Mel Andrews: Ontology of the Free Energy Principle and the Philosophy of Machine Learning
Jul 5, 2024
auto_awesome
Mel Andrews, a philosopher of science specializing in machine learning, delves into the philosophy of AI, ethics, and the ontology of the free energy principle. They discuss challenges in academia, capitalist influences on AI ethics, the evolution of machine learning research, and the intersection of philosophy with AI and critical theory.
Machine learning in science should align with standards of good science, avoiding pseudoscience pitfalls.
Historical ignorance in ML research contributes to oversimplified problem-solving approaches and substandard findings.
Data-driven methods must integrate conceptual frameworks to uphold scientific epistemology, bridging the gap with ML applications.
Deep dives
The Influence of Deep Learning on Scientific Discovery
There has been a historical debate in 20th-century philosophy of science regarding the possibility of automated or algorithmically implemented science. However, with the advent of the deep learning revolution, there has been a significant shift towards using intelligent computational systems for scientific discovery. The application of machine learning in extracting statistical patterns from data is viewed as a science-adjacent activity that should adhere to the standards of good science.
Challenges in Machine Learning and Science
Machine learning faces a pseudoscience problem due to its rapid adoption and the prevalence of hype narratives surrounding its capabilities. The lack of in-depth historical understanding in machine learning research leads to oversimplified approaches in addressing complex scientific problems. Additionally, the inadequate peer review process in machine learning venues contributes to the proliferation of substandard research, thus posing challenges to the credibility of scientific findings.
Theory-Free Ideal and Epistemic Issues in Machine Learning
The theory-free ideal in science, particularly in some disciplines like quantitative social sciences, leads to a belief that data-driven methods are more objective and scientific. However, this contradicts the essence of scientific epistemology, which requires prior conceptual infrastructure for empirical knowledge. Machine learning's sociological influences and insufficient understanding of its disciplinary history result in a disconnect between the application of machine learning techniques and the fundamental principles of scientific inquiry.
Exploring the Philosophical Interest in the Free Energy Principle
The Free Energy Principle (FEP) has garnered significant philosophical interest due to its unique approach as a thinking tool. Unlike traditional math used for scientific analysis, FEP allows conceptualization of target phenomena in new and philosophically generative ways. Concepts from Erwin Schrodinger's 'What is Life?' are echoed in FEP, challenging neo-Darwinian ideas by focusing on physical and structural exigencies of biology. FEP highlights the philosophical richness associated with new ways of thinking.
Discussing Conceptual Reification in Scientific Modeling
The concept of conceptual reification in scientific modeling is explored, emphasizing the distinction between models and the natural world. Mistaking the mathematical models for aspects of reality can lead to conceptual confusion. The discussion extends to the realism of conceptual tools used in science, questioning attributions of truth and reality to scientific mathematical constructs. The paper delves into philosophical perspectives on mathematical Platonism and the pragmatic application of mathematical concepts in scientific understanding.
Mel Andrews is a philosopher of science who primarily focuses on machine learning and the role of mathematical and computational methods in scientific modelling. Mel is currently a predoctoral research associate at the Department of Machine Learning at Carnegie Mellon University and doing a PhD in philosophy of science at the University of Cincinnati. They are also a visiting scholar at the Australian National University and the University of Pittsburgh. In this episode, we discuss the philosophy of artificial intelligence and machine learning, AI ethics and safety, scientific and mathematical realism, the ontology of the free energy principle and critical theory's relationship to AI research.
Copyright Disclaimer under section 107 of the Copyright Act 1976, allowance is made for “fair use” for purposes such as criticism, comment, news reporting, teaching, scholarship, education and research.
Fair use is a use permitted by copyright statutes that might otherwise be infringing.
If you are or represent the copyright owner of materials used in this video and have a problem with the use of the related material, please email me at trahulsam@gmail.com, and we can sort it out. Thank you.
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
