Combining Collective and Artificial Intelligence to Understand the Early Universe

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K14464
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 15010:Theoretical studies related to particle-, nuclear-, cosmic ray and astro-physics
• Research Institution: The University of Tokyo
• Principal Investigator: Hartwig Tilman 東京大学, 大学院理学系研究科(理学部), 助教 (00843434)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: Artificial Intelligence / Machine Learning / Astronomy / Data Science / Stellar Parameters

Outline of Final Research Achievements

We proposed to develop a novel spectral analysis pipeline, based on convolutional neural networks, to maximize the information gain for the upcoming Subaru Prime Focus Spectrograph. With an international team of collaborators, we started the development of such a tool in order to understand star formation in the early Universe by analyzing the chemical composition of old, metal-poor stars in the Milky Way. We wanted to optimize the convolutional neural network as a citizen science project, preceded by a series of public lectures on Artificial Intelligence (AI). However, we had to adapt our plans due to Covid.
The final pipeline will allow the community to extract stellar parameters, such as chemical composition, of thousands of stars in real time for upcoming observations.

Free Research Field

Astrophysics

Academic Significance and Societal Importance of the Research Achievements

We made significant contributions to the data analysis pipeline of the Prime Focus Spectrograph on the Japanese Subaru Telescope. Despite delays caused by the Covid pandemic, we developed new tools, shared our data with the public, and prepared the scientific data analysis.