Philosophical Foundations of Automated Scientific Discovery

Objetivo

"The Nobel Turing Challenge from 2021 calls for developing ``AI Scientists capable of autonomously carrying out research to make major scientific discoveries worthy of [a] Nobel Prize and beyond by 2050."" (https://www.nobelturingchallenge.org/(se abrirá en una nueva ventana)). A decade before, the Nobel Turing challenge would have seemed ridiculously optimistic: automated discovery was still mainly associated with rule-based systems like DENDRAL, EURISCO, or BACON, which had fallen short of delivering on any promise to enable more general approaches to scientific discovery. However, recent advancements in machine learning, particularly deep neural networks (DNNs) and large language models (LLMs), have significantly reignited work in automated scientific discovery. Symbolic regression has advanced significantly, AI systems are increasingly being presented as ``AI scientists"", and foundation models tailored for scientific applications are being developed.
The project will subject an innovative and potentially game-changing interdisciplinary field of science---that of automated scientific discovery---to a much needed philosophy of science investigation. The project has three major goals: The first goal is to assess the state-of-the-art of automated scientific discovery through the lens of the philosophy of science. The second goal is to design novel benchmarks of automated discovery systems based on so far neglected insights from the philosophy of science. The third goal is to integrate insights from the workings of automated discovery systems into our general understanding of how scientific discovery works. To keep the scope of the project under a controllable size, and to optimise synergy with the applicant’s background, the project deals with automated discovery in physics specifically.
The three goals will be addressed by the PI, a postdoc, and a PhD student by paying close attention to current scientific practice, supported by an international advisory board of experts."

Ámbito científico (EuroSciVoc)

Este proyecto aún no se ha clasificado con EuroSciVoc.
Sugiera los ámbitos científicos que considere más relevantes y ayúdenos a mejorar nuestro servicio de clasificación.