Nesar Ramachandra

Nesar Ramachandra

Computational Scientist, Argonne National Laboratory

Research

My research focuses on developing and applying computational methods at the intersection of astrophysics, cosmology, and machine learning. The work spans foundation models for scientific applications, advanced ML techniques for astronomical data analysis, cosmic structure investigation, and statistical inference methods.

Disclaimer: This section is automatically updated by Reasoning Language Models. Google Gemini is utilized to periodically go over recent publications, talks and activities to update the content. While the information is monitored, at times incorrect information may appear.

AI/ML for Astronomical Discovery

Nesar S Ramachandra's research extensively applies AI/ML to revolutionize astronomical discovery. Key methods include specialized Large Language Models (LLMs) like AstroMLab, demonstrating benchmark-topping performance in astronomy Q&A, achieving GPT-4o level reasoning, and interpreting spectroscopy. Deep learning and neural networks are employed for critical tasks such as point spread function deconvolution, peculiar velocity estimation from the kinetic SZ effect, and detecting/modeling strong gravitational lenses. Generative Adversarial Networks (GANs) are leveraged for anomaly detection in galaxy images. Machine learning also enables synthetic spectra generation for probabilistic redshift estimation (SYTH-Z) and unsupervised exploration of galaxy morphology. This work significantly enhances data analysis, knowledge retrieval, and the discovery of novel astrophysical phenomena.

Learn more about this research →
Figure from AstroMLab 1: Who Wins Astronomy Jeopardy!?
From: AstroMLab 1: Who Wins Astronomy Jeopardy!?
Figure from Neural Network Based Point Spread Function Deconvolution For Astronomical Applications
From: Neural Network Based Point Spread Function Deconvolution For Astronomical Applications
Figure from A Modular Deep Learning Pipeline for Galaxy-Scale Strong Gravitational Lens Detection and Modeling
From: A Modular Deep Learning Pipeline for Galaxy-Scale Strong Gravitational Lens Detection and Modeling
11 Publications 8 Figures Available

Cosmology & Large-Scale Structure

Nesar S Ramachandra's research illuminates Cosmology & Large-Scale Structure through diverse methods. A core focus involves characterizing the dark matter cosmic web using multi-stream analysis, unraveling its topology, geometry, and caustic structure, providing benchmarks for AI-generated structures. Simultaneously, the work constrains modified gravity theories like f(R) using k-cut cosmic shear analyses of Hyper Suprime-Cam data and matter power spectrum emulators. Significant contributions also encompass enhancing observational precision, optimizing galaxy selection for weak lensing cluster mass estimation from surveys like SPTpol. Furthermore, the research applies AI/ML to develop differentiable prediction models like SHAMNet, reducing model error and improving cosmological inferences.

Learn more about this research →
Figure from Matter Power Spectrum Emulator for f(R) Modified Gravity Cosmologies
From: Matter Power Spectrum Emulator for f(R) Modified Gravity Cosmologies
Figure from Multi-stream portrait of the Cosmic web
From: Multi-stream portrait of the Cosmic web
Figure from Topology and geometry of the dark matter web: a multistream view
From: Topology and geometry of the dark matter web: a multistream view
11 Publications 9 Figures Available

Advanced AI/ML Methodology & General Scientific Applications

Nesar S Ramachandra’s research advances AI/ML methodologies for complex scientific and engineering applications. A central focus is enhancing interpretability and robust uncertainty quantification, notably through statistically disentangled latent spaces and probabilistic neural networks. These techniques are applied across diverse domains, including High Energy Physics, high-dimensional stress fields, and particularly fluid dynamics via reduced-order surrogate modeling and global field reconstruction from sparse data. The work also establishes frameworks for evaluating AI models as scientific research assistants, demonstrating a comprehensive approach to creating reliable, explainable, and high-performing AI solutions critical for modern scientific discovery and complex system analysis.

Learn more about this research →
Figure from Application of probabilistic modeling and automated machine learning framework for high-dimensional stress field
From: Application of probabilistic modeling and automated machine learning framework for high-dimensional stress field
Figure from EAIRA: Establishing a Methodology for Evaluating AI Models as Scientific Research Assistants
From: EAIRA: Establishing a Methodology for Evaluating AI Models as Scientific Research Assistants
Figure from Enhancing Interpretability in Generative Modeling: Statistically Disentangled Latent Spaces Guided by Generative Factors in Scientific Datasets
From: Enhancing Interpretability in Generative Modeling: Statistically Disentangled Latent Spaces Guided by Generative Factors in Scientific Datasets
8 Publications 7 Figures Available

Stellar & Galactic Astrophysics

Nesar S Ramachandra’s research comprehensively explores Milky Way structure and stellar populations using vast astronomical datasets, primarily from Gaia. One significant work developed a photometric sample of 2.6 million Red Clump stars, extending from the inner to outer Galaxy. This extensive dataset provides a powerful tracer for mapping Galactic stellar density and architecture. Complementary research identifies Carbon-Enhanced Metal-Poor (CEMP) star candidates through BP/RP spectra from Gaia DR3. These rare, chemically peculiar stars are crucial for understanding early stellar nucleosynthesis and the chemical enrichment history of the universe. Collectively, these studies combine large-scale structural analysis with detailed chemical archaeology, offering vital insights into our Galaxy's formation and evolution.

Learn more about this research →
Figure from Carbon-Enhanced Metal-Poor star candidates from BP/RP Spectra in $Gaia$ DR3
From: Carbon-Enhanced Metal-Poor star candidates from BP/RP Spectra in $Gaia$ DR3
Figure from From the Inner to Outer Milky Way: A Photometric Sample of 2.6 Million Red Clump Stars
From: From the Inner to Outer Milky Way: A Photometric Sample of 2.6 Million Red Clump Stars
2 Publications 2 Figures Available