Soumyendu Sarkar
Senior Director & Senior Distinguished Technologist, Hewlett Packard Enterprise
Leads ML Research @ Hewlett Packard Enterprise Labs
Principal Investigator on US DOE funded AI Research for Fusion Energy
Principal Investigator for HPE at the U.S. AI Safety Institute Consortium (AISIC), NIST
Resaerch Profile and Publications
Color Codes: Reinforcement Learning, LLMs, Computer Vision, Nuclear Fusion
Best Paper Awards
[1] NeurIPS 2023 Workshop on Tackling Climate Change with Machine Learning,
Best Paper Award for ML Innovation,
Real-time Carbon Footprint Minimization in Sustainable Data Centers with Reinforcement Learning
Sarkar, Soumyendu and Naug, Avisek and Gutierrez, Ricardo Luna and Guillen, Antonio and Gundecha, Vineet and Babu, Ashwin Ramesh and Bash, Cullen,
https://s3.us-east-1.amazonaws.com/climate-change-ai/papers/neurips2023/28
[2] IEEE 18th International Conference on Automation Science and Engineering (CASE), 2022
Best Application Paper Award
Skip Training for Multi-Agent Reinforcement Learning Controller for Industrial Wave Energy Converters,
Sarkar, Soumyendu and Gundecha, Vineet and Ghorbanpour, Sahand and Shmakov, Alexander and Babu, Ashwin Ramesh and Pichard, Alexandre and Cocho, Mathieu,
Proceedings of the IEEE 18th International Conference on Automation Science and Engineering (CASE), 2022
Selected Papers (Recent):
[1] Reinforcement Learning Platform for Adversarial Black-box Attacks with Custom Distortion Filters
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Gundecha, Vineet and Ghorbanpour, Sahand and Naug, Avisek and Gutierrez, Ricardo Luna and Guillen, Antonio,
AAAI 2025, Proceedings of the AAAI Conference on Artificial Intelligence, volume=39, number=26, pages=27628--27635
https://doi.org/10.1609/aaai.v39i26.34976
[2] Hierarchical Multi-Agent Framework for Carbon-Efficient Liquid-Cooled Data Center Clusters
Sarkar, Soumyendu and Naug, Avisek and Guillen, Antonio and Gundecha, Vineet and Gutierrez, Ricardo Luna and Ghorbanpour, Sahand and Mousavi, Sajad and Babu, Ashwin Ramesh and Rengarajan, Desik and Bash, Cullen},
AAAI 2025, Proceedings of the AAAI Conference on Artificial Intelligence, volume=39, number=28, pages=29694--29696
https://doi.org/10.1609/aaai.v39i28.35370
[3] SustainDC: Benchmarking for Sustainable Data Center Control
Naug, Avisek and Guillen, Antonio and Luna, Ricardo and Gundecha, Vineet and Rengarajan, Desik and Ghorbanpour, Sahand and Mousavi, Sajad and Babu, Ashwin Ramesh and Markovikj, Dejan and Kashyap, Lekhapriya D and others,
NeurIPS 2024 Proceedings of the Advances in Neural Information Processing Systems, volume=37, pages=100630--100669
[4] Carbon Footprint Reduction for Sustainable Data Centers in Real-Time
Sarkar, Soumyendu and Naug, Avisek and Luna, Ricardo and Guillen, Antonio and Gundecha, Vineet and Ghorbanpour, Sahand and Mousavi, Sajad and Markovikj, Dejan and Babu, Ashwin Ramesh,
AAAI 2024, Proceedings of the AAAI Conference on Artificial Intelligence, volume=38, number=20, pages=22322--22330
[5] Robustness and visual explanation for black box image, video, and ECG signal classification with reinforcement learning
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Gundecha, Vineet and Naug, Avisek and Ghorbanpour, Sahand,
AAAI 2024, Proceedings of the AAAI Conference on Artificial Intelligence, volume=38, number=21, pages=23829--23831
[6] Sustainability of data center digital twins with reinforcement learning
