time-scales of magma hybridization during mixing processes in magma chambers. Lithos 125, 211-22

In 2001, I started the development of the LA-ICP-MS laboratory at the Department of Geology, University of Perugia. Since 2002, I have successfully run the LA-ICP-MS with activities that include the maintenance of the instrumentations, the development of analytical protocols, the analysis of natural and experimental samples, and the support of external users. Currently, the LA-ICP-MS lab in Perugia hosts two LA-ICP-MS systems, and it is actively involved in many scientific projects.

Selected Publications:

1.  Petrelli M., Laeger K., Perugini D. (2016). High spatial resolution trace element determination of geological samples by laser ablation quadrupole plasma mass spectrometry: implications for glass analysis in volcanic products. Geosciences Journal. DOI: 10.1007/s12303-016-0007-z

2. Alagna K.E., Petrelli M., Perugini D., Poli G. (2008). Micro‐analytical zircon and monazite U‐Pb isotope dating by laser ablation‐inductively coupled plasma‐quadrupole mass spectrometry. Geostandards and Geoanalytical research. DOI: 10.1111/j.1751-908X.2008.00866.x

3. Petrelli M., Caricchi L., Ulmer P. (2007). Application of high spatial resolution laser ablation ICP‐MS to crystal‐melt trace element partition coefficient determination. Geostandards and Geoanalytical Research. DOI: 10.1111/j.1751-908X.2007.00825.x