# Sandeep Madireddy

> Computer Scientist at Argonne National Laboratory

## Research Areas

- **Probabilistic Machine Learning**: Research at the intersection of Bayesian Inference and Information-Theoretic learning.
- **Scientific Machine Learning and AI for Science**: Theoretical and Applied Machine learning for Large-scale Science including Fusion Energy, Astrophysics, HPC, and Climate Science
- **Multi-modal Foundation Models**: Theoretical and Applied research on developing and evaluating (for skill and safety) Multimodal Language Models as scientific research assistants and Spatio-Temporal Foundation models for science.
- **Energy Efficient AI**: Energy-efficient AI with Bio-inspired architectures such as Neuromorphics for large scale AI models and life-long learning in this paradigm.

## Publications (113 total)

### Most Cited

- Biological underpinnings for lifelong learning machines (2022) [415 citations]
  - Nature Machine Intelligence 4 (3), 196-210, 2022
- Scaling transformer neural networks for skillful and reliable medium-range weather forecasting (2024) [165 citations]
  - Advances in Neural Information Processing Systems 37, 68740-68771, 2024
- Time-series learning of latent-space dynamics for reduced-order model closure (2020) [156 citations]
  - Physica D: Nonlinear Phenomena 405, 132368, 2020
- A Bayesian approach to selecting hyperelastic constitutive models of soft tissue (2015) [114 citations]
  - Computer Methods in Applied Mechanics and Engineering 291, 102-122, 2015
- Applications and techniques for fast machine learning in science (2022) [110 citations]
  - Frontiers in big Data 5, 787421, 2022
- Application of machine learning and artificial intelligence to extend EFIT equilibrium reconstruction (2022) [80 citations]
  - Plasma Physics and Controlled Fusion 64 (7), 074001, 2022
- DeepMerge–II. Building robust deep learning algorithms for merging galaxy identification across domains (2021) [75 citations]
  - Monthly Notices of the Royal Astronomical Society 506 (1), 677-691, 2021
- Bayesian calibration of hyperelastic constitutive models of soft tissue (2016) [67 citations]
  - Journal of the Mechanical Behavior of Biomedical Materials, Accepted for …, 2016
- DeepAstroUDA: semi-supervised universal domain adaptation for cross-survey galaxy morphology classification and anomaly detection (2023) [64 citations]
  - Machine Learning: Science and Technology 4 (2), 025013, 2023
- HPC I/O throughput bottleneck analysis with explainable local models (2020) [49 citations]
  - SC20: International Conference for High Performance Computing, Networking …, 2020
- DeepAdversaries: examining the robustness of deep learning models for galaxy morphology classification (2022) [45 citations]
  - Machine Learning: Science and Technology 3 (3), 035007, 2022
- Improving Scalability of Parallel CNN Training by Adjusting Mini-Batch Size at Run-Time (2019) [45 citations]
  - 2019 IEEE International Conference on Big Data (Big Data), 830-839, 2019
- A domain-agnostic approach for characterization of lifelong learning systems (2023) [43 citations]
  - Neural Networks 160, 274-296, 2023
- Machine learning based parallel I/O predictive modeling: A case study on Lustre file systems (2018) [41 citations]
  - International Conference on High Performance Computing, 184-204, 2018
- Ailuminate: Introducing v1. 0 of the ai risk and reliability benchmark from mlcommons (2025) [37 citations]
  - arXiv preprint arXiv:2503.05731, 2025
- Adaptive Learning for Concept Drift in Application Performance Modeling (2019) [29 citations]
  - Proceedings of the 48th International Conference on Parallel Processing, 1-11, 2019
- Analysis and correlation of application I/O performance and system-wide I/O activity (2017) [27 citations]
  - 2017 International Conference on Networking, Architecture, and Storage (NAS …, 2017
- Phase segmentation in atom-probe tomography using deep learning-based edge detection (2019) [26 citations]
  - Scientific reports 9 (1), 20140, 2019
- Hpc storage service autotuning using variational-autoencoder-guided asynchronous bayesian optimization (2022) [24 citations]
  - 2022 IEEE International Conference on Cluster Computing (CLUSTER), 381-393, 2022
- AstroMLab 1: Who wins astronomy jeopardy!? (2025) [23 citations]
  - Astronomy and Computing 51, 100893, 2025

### Recent Publications

- GLANCED-IO: Taming I/O Optimization for Deep Learning at Scale (2026)
- KORAL: Knowledge Graph Guided LLM Reasoning for SSD Operational Analysis (2026)
  - arXiv preprint arXiv:2602.10246, 2026
- Multi-task Modeling for Engineering Applications with Sparse Data (2026)
  - arXiv preprint arXiv:2601.05910, 2026
- Uncovering Physical Drivers of Dark Matter Halo Structures with Auxiliary-Variable-Guided Generative Models (2026)
  - arXiv preprint arXiv:2602.23518, 2026
- Aeris: Argonne earth systems model for reliable and skillful predictions (2025)
  - Proceedings of the International Conference for High Performance Computing …, 2025
- Ailuminate: Introducing v1. 0 of the ai risk and reliability benchmark from mlcommons (2025)
  - arXiv preprint arXiv:2503.05731, 2025
- AstroMLab 1: Who wins astronomy jeopardy!? (2025)
  - Astronomy and Computing 51, 100893, 2025
- AstroMLab 1 (2025)
- AuroraGPT Data Collection Interface (2025)
  - Argonne National Laboratory (ANL), Argonne, IL (United States), 2025
- Automated MCQA Benchmarking at Scale: Evaluating Reasoning Traces as Retrieval Sources for Domain Adaptation of Small Language Models (2025)
  - Proceedings of the SC'25 Workshops of the International Conference for High …, 2025

