| Name: Shatanik Bhattacharya |
| Affiliation: Tata Institute of Fundamental Research |
| Conference ID: ASI2026_474 |
| Title: Accelerating Peak Bagging of red-giant oscillation spectra using Supervised and Reinforcement Learning |
| Abstract Type: Poster |
| Abstract Category: Facilities, Technologies and Data science |
| Author(s) and Co-Author(s) with Affiliation: Shatanik Bhattacharya(Tata Insititute of Fundamental Research, Mumbai-400005, India), Meenakshi Gaira(Tata Insititute of Fundamental Research, Mumbai-400005, India), Rajarshi Barman(Indian Institute of Science, Bengaluru-560012, India; Tata Institute of Fundamental Research, Mumbai-400005, India), Sapnodip Pramanik(Indian Institute of Science Education and Research, Kolkata-741246, India), Shravan M. Hanasoge(Tata Institute of Fundamental Research, Mumbai-400005, India) |
| Abstract: Red-giants (RGs) exhibit mixed dipolar modes of oscillations, which are highly sensitive to the internal structure and dynamics of their cores – rendering them excellent astrophysical laboratories for probing physical phenomena like differential rotation, internal magnetic fields. Despite the availability of an extensive sample of RGs observed by the Kepler telescope—and the anticipated influx of additional data from missions such as TESS and the forthcoming PLATO—the full scientific exploitation of these datasets remains constrained by the conventional methods which require substantial computational time and resources for detailed analysis.
I will demonstrate how we leverage advanced machine learning techniques, including supervised learning methods (like convolutional neural networks, gaussian process optimization, etc.) and reinforcement learning, to automate and accelerate the process of peak bagging to the observed oscillation spectra of RGs without the need for explicit human intervention, smaller number of iterations and thereby smaller computation time. |