Sarkar, Soumyendu and Naug, Avisek and Guillen, Antonio and Luna, Ricardo and Gundecha, Vineet and Babu, Ashwin Ramesh and Mousavi, Sajad,
AAAI 2024, Proceedings of the AAAI Conference on Artificial Intelligence, volume=38, number=21, pages=23832--23834
[7] Benchmark generation framework with customizable distortions for image classifier robustness
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Carmichael, Zachariah and Gundecha, Vineet and Ghorbanpour, Sahand and Gutierrez, Ricardo Luna and Guillen, Antonio and Naug, Avisek,
WACV 2024, Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2024, pages=4418--4427
[8] Function Approximation for Reinforcement Learning Controller for Energy from Spread Waves
Sarkar, Soumyendu and Gundecha, Vineet and Ghorbanpour, Sahand and Shmakov, Alexander and Babu, Ashwin Ramesh and Naug, Avisek and Pichard, Alexandre and Cocho, Mathieu,
IJCAI 2023, Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence, 2023
pages=6201--6209,
International Joint Conferences on Artificial Intelligence Organization}
[9] Multi-Agent Reinforcement Learning Controller to Maximize Energy Efficiency for Multi-Generator Industrial Wave Energy Converter
Soumyendu Sarkar, Vineet Gundecha, Alexander Shmakov, Sahand Ghorbanpour, Ashwin Ramesh Babu, Paolo Faraboschi, Mathieu Cocho, Alexandre Pichard, Jonathan Fievez,
AAAI '22, Proceedings of the AAAI Conference on Artificial Intelligence, volume=36, number=21, pages=23832--23834
Other Papers
2025
Robustness Evaluation for Video Models with Reinforcement Learning,
Babu, Ashwin Ramesh and Mousavi, Sajad and Gundecha, Vineet and Ghorbanpour, Sahand and Naug, Avisek and Guillen, Antonio and Gutierrez, Ricardo Luna and Sarkar, Soumyendu},
Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR) Workshop, 2025, pages=4362--a4370
Coordinated Robustness Evaluation Framework for Vision-Language Models
Babu, Ashwin Ramesh and Mousavi, Sajad and Gundecha, Vineet and Ghorbanpour, Sahand and Naug, Avisek and Guillen, Antonio and Luna, Ricardo and Sarkar, Soumyendu
Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR) Workshop, 2025, pages={712--720}
2024
iART-Imitation guided Automated Red Teaming,
Mousavi, Sajad and Rengarajan, Desik and Babu, Ashwin Ramesh and Gundecha, Vineet and Naug, Avisek and Ghorbanpour, Sahand and Gutierrez, Ricardo Luna and Guillen, Antonio and Faraboschi, Paolo and Sarkar, Soumyendu,
NeurIPS 2024 Red Teaming GenAI Workshop - Oral Presentation
Informed Tree of Thought: Cost-efficient Problem Solving with Large Language Models,
Mousavi, Sajad and Rengarajan, Desik and Babu, Ashwin Ramesh and Ghorbanpour, Sahand and Gundecha, Vineet and Naug, Avisek and Sarkar, Soumyendu,
NeurIPS 2024 Workshop on Adaptive Foundation Models: Evolving AI for Personalized and Efficient Learning
LLM Enhanced Bayesian Optimization for Scientific Applications like Fusion,
Ghorbanpour, Sahand and Gutierrez, Ricardo Luna and Gundecha, Vineet and Rengarajan, Desik and Babu, Ashwin Ramesh and Sarkar, Soumyendu,
NeurIPS 2024 Workshop on Physical Sciences
Optimizing the Performance of Direct-Drive Implosion Experiments Using Meta-Bayesian Optimization,
Ejaz, Rahman and Gopalaswamy, Varchas and Luna, Ricardo and Gundecha, Vineet and Lees, Aarne and Betti, Riccardo and Ghorbanpour, Sahand and Sarkar, Soumyendu and Kanan, Christopher,
Bulletin of the American Physical Society, 2024, Publisher: American Physical Society
Explainable meta bayesian optimization with human feedback for scientific applications like fusion energy,