## Talks (52 total: 11 invited, 41 contributed)

- Language Model Evaluation and Safety for Scientific Tasks [Invited] — Argonne Training Program on Extreme-Scale Computing (2025)
- Half Day Tutorial: Evaluation of AI Model Scientific Skills — Trillion Parameter Consortium's 2025 all-hands conference and exhibition (2025)
- Multi-diagnostic Generative modeling of Edge-Localized Modes in Tokamak plasma — SciDAC PI Meeting (Poster) (2025)
- Scalable Flow based Generative Models for Science — SciDAC PI Meeting (Poster) (2025)
- Half Day Tutorial: Leveraging and Evaluation of LLMs For HPC Research — ISC High Performance Computing Conference (2025)
- Uncertainty Quantifying and Propagation in Large Language Models for Single and Multi Step tasks [Invited] — 17th Joint Laboratory on Extreme Scale Computing Workshop (2025)
- Full Day Tutorial: Leveraging and Evaluation of LLMs as Scientific Assistants — Trillion Parameter Consortium Winter Workshop (2025)
- Question Rephrasing for Quantifying Uncertainty in Large Language Models: Applications in Molecular Chemistry Tasks — NeurIPS 2024 Workshop on Statistical Foundations of LLMs and Foundation Models (Poster and short talk) (2024)
- Establishing a Methodology to Evaluate Large Language Models for Science — Workshop on Accelerating the Development and Use of Generative AI for Science and Engineering at Super Computing 2024 (2024)
- AI Safety Research Directions at Argonne — Trillion Parameter Consortium Hackathon (2024)

## Projects

- **Atoms to Manufacturing**: Deep transfer learning to automatically segment the precipitate from matrix in 3D Atom Probe Tomography data.
- **CETOP - A Center for Edge of Tokamak OPtimization**: The project overarching objective is to develop the simulation capability and to perform extended MHD and (drift-gyro) kinetic simulations of non-ELMing (and some ELMing) regime operating points to close gaps in understanding, prediction, and optimization of edge stability for an FPP. I am leading a team to develop ML techniques for extracting reduced-order models, data reduction, and feature extraction to the existing non-ELM database (based on interpolation of data), and to extrapolate to new parameter regimes (such as coil currents for negative-triangularity shaping) for ELM-free optimization.
- **Dynamic architectures through introspection and neuromodulation (DARPA's Lifelong Learning Machines Program)**: Employing architechtures inspired by insect brain to devise efficient, life-long learning machines.
- **ML Assisted Equilibrium Reconstruction for Tokamak Experiments and Burning Plasmas**: Deelop a framework for efficient and accurate equilibrium reconstructions, by automating and maximizing the information extracted from measurements and by leveraging physics-informed ML models constructed from experimental and synthetic solution databases to guide the search for the solution vector.
- **High-Velocity Artificial Intelligence for HEP**: Develop cross-cutting artificial intelligence framework for fast inference and training on heterogeneous computing resources as well as algorithmic advances in AI explainability and uncertainty quantification.
- **Self-Aware Adaptive Workflow and Data Management Services for Future HPC Systems**: Develop modular characterization approaches, where we can examine key performance parameters and application execution similarities.
- **Improving Computational Science Throughput via Model-Based I/O Optimization (SciDAC SUPER-SDAV)**: Machine learning-based probabilistic I/O performance models that take the background traffic, and system state into account while prediciting application performance on HPC system.
- **Foundations for Correctness Checkability and Performance Predictability of Systems at Scale (ScaleSTUDS)**: Multi-dimensional automated scalability tests, program analysis, performance learning and prediction at various levels of the software/hardware stack.
- **Probabilistic Machine Learning for Rapid Large-Scale and High-Rate Aerostructure Manufacturing (PRISM)**: he project goal is to develop a probabilistic ML framework – PRISM – to improve manufacturing efficiency and demonstrate the technologies on wing spars.
- **Accelerating HEP Science:-Inference and Machine Learning at Extreme Scales**: This project brings together ASCR and HEP researchers to develop and apply new methods and algorithms in the area of extreme-scale inference and machine learning. The research program melds high-performance computing and techniques for ​“big data” analysis to enable new avenues of scientific discovery.
- **Enabling Cosmic Discoveries in the Exascale Era**: The overarching objective of this SciDAC-5 project is to create consistent predictions of the dark and visible Universe across redshifts, length scales and wavebands based on state-of-the-art cosmological simulations. The simulation suite will encompass large-volume, high-resolution gravity-only simulations and hydrodynamical simulations equipped with a comprehensive set of subgrid models covering both small and large volumes. The simulations will be coupled to a powerful analysis framework and associated tools to maximize the analysis flexibility and science return.
- **RAPIDS:- A SciDAC Institute for Computer Science and Data**: The goal of RAPIDS (a SciDAC Institute for Resource and Application Productivity through computation, Information, and Data Science) institute is to assist Office of Science (SC) application teams in overcoming computer science and data challenges in the use of DOE supercomputing resources to achieve science breakthroughs.
- **RAPIDS2:- SciDAC Institute for Computer Science, Data, and Artificial Intelligence**: The objective of RAPIDS2 is to assist the Office of Science (SC) application teams in overcoming computer science, data, and AI challenges in the use of DOE supercomputing resources to achieve scientific breakthroughs.
- **A Transformative Co-Design Approach to Materials and Computer Architecture Research (Threadwork)**: Co-design approach that encompasses neuromorphic computing, systems architecture, and datacentric applications. Focus on high energy physics (HEP) and nuclear physics (NP) detector experiments

---

This file is auto-generated at build time from structured content.
For programmatic access, use the MCP endpoint at /api/mcp