To check my works in other research areas go here.

My research has consistently been guided by a naturalistic attitude to knowledge and by the belief that science has the epistemic authority to reveal the structure of our physical reality. However, how can we ensure that we are accurately deriving the physical ontology from science? The history of science shows that we often believed in entities that turned out not to exist. Deriving ontological commitments from science is a complex task which needs to take into consideration the problems of interpreting theories, deciding among them, and the limits of the methods that lead to their creation. Much of my work is therefore devoted to implementing a sophisticated and cautious form of scientific realism, by pursuing a multi-faceted research agenda that integrates issues arising in the metaphysics of science on ontological interpretations of physical theories with critical examination of aspects of the scientific method. 

In the metaphysics of science, I have conducted extensive research on deriving ontological commitments from physical theories, primarily within non-relativistic quantum mechanics. For my PhD, I focused on Bohmian mechanics, developing a framework for deriving physical ontology from physical theories and presenting my own proposal, which aligns with ontic structural realism. Afterward, I expanded my focus to quantum mechanics in general, exploring a range of ontological perspectives—from wavefunction realism to super-humean interpretations—and analyzing the virtues and challenges of each. I further applied this approach to quantum gravity, specifically examining loop quantum gravity. Given that both quantum mechanics and quantum gravity grapple with the problem of underdetermination, I sought metaphysical principles that could address underdetermination. For cases like non-relativistic quantum mechanics, in which underdetermination is inescapable, I developed a fictionalist account of the objects whose ontological status is controversial. The fictionalist view that I developed is not inherently anti-realist, as it claims that as long as underdetermination persists, we cannot but only recognize the fictionalist status of the controversial entities. Fictionalism has already been applied in scientific model literature, and it was also inspired by my collaboration with astrophysicists during my second postdoc, who frequently employ fictional entities in their models. 

Working with astrophysicists and talking to them about scientific realism applied to models also spurred my interest in the scientific method. Given the limited data available in astrophysics, I worked on meta-empirical confirmation as a way of reasoning that can be used to test the viability of certain hypotheses regarding the nature of astrophysical objects. It can also be used to understand how atomspheric retrievial models functions, which are models that use ‘intelligent’ alghoritms to find the best specification of the atmosphere of exoplanets. Given the high-replicability nature of the field, I also worked on replicability, and extended the epistemic considerations that I learnt from how astrophysicists implement replicability to other fields of research, such as psychology. My work in the philosophy of astrophysics has led me to co-edit the first volume on the philosophy of astrophysics with Nora Boyd (Siena College), Kevin Heng (LMU), and Siska de Baedermaker (Stockholm University).

In my next project, I plan to extend my research by investigating what realist commitments we should adopt for the large-scale universe, addressing and integrating questions within both metaphysics of science and the scientific method. In the metaphysics of science, I intend to explore dark energy, which according to the standard cosmological models makes up the 75% of our universe) using interpretative tools developed in my work on quantum mechanics and quantum gravity. Dark energy presents a notable challenge due to the underdetermination of different empirical models (e.g., the Λ-constant, quintessence, and phantom energy models). My approach is to develop functionalist and fictionalist perspectives: a functionalist account would focus on dark energy’s causal role in accelerating the universe’s expansion, while a fictionalist account would emphasize its provisional status as a theoretical placeholder. I believe these complementary approaches can provide metaphysical insights into dark energy within the ΛCDM model, accommodating the underdetermination issue currently faced. I will also address crucial questions about the ΛCDM model itself. While this model is commonly regarded as a significant scientific achievement, I have suggested that it exhibits features like plasticity and path-dependency, which are often indicators that the model is not truly representing its target system. If the ΛCDM model indeed possesses anti-realist characteristics, a significant challenge arises: how can we justify realism within astrophysics and cosmology if our primary model for understanding the universe is not robustly realist? I am committed to developing a nuanced and justified form of realism to address this concern.

Given the increasing reliance on AI-driven models and simulations in astrophysics, my future research will also include two related projects on AI’s impact on scientific discovery and understanding, particularly. There is a concern that AI’s role in accelerating scientific discovery may outpace our capacity for understanding, potentially creating a gap between discovery and comprehension. Building on my previous research on the epistemic status of retrieval models, I aim to conduct case studies to determine whether AI can genuinely produce objective, truth-conducive knowledge. I will examine how AI has altered concepts of scientific discovery, scientific knowledge, and the role of the epistemic agent. In a recent paper, I argued that machine learning's reliance on optimization poses risks, potentially transforming science into a purely optimization-driven process. To counteract this, I proposed implementing 'scientific well-being' through AI, particularly by fostering diversity within scientific research. In another paper, I explored whether machine learning can provide understanding in science, ultimately advocating for a positive view of AI’s potential to foster understanding.