Gutierrez, Ricardo Luna and Ghorbanpour, Sahand and Gundecha, Vineet and Ejaz, Rahman and Gopalaswamy, Varchas and Betti, Riccardo and Naug, Avisek and Rengarajan, Desik and Babu, Ashwin Ramesh and Faraboschi, Paolo and others},
NeurIPS 2024 Workshop on Tackling Climate Change with Machine Learning
Meta-Learned Bayesian Optimization for Energy Yield in Inertial Confinement Fusion,
Gundecha, Vineet and Gutierrez, Ricardo Luna and Ghorbanpour, Sahand and Ejaz, Rahman and Gopalaswamy, Varchas and Betti, Riccardo and Rengarajan, Desik and Sarkar, Soumyendu,
NeurIPS 2024 Workshop on Physical Sciences
Digital twins for data centers,
Athavale, Jyotika and Bash, Cullen and Brewer, Wesley and Maiterth, Matthias and Milojicic, Dejan and Petty, Harry and Sarkar, Soumyendu,
IEEE Computer Magazine, Volume={57}, Issue={10}, pages={151--158}, 2024
Enhancing Reinforcement Learning-Based Control of Wave Energy Converters Using Predictive Wave Modeling,
Gundecha, Vineet and Paap, Arie and Cocho, Mathieu and Ghorbanpour, Sahand and Pichard, Alexandre and Babu, Ashwin Ramesh and Sarkar, Soumyendu
NeurIPS 2024 Workshop on Tackling Climate Change with Machine Learning
Carbon-Aware Spatio-Temporal Workload Distribution in Cloud Data Center Clusters Using Reinforcement Learning,
Sarkar, Soumyendu and Guillen-Perez, Antonio and Gundecha, Vineet and Naug, Avisek and Gutierrez, Ricardo Luna and Mousavi, Sajad and Faraboschi, Paolo and Bash, Cullen
NeurIPS 2024 Workshop on Tackling Climate Change with Machine Learning
Enhancing Sustainability in Liquid-Cooled Data Centers with Reinforcement Learning Control,
Naug, Avisek and Perez, Antonio Guillen and Gundecha, Vineet and Gutierrez, Ricardo Luna and Babu, Ashwin Ramesh and Mousavi, Sajad and Faraboschi, Paolo and Bash, Cullen and Sarkar, Soumyendu,
NeurIPS 2024 Workshop on Tackling Climate Change with Machine Learning
CFD Surrogates for Data Center Sustainability Using 3D U-Net Convolutional Neural Network,
Sarkar, Soumyendu and Guillen-Perez, Antonio and Carmichael, Zachariah and Gundecha, Vineet and Naug, Avisek and Gutierrez, Ricardo Luna and Babu, Ashwin Ramesh and Bash, Cullen,
2024 23rd IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm),
Imitation Guided Automated Red Teaming,
Rengarajan, Desik and Mousavi, Sajad and Babu, Ashwin Ramesh and Gundecha, Vineet and Naug, Avisek and Ghorbanpour, Sahand and Guillen, Antonio and Gutierrez, Ricardo Luna and Sarkar, Soumyendu,
NeurIPS 2024 Workshop on Safe Generative AI
2023
Direct-Drive Implosion Performance Optimization Using Gaussian Process Modeling and Reinforcement Learning,
Ejaz, Rahman and Gopalaswamy, Varchas and Lees, Aarne and Cao, Duc and Sarkar, Soumyendu and Kanan, Christopher,
APS Division of Plasma Physics Meeting Abstracts, 2023
Pydcm: Custom data center models with reinforcement learning for sustainability,
Naug, Avisek and Guillen, Antonio and Luna Gutierrez, Ricardo and Gundecha, Vineet and Ghorbanpour, Sahand and Dheeraj Kashyap, Lekhapriya and Markovikj, Dejan and Krause, Lorenz and Mousavi, Sajad and Babu, Ashwin Ramesh and Sarkar, Soumyendu,
Proceedings of the 10th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation, 2023, pages={232--235},
Robustness with query-efficient adversarial attack using reinforcement learning,
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Ghorbanpour, Sahand and Gundecha, Vineet and Guillen, Antonio and Luna, Ricardo and Naug, Avisek,
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshop, 2023, pages={2330--2337},
Rl-cam: Visual explanations for convolutional networks using reinforcement learning,
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Ghorbanpour, Sahand and Gundecha, Vineet and Guillen, Antonio and Luna, Ricardo and Naug, Avisek,
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshop, 2023, pages=3861--3869
Reinforcement learning based black-box adversarial attack for robustness improvement,
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Ghorbanpour, Sahand and Gundecha, Vineet and Gutierrez, Ricardo Luna and Guillen, Antonio and Naug, Avisek,
2023 IEEE CASE, 19th International Conference on Automation Science and Engineering
Concurrent carbon footprint reduction (c2fr) reinforcement learning approach for sustainable data center digital twin,
Sarkar, Soumyendu and Naug, Avisek and Guillen, Antonio and Gutierrez, Ricardo Luna and Ghorbanpour, Sahand and Mousavi, Sajad and Babu, Ashwin Ramesh and Gundecha, Vineet,
2023 IEEE CASE, 19th International Conference on Automation Science and Engineering
Rtdk-bo: High dimensional bayesian optimization with reinforced transformer deep kernels,
Shmakov, Alexander and Naug, Avisek and Gundecha, Vineet and Ghorbanpour, Sahand and Gutierrez, Ricardo Luna and Babu, Ashwin Ramesh and Guillen, Antonio and Sarkar, Soumyendu},
2023 IEEE CASE, 19th International Conference on Automation Science and Engineering
Natural language processing workflow,
Sarkar, Soumyendu and Rawat, Rohit and Gundecha, Vineet and Ghorbanpour, Sahand and Jindal, Abhishek and Faraboschi, Paolo,
2023, US Patent App. 17/498,124
Robustness with Black-Box Adversarial Attack using Reinforcement Learning,
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Sajad and Gundecha, Vineet and Ghorbanpour, Sahand and Shmakov, Alexander and Gutierrez, Ricardo Luna and Guillen, Antonio and Naug, Avisek,
AAAI 2023: Proceedings of the Workshop on Artificial Intelligence Safety 2023 (SafeAI 2023), Oral Presentation, volume={3381}, year={2023},
https://ceur-ws.org/Vol-3381/8.pdf
N-Critics: Self-Refinement of Large Language Models with Ensemble of Critics,
Mousavi, Sajad and Guti{\'e}rrez, Ricardo Luna and Rengarajan, Desik and Gundecha, Vineet and Babu, Ashwin Ramesh and Naug, Avisek and Guillen, Antonio and Sarkar, Soumyendu,
NeurIPS 2023 Foundation Models for Decision Making Workshop,
https://openreview.net/forum?id=L7vC3OYb2p
Enhancing Large Language Models with Ensemble of Critics for Mitigating Toxicity and Hallucination,
Mousavi, Sajad and Gutierrez, Ricardo Luna and Rengarajan, Desik and Gundecha, Vineet and Babu, Ashwin Ramesh and Naug, Avisek and Guillen, Antonio and Sarkar, Soumyendu,
NeurIPS R0-FoMo Workshop: Robustness of Few-shot and Zero-shot Learning in Large Foundation Models
A configurable pythonic data center model for sustainable cooling and ml integration,
Naug, Avisek and Guillen, Antonio and Gutierrez, Ricardo Luna and Gundecha, Vineet and Ghorbanpour, Sahand and Mousavi, Sajad and Babu, Ashwin Ramesh and Sarkar, Soumyendu,
NeurIPS 2023 Workshop on Tackling Climate Change with Machine Learning,
https://s3.us-east-1.amazonaws.com/climate-change-ai/papers/neurips2023/15
Sustainable data center modeling: A multi-agent reinforcement learning benchmark,
Sarkar, Soumyendu and Naug, Avisek and Guillen, Antonio and Luna Gutierrez, R and Gundecha, Vineet and Ghorbanpour, Sahand and Mousavi, Sajad and Ramesh Babu, Ashwin,
NeurIPS 2023 Workshop on Tackling Climate Change with Machine Learning
Enhancing data center sustainability with a 3d cnn-based cfd surrogate model,
Sarkar, Soumyendu and Guillen, Antonio and Carmichael, Zachariah and Gundecha, Vineet and Naug, Avisek and Ramesh Babu, A and Luna Gutierrez, Ricardo,
NeurIPS 2023 Workshop on Tackling Climate Change with Machine Learning,
2022
Multi-agent reinforcement learning controller to maximize energy efficiency for multi-generator industrial wave energy converter,
Sarkar, Soumyendu and Gundecha, Vineet and Shmakov, Alexander and Ghorbanpour, Sahand and Babu, Ashwin Ramesh and Faraboschi, Paolo and Cocho, Mathieu and Pichard, Alexandre and Fievez, Jonathan},
AAAI, Proceedings of the AAAI Conference on Artificial Intelligence, 2022, volume={36}, number={11}, pages={12135--12144}, year={2022}
Function approximations for reinforcement learning controller for wave energy converters,
Sarkar, Soumyendu and Gundecha, Vineet and Ghorban-pour, Sahand and Shmakov, Alexander and Babu, Ashwin Ramesh and Pichard, Alexandre and Cocho, Mathieu and Enterprise, Hewlett Packard,
NeurIPs Tackling Climate Change with Machine Learning Workshop, 2022,
https://s3.us-east-1.amazonaws.com/climate-change-ai/papers/neurips2022/1
Measuring robustness with black-box adversarial attack using reinforcement learning,
Sarkar, Soumyendu and Mousavi, Sajad and Babu, Ashwin Ramesh and Gundecha, Vineet and Ghorbanpour, Sahand and Shmakov, Alexander K,
NeurIPS ML Safety Workshop, 2022
2021 and Earlier
Effectiveness of deep networks in NLP using BiDAF as an example architecture,
Sarkar, Soumyendu @ Stanford, arXiv preprint arXiv:2109.00074, year={2021}
Multi-objective Reinforcement Learning Controller for Multi-Generator Industrial Wave Energy Converter,
Sarkar, Soumyendu and Gundecha, Vineet and Shmakov, Alexander and Ghorbanpour, Sahand and Ramesh Babu, Ashwin and Faraboschi, Paolo and Cocho, Mathieu and Pichard, Alexandre and Fievez, Jonathan,
NeurIPs Tackling Climate Change with Machine Learning Workshop, 2021
https://s3.us-east-1.amazonaws.com/climate-change-ai/papers/neurips2021/51
Image tracking objects associated with objects of interest,
Sarkar, Soumyendu, 2021,
US Patent 20,210,289,171
All-in-focus implementation,
Sarkar, Soumyendu and Burns, Geoffrey Francis}, 2017
publisher={Google Patents},
US Patent 9,554,039
REAL-TIME CARBON FOOTPRINT REDUCTION CONTROLLER,
Sarkar, Soumyendu and Naug, Avisek and Luna Gutierrez, Ricardo and Guillen Perez, Antonio and Ghorbanpour, Sahand},
year={2024},
US Patent App. 18/492,176
HPE’s Natural Language Processing platform for Question and Answer,
Sarkar, Soumyendu,
year={2020}
Survey of LLM Agent Communication with MCP: A Software Design Pattern Centric Review,
Sarkar, Anjana and Sarkar, Soumyendu},
url={https://www.researchgate.net/publication/392083326\_Survey\_of\_LLM\_Agent\_Communication\_with\_MCP\_A\_Software\_Design\_Pattern\_Centric\_Review}, year={2025}
Review of LLM based Control,
Sarkar, Soumyendu and Ghorbanpour, Sahand and Gutierrez, Ricardo Luna and Guillen-Perez, Antonio and Gundecha, Vineet and Babu, Ashwin Ramesh and Naug, Avisek},
url={doi.org/10.13140/RG.2.2.28595.54567}, year={2025}
Reinforcement learning agent to measure robustness of black-box image classification models,
Sarkar, Soumyendu and Babu, Ashwin Ramesh and Mousavi, Seyed Sajad and Gundecha, Vineet and Ghorbanpour, Sahand and Naug, Avisek},
2024,
US Patent App. 18/486,756